Members: Login / Logout
Abstracts

AUDIT POSTER PRIZE SHORTLIST

Endocrinology

  • P001 FSH requesting and NICE guideline 23 in York Teaching NHS Foundation Trust

    D Turnock (Department of Clinical Biochemistry, York Teaching Hospital NHS Foundation Trust, York, United Kingdom)

    Background: NICE Guideline 23 (NG23), Menopause: diagnosis and management was published in November 2015 and made clear recommendations about the use of laboratory tests in the diagnosis of menopause. The guidelines recommend laboratories actively discourage FSH requesting in women over the age of 45 y and suggested that this would provide cost savings. The aim of this audit was to assess local compliance of FSH requesting against NG23 and the effectiveness of strategies to discourage inappropriate FSH requesting.

    Methods: FSH requests, patient demographics and clinical details were obtained from the LIMS (Telepath) for a period of three months. Audit criteria were selected based upon NG23: 1) Women aged >45 y should not have FSH requested to diagnose menopause, 2) A repeat blood sample 4-6 weeks later should be taken if premature ovarian failure is suspected in a woman aged <40 y.

    Results: A total of 819 FSH requests were made in patients aged >45 y in the 3 month audit period, representing 43% of requests made for FSH during this time. Examination of clinical details revealed that 340 (42%) of these requests mentioned menopause or related symptoms. There were 18 patients aged <40 y with results suggestive of premature ovarian failure; 8 patients (44%) had a repeat within 4-6 weeks. A GP information sheet was produced, a pop-up message was added to the electronic pathology ordering system (ICE) and coded comments were added to reports in the LIMS. Re-audit identified 539 FSH requests in patients >45 y in a 3 month period; a reduction of 34% compared to previous.

    Conclusions: There is poor compliance with NG23 in our local area. Education and electronic requesting prompts can reduce inappropriate and promote appropriate FSH requesting but a sustained and targeted approach is required.

Gut Nutrition Trace Elements

  • P002 An audit of nutritional trace element requesting within York Teaching Hospitals NHS Foundation Trust

    R Bramley (Specialist Laboratory Medicine, Leeds Teaching Hospitals NHS Trust, Leeds, United Kingdom), D Turnock (Department of Clinical Biochemistry, York Teaching Hospitals NHS Foundation Trust, York, United Kingdom)

    Background: Measurement of trace elements (zinc, copper and selenium) is often useful in nutritional monitoring, for example post-bariatric surgery or in the context of a malabsorptive disorder. There are limited indications for isolated zinc measurement and inappropriate requests may generate unhelpful and confusing results. At York, requests for zinc-only are processed in house whereas combined copper, zinc and selenium requests are analysed externally. This audit aimed to investigate the reasons for, and frequency of, trace element requesting in York Teaching Hospitals NHS Foundation Trust.

    Methods: Data were gathered from the pathology LIMS and electronic patient record. Audit standards for retesting intervals were taken with reference to NICE guidelines for monitoring of patients on nutritional support (which advise monthly follow up) and the British Obesity and Metabolic Surgery Society (BOMSS) guidelines (which advise annual follow up unless there are specific signs of nutritional deficiency).

    Results: Most (90%) full trace element profiles were appropriate although patients were monitored for bariatric follow up more frequently than guidelines advised (time between repeat samples was ≤6 months in 51% of patients, 6 months to 12 months in 37% and ≥12 months in 12%). Isolated zinc requests were frequently ordered for reasons with an unclear evidence base or had no clear reason for zinc testing provided (40%). Minimum retest intervals were implemented, information sheets about appropriate zinc requesting were sent to GPs and an information box was added to zinc electronic ordering on ICE. A re-audit 3 months later showed a 28% reduction in zinc-only requesting.

    Conclusions: Many reasons for ordering zinc in isolation are inappropriate and there is considerable variation in order frequency for trace element profiles. Minimum re-testing intervals and GP education can be effect in demand management of trace element requests.

  • P003 Copper deficiency: are we identifying it and are we acting on it?

    R Bramley (Specialist Laboratory Medicine, Leeds Teaching Hospitals NHS Trust, Leeds, United Kingdom), C Lippiatt (Specialist Laboratory Medicine, Leeds Teaching Hospitals NHS Trust, Leeds, United Kingdom), E Fox (Specialist Laboratory Medicine, Leeds Teaching Hospitals NHS Trust, Leeds, United Kingdom)

    Background and aims: Copper deficiency is an unusual but important cause of refractory anaemia and permanent neurological damage. Those at risk of copper deficiency include patients on long term parenteral nutrition, patients with malabsorptive conditions (such as coeliac disease or bariatric surgery-induced malabsorption) and individuals taking excess zinc supplements. Our laboratory uses ICP-MS to analyse plasma copper, zinc and selenium simultaneously, which can generate incidental findings of significant copper deficiency when copper measurement has not been requested. This audit aimed to answer the following questions: what were the causes of low plasma copper (<11 µmol/L), how many low results were incidental findings and how were significantly low results followed up?

    Methods

    All trace elements results over a 3 month period were included. For copper results <11 µmol/L with available clinical details, the presumptive cause of low copper was recorded. Where the copper concentration was <7 µmol/L, patient follow-up was investigated in detail.

    Results: Of the 91 low copper cases, 45 were attributable to malabsorptive conditions including bowel resection (17/91), post-bariatric surgery (12/91), Crohn’s disease (10/91) and other malabsorptive conditions (6/91). Other causes included liver disease (16/91) and one case of suspected zinc-induced copper deficiency (ZICD). The cause was undetermined in 15/91 cases. Low copper was discovered incidentally and not initially reported in 15/91 results <11 µmol/L and 2/14 results <7 µmol/L, which led to failure to recognise the ZICD case. All reported results <7 µmol/L were highlighted to clinicians as very low but fewer than half had measurement repeated.

    Conclusions: The reasons for low copper are varied with poor follow up of results by some clinicians. In light of this audit all results <3 µmol/L are now telephoned to clinicians, all incidental findings of plasma copper <7 µmol/L are reported with a comment and a comment is also added where zinc >18 µmol/L and copper <7 µmol/L to query ZICD.

Lipids

  • P004 Audit: the prescription of PCSK9 inhibitors (alirocumab and evolocumab) in a lipid clinic, assessed against NICE guidelines TA393 & TA394

    IAA Appianing (Dept of Clinical Biochemistry, Royal Liverpool University Hospital, Liverpool, United Kingdom), V Mishra (Dept of Clinical Biochemistry, Royal Liverpool University Hospital, Liverpool, United Kingdom)

    Introduction: NICE guidance (June 2016) recommends the use of alirocumab and evolocumab for treating primary hypercholesterolaemia and mixed dyslipidaemia in patients with cardiovascular disease (CVD). These monoclonal antibodies reduce low-density lipoprotein cholesterol (LDL-C) by inhibiting proprotein convertase subtilisin/kexin type 9 (PCSK9), an enzyme involved in down-regulation of low-density lipoprotein receptors.

    AIM: To determine whether PCSK9 inhibitor prescription complied with NICE recommendations.

    Method: This retrospective audit assessed whether all PCSK9-inhibitor prescriptions (July 2016 - Jan 2017) in our Lipid clinic (Aintree University Hospital) occurred at LDL-C concentrations recommended by NICE (TA393 & TA394).

    Results: Twenty-three patients (9 males, 14 females), aged 45 to 88 years, were prescribed evolocumab (n=11) or alirocumab (n=12). They were categorized as follows: primary non-familial hypercholesterolaemia (non-FH) or mixed dyslipidaemia (MD) without CVD (n=3); non-FH or MD with CVD (n=5); primary heterozygous familial hypercholesterolaemia (HeFH) without CVD (n=5); HeFH with CVD (n=10).

    Compliance with NICE recommendations:

    Standard 1: For non-FH or MD without CVD, PCSK9 inhibitors are not recommended. Compliance=0%: Two patients had possible familial combined hyperlipidaemia, LDL-C>5mmol/L, and intolerance to all lipid lowering therapy (LLT). One patient had uncontrolled type 2 diabetes, recurrent pancreatitis, hypercholesterolaemia >10mmol/L, hypertriglyceridaemia > 40 mmol/L despite maximum LLT.

    Standard 2: For non-FH or MD with CVD, PCSK9 inhibitors recommended if LDL-C >4 mmol/L (high risk CVD) or LDL-C >3.5 mmol/L (very high-risk CVD). Compliance=100%

    Standard 3: For HeFH without CVD, PCSK9 inhibitors recommended if LDL-C >5 mmol/L. Compliance=80%. Non-compliant patient’s LDL-C 4.1mmol/L on maximum tolerated therapy

    Standard 4: For HeFH with CVD, PCSK9 inhibitors recommended if LDL-C >3.5 mmol/L. Compliance 70%: 3 non-compliant patients had LDL-C ≥2.7 on maximum tolerated therapy

    Conclusion: Non-compliance with NICE recommendations in this audit illustrates the need for additional LLT in patients intolerant to all LLT and those on maximally tolerated LLT who do not meet CVD LDL-C targets (<1.8 mmol/L).

Management

  • P005 Electronic reflective test requesting at a tertiary oncology centre: results of a pilot study

    P Monaghan (Blood Sciences Department, The Christie Pathology Partnership, Manchester, United Kingdom), S Thirkettle (Blood Sciences Department, The Christie Pathology Partnership, Manchester, United Kingdom), D Swindells (Blood Sciences Department, The Christie Pathology Partnership, Manchester, United Kingdom), A Hussey (Blood Sciences Department, The Christie Pathology Partnership, Manchester, United Kingdom), A Newbery (Transformation Team, The Christie NHS Foundation Trust, Manchester, United Kingdom), R Bramley (Department of Radiology, The Christie NHS Foundation Trust, Manchester, United Kingdom)

    Clinical users of the blood sciences laboratory currently telephone the laboratory to reflectively add-on additional test requests. This presents a challenge for the laboratory as verbal requests are difficult to document to satisfy ISO15189 standards.

    To address this problem we developed an electronic ordering system in the HIS (Clinical Web Portal; CWP). With user convenience in mind, we designed this process so the clinician when viewing results on the CWP could, with a single click order additional tests from an A-Z drop down menu or keyword search term. The system also shows sample type and permitted add-on period (based on in vitro stability) for each test, restricting requests beyond the add-on period. The system also provides status updates to show clinicians which tests have already been ordered, thus averting duplicate requesting.

    The laboratory piloted the new system over 2 weeks with 20 clinicians from 5 clinical specialties (neuroendocrinology, endocrinology, melanoma, lymphoma, breast cancer teams). The 2 most commonly requested tests were TFTs and CA19-9, reflecting the clinical areas participating in the pilot. Endocrine tests accounted for 38% and tumour markers 31% of total tests requested. The largest proportion of tests (65%) requested by a consultant. The neuroendocrine and hepatobiliary clinics requested 37% of all tests, followed by the melanoma team (30%), where appropriate endocrine tests were added in the context of immunotherapy.

    User feedback was very positive; 100% of users rating their first experience of the process good (20%) to excellent (80%), with all users finding the system extremely user-friendly. The new process enables accurate recording of add-on tests to audit requesting trends and laboratory service provision. These data will also be of value to refine test order sets. Electronic reflective requesting improved the efficiency of the end-to-end process; for clinicians requesting and the laboratory processing.

Proteins Enzymes

  • P006 Implications for the monitoring of patients with multiple myeloma during treatment with the immunotherapy daratumumab

    S Thirkettle (Department of Blood Sciences, Christie NHS Foundation Trust, Manchester, United Kingdom), J Cavet (Haematology, Christie NHS Foundation Trust, Manchester, United Kingdom), P Monaghan (Department of Blood Sciences, Christie NHS Foundation Trust, Manchester, United Kingdom)

    Daratumumab is a human IgG1 kappa antibody that binds with high affinity to a unique epitope on a transmembrane glycoprotein, CD38. It is a targeted immunotherapy directed against plasma cells which express high levels of CD38, such as those produced in patients with multiple myeloma.

    An audit of electrophoresis and immunofixation reports of 6 patients with multiple myeloma being treated with daratumumab was carried out. In all 6 cases an IgG kappa paraprotein band of <2.0 g/L was reported, in addition to the original paraprotein, within 2 months of commencing on daratumumab. We have found that during treatment daratumumab becomes visible as a discrete band on gel electrophoresis and immunofixation, which could be interpreted as a paraprotein. In one patient the IgG kappa of daratumumab migrated to the same electrophoretic region as the patient’s original IgG lambda paraprotein band. Quantification of the paraprotein band therefore also included the IgG kappa of the daratumumab. Two of the six patients had measurable Bence Jones Protein, but the IgG kappa paraprotein was not detected in the urine of either patient.

    The appearance of an apparent second paraprotein band in a patient on treatment for multiple myeloma will cause anxiety and confusion for the patient and may underestimate complete remission rates.

    Thus far, daratumumab has only been used in clinical trials. However, NICE approved the use of daratumumab, via the Cancer Drugs Fund, on 17th January 2018, so clinical use will increase. Therefore it is important to develop strong working relationships between Consultant Haematologists and the laboratory to prevent confusion and anxiety for patients, and to improve the electrophoresis service provided by the laboratory.

  • P007 Modelling CRP minimal retest intervals

    C Cockcroft (Blood Sciences, Bradford Teaching Hospitals NHS Foundation Trust, Bradford, United Kingdom), R Azad (Blood Sciences, Bradford Teaching Hospitals NHS Foundation Trust, Bradford, United Kingdom)

    Our trust is developing an EPR system that can alert and guide clinicians on inappropriate repeat requests. Retesting CRP within 24 hours is not recommended by the Royal Collage of Pathologists guidance on national minimal retesting intervals. The aim of this study is to model how introduction of a minimal retest interval for CRP would benefit demand management and impact patient monitoring.

    The data were gathered from the pathology database for adult CRP requests in 2017. Initial results less than 10 mg/L were compared to the retest result using a CVI of 42.2% (Ricos et al, 2014) and CVA 10% to generate a one-tailed reference change value of 16 mg/L with 95 % significance. Increased CRP levels when initial results was greater than 10 mg/L were identified by a percentage difference greater than 100%.

    The proportion of retest requests received within 24 hours was 16.7% (14851/89040) of the CRP workload for 2017. However, 57% (8424/14851) of these were received between 20 to 24 hours after the initial request and may have reflected variation in daily monitoring routines. Whereas, the proportion of retest requests received within 20 hours was 7.2 % (6427/89040) of the CRP workload. The frequency of a significant increase within 20 hours for a CRP initially tested less than 10 was 8% (128/1610). The frequency of significant increase within 20 hours when the initial result was 10-100 mg/L was 3.5% (182/2657) and 0% when greater than 100 mg/L.

    These findings support the recommendation of a minimal retest interval of 20 hours. However, the ability to retest after 8 hours if the initial result was less than 10 mg/L should be available if clinically indicated. Implementation would potentially decrease the number requests by approximately 5000 per year.

Quality Assurance

  • P008 Focus on effective ascitic and pleural fluid analysis

    K Perkins (Department of Pathology, Royal Preston Hospital, Lancashire Teaching Hospitals), M Myers (Department of Pathology, Royal Preston Hospital, Lancashire Teaching Hospitals) , S Haslam (Department of Pathology, Royal Preston Hospital, Lancashire Teaching Hospitals) NHS Trust, Preston, United Kingdom

    As fluid analysis comprises a small part of the work we do in the laboratory, it is easily ignored as a useful diagnostic service. Additionally, sample collection from patients can be painful and stressful.

    Biochemical testing of ascitic and pleural fluids can help differentiate between a variety of causes. We audited biochemical testing of ascitic and pleural fluids in our laboratory over the period of August – September 2017 to assess appropriate test requesting according to published guidelines. One hundred and thirty six fluid samples were sent to the laboratory during this period: pleural fluids = 83 (61%); ascitic fluids = 39 (29%); unidentified/other (14 samples).

    Differentiating transudate and exudate pleural fluids uses Light’s Criteria. This requires a fluid LDH and protein plus a paired serum LDH and protein. Paired serum requests were made with only 2 of 83 fluid samples received during the audit period. Thirty requests for fluid albumin were made, with no mention of the clinical relevance of this in any current guidelines. Other relevant requested tests included pH (infection), glucose (empyaema), amylase (pancreatitis/oesophageal rupture), lipids (chylothorax/pseudochylothorax) and creatinine (urinothorax).

    Serum Ascites Albumin Gradient is used to differentiate hepato-renal causes of ascites from infection/malignancy and requires a paired serum albumin sample. Of 84 ascitic fluid test requests, 17 were for albumin; 6 of these had a paired serum albumin. LDH and protein can also be used as per Light’s criteria, but the 8 requests for LDH did not match the 9 requests for protein. Other relevant tests requested were amylase, glucose and pH. Fluid U&E panels were requested on some samples. These tests have no clear clinical use.

    Local guidelines for sample collection and analysis are clearly required to provide effective clinical and biochemical assessment of the patient and to avoid painful and stressful resampling.

Toxicology TDM

  • P009 Drugs of abuse screening: effect of lowering positive cut-off thresholds

    RL Griffiths (Department of Clinical Biochemistry, Sandwell & West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), O Roberts (Department of Clinical Biochemistry, Plymouth Hospitals NHS Trust, Plymouth, United Kingdom), L Ford (Department of Clinical Biochemistry, Sandwell & West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), J Berg (Department of Clinical Biochemistry, Sandwell & West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom)

    Introduction: We provide a clinical toxicology screening service for urine drugs of abuse by LC-MS/MS. Our positive cut-off thresholds are based upon the European Workplace Drug Testing Society (EWDTS) guidelines, where possible. However our assay is capable of detecting drugs well below these levels and the EWDTS guidance is not applicable for analytes detected as ‘free’ drug. As part of a recent review of our assay we reduced some of our positive cut-off thresholds. Here we review the effect of these changes.

    Methods: Semi-quantitative results were reviewed to determine the number of additional positive results obtained using these reduced cut-off thresholds, which were as follows: opiates: codeine, dihydrocodeine (300 ng/mL to 50 ng/mL); opioids: EDDP (250 ng/mL to 75 ng/mL), buprenorphine, norbuprenorphine (5 ng/mL to 2 ng/mL); cocaine: benzoylecgonine (150 ng/mL to 100 ng/mL); benzodiazepines: diazepam, nordiazepam, temazepam, oxazepam, lorazepam (5 ng/mL to 2 ng/mL).

    Results: Two thousand three hundred and ninety requests were reviewed. An increase in the positive detection rate was seen for all analytes, with an additional 850 positive results reported. Benzodiazepine reporting rates were significantly increased, with 537 positive results across all analytes. The detection rate of nordiazepam increased the most from 1.9% to 12.0%. Increases in reporting rates for opiates were also observed, with an extra 162 positive codeine and 30 positive dihydrocodeine positive results reported. Furthermore, we detected cocaine use in an additional 32 cases which would have been missed under the previous cut-offs.

    Conclusions: By lowering our cut-off thresholds we have increased our positive reporting rate, thereby reducing the number of false-negative results. The audit demonstrates that laboratories should take into account the clinical requirements of their service and the technology used when deciding on positive cut-off thresholds for drugs of abuse screening. The overall suitability of applying EWDTS cut-offs for toxicology screening in a clinical setting remains questionable.

CLINICAL CASES POSTER PRIZE SHORTLIST

Cardiovascular

  • P010 Macrotroponin T: a cause of consistently elevated troponin T in the investigation of acute coronary syndrome

    A Read (Clinical Blood Sciences, South West London Pathology, St George's Hospital, London, United Kingdom), P Collinson (Chemical Pathology, South West London Pathology, St George's Hospital, London, United Kingdom), C Hunt (Clinical Blood Sciences, South West London Pathology, St George's Hospital, London, United Kingdom)

    Presentation: A 54 year old Asian British male was admitted with a 4 day history of chest pain for 7 days.

    Past medical history: Chronic hepatitis B (e-Antigen negative), non-alcoholic steatohepatitis, hypertension and type-2 diabetes. Previous admissions for chest pain 4 and 2 years prior to this episode.

    Family history: Type 2 diabetes, hypertension and hypercholesterolaemia.

    Clinical course: Troponin T (cTnT) Roche Diagnostics Cobas 8000 (10% CV = 13 ng/L, 99th percentile URL = 14 ng/L) was elevated on admission and noted to remain consistently elevated (1588-1842 ng/L).

    Investigations: Imaging by thoracic CT (aortic imaging, no abnormality detected), coronary angiography (no obstructive epicardial disease) and cardiac MRI (no evidence of injury or wall motion abnormality) did not support acute myocardial injury as a cause for the raised troponin. An analytical interference was suspected and serial dilution and polyethylene glycol (PEG) precipitation of the sample was performed and the sample was analysed for cardiac troponin I (Abbott Diagnostics hs cTnT, 10% CV = 4.7 ng/L, 99th percentile URL = 26.2 ng/L).

    Results: cTnI was <2 ng/L. Serial dilution showed comparable recovery with a known native high troponin sample (y = 0.9856x - 60.759, R² = 0.9986). PEG precipitation showed a large disparity in recovery between original measurements and against known native troponin samples (Day 4 = 1.13% and Day 5 = 1.36% recovery, control samples = 94.46% and 85.09% recovery).

    Conclusion: The results were consistent with macrotroponin and not acute cardiac injury. Macrotroponin T has not been widely reported.

Diabetes

  • P011 Diabetic ketoacidosis presentation in a patient with undocumented cystic fibrosis

    A Jackson-Crawford (Blood Sciences, Royal Gwent Hospital, Newport, United Kingdom)

    Context: Diabetic ketoacidosis (DKA) is an acute presentation of poorly controlled diabetes, and is a common initial presentation in novel type one diabetes (T1DM). A lack of insulin results in hyperglycaemia and causes tissues to revert to fat metabolism for energy. Hydrolysis of fat produces ketones, and can cause a metabolic acidosis. While the most common mechanism of T1DM is autoimmune destruction of pancreatic islet cells, other conditions that can affect insulin production or secretion can also be implicated.

    Case: A 26 year old male presented in DKA serious enough to require intubation and admission to intensive care (ICU). On clinical examination he was confused, disruptive, tachycardic and very underweight (BMI 13). His glucose was 67.6 mmol/L, HbA1c 126 mmol/mol and pH 7.11. He also had acute kidney injury due to dehydration, and a significant pneumonia. During treatment on ICU, the patient was fed via nasogastric tube, and developed refeeding syndrome which complicated his care. Towards the end of his six day ICU stay, it was identified that he had been diagnosed with cystic fibrosis (CF) as a child, but due to little interaction with primary or secondary care during the intervening 20 years this hadn’t been managed. This would have had implications for his acute management during the admission. Pancreatic insufficiency secondary to CF was confirmed in this patient, posing an alternative source of islet cell damage to the typical autoimmune pathway.

    Conclusion: The late disclosure of CF in this patient highlights that mildly affected patients diagnosed before the advent of genetic diagnosis may still present as adults without knowledge that it is a chronic condition. This case also raises awareness of potential atypical mechanisms of T1DM. The patient is now being managed by a CF specialist team.

Endocrinology

  • P012 A case of neonatal iodine-induced hypothyroidism

    C Pritchard (Department of Clinical Biochemistry, University Hospitals Bristol, Bristol, United Kingdom), R Hopkins (Department of Clinical Biochemistry, University Hospitals Bristol, Bristol, United Kingdom)

    This is a case of a male born at 37.5 weeks by semi elective C-section due to early contractions. He had an antenatal diagnosis of congenital heart disease and was noted to have dysmorphic features at birth. A full neonatal screen was taken on day 5 (insufficient) and repeated on day 7 of life, with all disorders reported as not suspected, including congenital hypothyroidism.

    He was admitted to the paediatric intensive care unit on day 12 of life due to apnoeic episodes with desaturations. As part of his congenital heart disease management, a CT scan with IV contrast was performed on day 15, after an initial delay due to mild renal impairment.

    He was discharged at 1 month of age, but readmitted the following day due to poor oral feeding. At this time, thyroid function tests revealed profound hypothyroidism, with TSH >100 mIU/L and fT4 5.3 pmol/L. A referral was made to the paediatric endocrinologists and on review he was diagnosed with iodine-induced hypothyroidism following CT contrast. He was commenced on levothyroxine 37.5 mcg daily.

    The acute response to excess iodine leads to reduced thyroid peroxidase activity and reduced fT4 production known as the Wolff-Chaikoff effect. With prolonged excess iodine exposure there is down regulation of the sodium-iodine cotransporter which reduces iodine transport into the thyroid gland. This downregulation results in an escape from this acute response, resuming normal thyroid hormone synthesis. In neonates this escape mechanism is delayed, most likely due to immaturity of the thyroid gland, putting these patients at increased risk of iodine-induced hypothyroidism. There is no clear guidance in place currently on the use of iodine contrast in the paediatric population at risk of hypothyroidism. Improved guidance is required for high risk patient groups such as the congenital heart disease population to avoid this presentation.

  • P013 Paradoxical increase of 17-hydroxyprogesterone in 17-hydroxylase deficiency

    L Ghataore (Clinical Biochemistry, Viapath Analytics, King's College Hospital, London, United Kingdom), N Taylor (Clinical Biochemistry, Viapath Analytics, King's College Hospital, London, United Kingdom), M Rahman (Diabetes and Endocrinology, Northwick Park Hospital, London, United Kingdom), R Padidela (Paediatric Endocrinology, Central Manchester University Hospitals, Manchester, United Kingdom), R Bhake (Endocrinology, University Hospitals of Leicester, Leicester, United Kingdom), A Burns (Chemical Pathology and Metabolic Diseases, University Hospitals of Leicester, Leicester, United Kingdom)

    Background: 17-Hydroxylase deficiency is a rare cause of congenital adrenal hyperplasia, due to a mutation in the gene CYP17A1. The enzyme combines 17-hydroxylase and 17,20-lyase activity of both adrenal and gonadal steroidogenesis pathways. Here we describe 3 cases where urine steroid metabolite profiling by gas chromatography-mass spectrometry was informative on partial forms of this disorder and we offer an explanation for observed paradoxical increases of 17-hydroxyprogesterone.

    Patients: Patient 1: 16 y F, 1° amenorrhoea, no axillary or pubic hair; Patient 2: 15 y M, micropenis, hypospadias and gynaecomastia; Patient 3: 20 y M, hypospadias and gynaecomastia (deleterious CYP17A1 mutation later confirmed).

    Results: Patients 1 and 2 showed higher deficits of corticosteroid 17-hydroxylation than patient 3: corticosterone/cortisol metabolites were 15, 13 & 1.2 respectively (normal <0.6). Serum cortisol at 09:00 was 124, 70 & 419 nmol/L, with no response to synacthen in any and increased ACTH recorded in patients 1 & 3. Ratios of pregnanetriol, derived from 17-hydroxyprogesterone, over pregnanediol, derived from progesterone, (P3/P2), were: 0.2, 0.9 & 1.1 (normal: F, <5.0, M, <6.2). Serum 17-hydroxyprogesterone was 4.1, ‘marginally high’ and 31.3 nmol/L (normal <18). Rates of 17,20-lyase activity, based on androgen/corticosterone+cortisol metabolites, were low, at 0.003, 0.010 & 0.05 (normal: F, >0.1, M, >0.3).

    Discussion: Patients 1 & 2 have a similar 17-hydroxylation deficit for corticosteroids, but normal P3/P2 ratios. We suggest that the latter steroids are mostly derived from enhanced gonadotrophin-driven gonadal production resulting from diminished sex steroid feedback inhibition. Male patients 2 & 3 show much higher ratios than female patient 1. This might be explained by males having a more metabolically active pathway to androgens than females. Although Patient 3 appears to have a lesser deficit of 17,20-lyase activity than patient 2 and so would be expected to have less testicular hyperstimulation, 17-hydroxyprogesterone concentration is paradoxically much higher.

  • P014 An unexpected cause of secondary adrenal insufficiency

    B Johnson (Clinical Chemistry and Immunology, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), C Webster (Clinical Chemistry and Immunology, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), S George (Toxicology, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), A Bates (Endocrinology and Diabetes, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), L Shepherd (Endocrinology and Diabetes, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), A Lawson (Clinical Chemistry and Immunology and Toxicology, Heart of England NHS Foundation Trust, Birmingham, United Kingdom)

    Background: Secondary adrenal insufficiency is hypoadrenalism which can be caused by a lack of adrenocorticotropic hormone (ACTH), due to a hypothalamic defect, hypopituitarism, defective synthesis/processing of ACTH, or chronic glucocorticoid use. Glucocorticoids are used to treat various medical conditions owing to their anti-inflammatory and immunosuppressive effects; however, chronic use can cause suppression of the hypothalamic-pituitary-adrenal (HPA) axis.

    Case Description: A 60 year-old woman referred to endocrinology following an undetectable cortisol result (<28 nmol/L) in May 2016. She complained of feeling tired and had noticed slight weight loss. Her medical history includes type II diabetes, hypertension, vertigo, migraine, asthma and depression. The clinic notes state there was no obvious abnormal pigmentation. Biochemistry investigations showed a baseline cortisol <28 nmol/L with no response to Synacthen, suppressed ACTH (<5 ng/L) and normal renin (23 mU/L). A diagnosis of secondary adrenal insufficiency was made, most likely attributable to chronic steroid inhaler use. She was started on hydrocortisone and referred for an asthma medication review. She had used a Seretide inhaler but not regularly, and the respiratory physician advised regular steroid inhalers were not needed.

    In July 2017, 7 months since stopping the inhaled steroids, a repeat Synacthen test gave the same failed response; however, the patient mentioned she was taking herbal medication for knee osteoarthritis, and wondered if this could be tested for steroids. These tablets were analysed in the Toxicology department at Heart of England NHS Foundation Trust (HEFT) using a Waters Xevo G2-XS QTof System, and found to contain dexamethasone, ciprofloxacin, paracetamol, diclofenac, ibruprofen and cimetidine. Dexamethasone is a potent glucocorticoid, and likely the cause of her adrenal suppression.

    Conclusion: This case highlights the importance in asking patients about over-the-counter (OTC) and herbal medications when taking their drug history, particularly when the patient has multiple co-morbidities.

  • P015 Florid Cushing's disease in an adolescent male

    J Flatt (Department of Medical Biochemistry and Immunology, University Hospital of Wales, Cardiff, United Kingdom), S Tennant (Department of Medical Biochemistry and Immunology, University Hospital of Wales, Cardiff, United Kingdom), A Lansdown (Endocrinology, University Hospital of Wales, Cardiff, United Kingdom), C Evans (Department of Medical Biochemistry and Immunology, University Hospital of Wales, Cardiff, United Kingdom)

    A 17 year old male was investigated by endocrinology after initial investigations revealed a random cortisol of 798 nmol/L (1:30 pm) with suppression of LH, FSH and testosterone. The patient was hypokalaemic (3.2 mmol/L) and had normal TFTs, prolactin and HbA1c (30 mmol/mol). On examination he was noted to be markedly cushingoid. He had purple striae on his upper arms and lower abdomen, which he had noticed about a year previously, and had gained 2 stone in weight over the last 6 months. He reported anxiety and trouble sleeping. He was hypertensive (176/102) and had ankle oedema that limited his mobility. He had experienced ongoing back pain, and reported a loss in height. He suffered from proximal muscle weakness, especially when standing, and had difficulty climbing stairs.

    Cushing’s syndrome was suspected and subsequent investigations showed a 24 hour urine free cortisol of 4,411 nmol (<146). This was followed by a low dose dexamethasone test; 48 hour cortisol was 696 nmol/L (<50). ACTH was 147 ng/L (7-63), which indicated ACTH-dependent Cushing’s. Initial imaging of the pituitary, adrenals, chest, thorax and abdomen did not reveal any suspicious lesions. Inferior petrosal sinus sampling was consistent with a pituitary source. Subsequent high resolution MRI identified a 5 mm right-sided microadenoma in the pituitary. The patient was medically managed with anti-hypertensives and a block and replace regime of metyrapone and dexamethasone. These alleviated his symptoms until adenomectomy, which was performed trans-sphenoidally.

    His 5 day post-operative 9 am cortisol was 70 nmol/L, indicating successful resolution of the Cushing’s disease. He is being treated with glucocorticoid replacement until normal hypothalamic-pituitary-adrenal function is re-established. This was an interesting case as it featured unusually severe Cushing’s in a young male.

Miscellaneous

  • P016 Assay interference could lead to potentially life-threatening delay in diagnosis in ethylene glycol poisoning

    L Evans (Department of Clinical Biochemistry, Liverpool Clinical Laboratories, Liverpool, United Kingdom), A Davison (Department of Clinical Biochemistry, Liverpool Clinical Laboratories, Liverpool, United Kingdom)

    Background: The differential diagnosis of a raised anion gap metabolic acidosis can be complex, ranging from increased acid production associated with keto- and lactic acidosis, ingestion of alcohols and poisons and decreased acid excretion as a result of renal and liver impairment. Clinical history together with examination and pathology results aid in summation of likely clinical diagnosis in patients presenting with this and direct the appropriate course of action. In the absence of reliable results, incorrect conclusions can be drawn which ultimately leads to mismanagement.

    Case Description: A 71 year old male presented to A+E with raised anion gap metabolic acidosis of unknown cause (pH=7.15, [H+]=71 nmol/L, [HCO3-]=10 mmol/L, osmolality=390 mOsm/kg, osmolal gap=83.7, anion gap=38). A decreased GCS (GCS 5; E2M2V1) rendered it impossible to obtain a clinical history. Arterial blood gas and peripheral blood samples were collected two minutes apart and lactate was measured simultaneously on each using a Roche b123 POCT blood gas analyser and Roche Cobas 501 analyser respectively. Results obtained using the two different platforms were found to be significantly discordant (ABG= 2.2 mmol/L vs Roche c8000= 40.9 mmol/L) causing confusion amongst clinical staff. Investigation determined that the patient’s condition resulted from intentional ethylene glycol ingestion (788 mg/L on admission), not lactic acidosis and he was treated with formepazole and haemodialysis. It is believed that the discrepancy in lactate measurements was due to cross-reactivity of glycolate with the oxidase enzyme utilised in the automated enzymatic assay.

    Conclusion: Timely treatment of ethylene glycol poisoning is important to prevent/limit neurological, cardiovascular and renal complications resulting from the accumulation of toxic metabolites glycolic and oxalic acid. Inability to collate an adequate patient history together with falsely elevated lactate results as observed in this case have the potential to confuse interpretation of the clinical picture resulting in delayed diagnosis or even misdiagnosis and inappropriate patient management.

Oncology

  • P017 Chloral hydrate and chlorphenamine-induced interference in an HPLC method for catecholamines

    C Pitt (Department of Clinical Biochemistry, NHS Ayrshire & Arran, Kilmarnock, United Kingdom)

    The aim of this work was to report possible interference in an HPLC method for catecholamines and metabolites (cats, mets and VMA). These were measured using an HPLC kit with electrochemical detection. The column was a Spherisorb ODS2 150 x 4.6 mm i.d. with 5 µm particles and the mobile phase was diammonium hydrogen orthophosphate with heptane sulphonic acid, EDTA and methanol. Flow-rate was 2 mL/min.

    The patient, a 2 year old girl, has Leigh’s syndrome, a rare genetic neurometabolic condition with a possible link to phaeochromocytoma. She had had a three day exposure to chlorphenamine eight days prior to sample collection and had been commenced on chloral hydrate three days prior to the collection. The chlorphenamine had been stopped but the chloral hydrate continued throughout the collection period. Although raised, the chromatography results showed an unusual pattern for phaeochromacytoma/neuroblastoma.

    Typical phaeochromocytoma results include elevated noradrenaline and adrenaline with elevated normetadrenaline and metadrenaline. In the current case, four urine samples were collected within a fortnight. Results on these varied. Initially, adrenaline, noradrenaline, dopamine, free metadrenaline and VMA were elevated, in-keeping with a neuroblastoma. A similar pattern was seen on a sample 9 days later. A day 10 sample, however, showed dopamine and 3MT were decreasing and a final sample on day 12 showed only elevated VMA and HVA with all other results within reference ranges. Chloral hydrate (tri-chloro acetaldehyde) is an anxiolytic and is converted to trichloroethanol which enhances GABA-mediated inhibition with a risk of gastric irritation. Chlorphenamine is a weakly sedative anti-histamine. It is an alkyl amine with two phenyl rings attached to a central cyano-containing carbon. Whether these structures cause on-column interference or physiological stimulation of catecholamine release remains to be clarified.

Paediatrics and IEM

  • P018 A case of 3-hydroxy-3-methylglutaryl coenzyme A deficiency: the role of the metabolic biochemist in expediting specialist testing

    SJ Bennett (Department of Chemical Pathology, Great Ormond Street Hospital NHS Foundation Trust, London, United Kingdom), R Wigley (Department of Chemical Pathology, Great Ormond Street Hospital NHS Foundation Trust, London, United Kingdom), H Prunty (Department of Chemical Pathology, Great Ormond Street Hospital NHS Foundation Trust, London, United Kingdom), M Cleary (Metabolic Medicine Department, Great Ormond Street Hospital NHS Foundation Trust, London, United Kingdom), H Aitkenhead (Department of Chemical Pathology, Great Ormond Street Hospital NHS Foundation Trust, London, United Kingdom)

    3-Hydroxy-3-methylglutaryl coenzyme A (3-HMG CoA) lyase deficiency is a rare inborn error of leucine metabolism which leads to impaired ketogenesis. It usually presents within the first year of life and is characterised by metabolic decompensation episodes which are fatal in approximately 20% of cases and result in permanent neurological injury in the majority of those who survive. Early diagnosis and initiation of appropriate treatment is critical for a good outcome.

    We describe a pre-term baby presenting on day two with mild hypoglycaemia, hyperammonaemia, metabolic acidosis and hypothermia. The hypoglycaemia and hypothermia resolved following 90 mL/kg/d of 10% dextrose, IV fluids and antibiotics; however the metabolic acidosis persisted and was considered secondary to renal tubular acidosis. The baby was clinically well and discharged on day nine with improved ammonia levels and blood gases. Bloodspot acylcarnitines requested as part of a hypoglycaemia screen showed moderately increased 3-hydroxyisovalerylcarnitine with mildly increased adipyl/methylglutarylcarnitine. Based on these abnormalities the metabolic biochemist alerted the clinical team and requested urinary organic acid analysis which showed gross excretion of 3-hydroxy-3-methylglutaric, 3-methylglutaric, 3-methylglutaconic and 3-hydroxyisovaleric acids - consistent with 3-HMG CoA lyase deficiency. The baby was urgently reviewed and appropriate clinical management implemented.

    This case demonstrates the crucial role of the metabolic biochemist in expediting appropriate specialist testing to aid prompt diagnosis and patient management. It also highlights (1) that a milder neonatal presentation of hypoglycaemia and metabolic acidosis, with the former resolving, may indicate a rarer underlying metabolic cause and (2) the importance of measuring bloodspot/plasma acylcarnitines and urinary organic acids in parallel, rather than in isolation.

Proteins Enzymes

  • P019 A case of unusual serum protein electrophoresis and CSF oligoclonal band results following bone marrow transplant

    K Birch (The Neuroscience Laboratories, The Walton Centre NHS Foundation Trust, Liverpool, United Kingdom), L Bailey (Liverpool Clinical Laboratories, The Royal Liverpool and Broadgreen University Hospital NHS Trust & Aintree University Hospital NHS Foundation Trust, Liverpool, United Kingdom), J Holt (Department of Neurology, The Walton Centre NHS Foundation Trust, Liverpool, United Kingdom), G Keir (The Neuroscience Laboratories, The Walton Centre NHS Foundation Trust, Liverpool, United Kingdom)

    A 30 year old gentleman was diagnosed with chronic myeloid leukaemia in April 2016. He underwent myeloablative chemotherapy culminating in allogenic stem cell therapy in October 2016. He had a number of post-transplant complications and developed graft versus host disease affecting his skin and bowel.

    In October 2017 he was referred to Neurology for a number of new symptoms. He reported having tingly feet that progressed over the following days with weakness in an ascending pattern so that his hands were then also affected and he was unable to walk unassisted.

    During the investigation of these symptoms, the laboratory received requests for serum protein electrophoresis and CSF oligoclonal bands. The serum protein electrophoresis results showed 3 clear monoclonal bands demonstrated to be a mixture of IgG lambda and IgG kappa isotypes by immunofixation. Iso-electric focussing revealed bands corresponding to the 3 dominant monoclonal products all of which were mirrored in CSF.

    His history was felt to be in keeping with a diagnosis of chronic inflammatory demyelinating polyradiculoneuropathy (CIDP). This was supported by his neurophysiology results, his CSF results which showed a raised protein of 0.84 g/L and raised white cell count of 10 cells per mm3, the absence of any glycolipid antibodies and his excellent response to a treatment plan including oral prednisolone and intravenous immunoglobulin.

    It was thought that CIDP had developed in the context of the reboot of his immune system by the bone marrow transplant. It was also felt that the unusual pattern of multiple monoclonal immunoglobulin bands was more likely to represent normal regeneration of the bone marrow following transplant rather than a malignant process. Although this phenomenon has been described previously, it represents an important learning point for clinical scientists reporting serum protein electrophoresis and CSF oligoclonal band results.

THURSDAY

Diabetes

  • P020 Fatty acid profile of Mexican and Nigerian type 2 diabetes mellitus subjects

    EK Oghagbon (Department of Chemical Pathology, Benue State University Teaching Hospital, Makurdi, Nigeria), LS Harbige (Lipidomics and Nutrition Research Centre, School of Human Sciences, London Metropolitan University, London, United Kingdom), B Chowdhry (Department of Sports and Life Science, Faculty of Engineering & Science, University of Greenwich at Medway, Medway, Kent, United Kingdom), E Sedlak (Lipidomics and Nutrition Research Centre, School of Human Sciences, London Metropolitan University, London, United Kingdom), R Valdes-Ramos (Faculty of Medicine, Universidad Autónoma del Estado de México, Paseo Tollocan esq. Jesús Carranza S/N, Col. Moderna de la Cruz, Toluca, Toluca, Mexico), K Ghebremeskel (Lipidomics and Nutrition Research Centre, School of Human Sciences, London Metropolitan University, London, United Kingdom)

    Objective: To investigate fatty acid status of type 2 diabetes mellitus (DM2) subjects from two populations with contrasting diet and genetics, but have rapidly increasing prevalence in DM2.

    Methods: Fasting blood samples were obtained from Nigerian (n=15) and Mexican (n=15) subjects. Plasma total lipids were extracted, fractionated and fatty acid composition of choline phosphoglycerides assayed.

    Results: The Nigerian subjects compared with their Mexican counterparts had higher levels of heptadecanoic (0.46±0.14% vs. 0.34%±0.05, p=0.035), stearic (15.39±1.90% vs. 13.52±1.24%, p=0.004), arachidic (0.13±0.06 % vs. 0.05±0.05%), total saturates (48.37±4.10% vs. 45.28±3.62%, p=0.037), arachidonic (9.22±2.17% vs. 5.64±1.21%, p=<0.001), eicosapentaenoic (0.67±0.39% vs. 0.28±0.15%, p=0.002), n-3 docosapentaenoic (0.53±0.18% vs. 0.34± 0.13%, p=0.002), docosahexaenoic (3.62±1.23% vs. 1.49±0.55%, p=<0.001) and total n-3 PUFA (5.00±1.49% vs. 2.50±0.70%, p=<0.001). Conversely, the Nigerians had lower pentadecanoic (0.10±0.04% vs. 0.15±0.05%), linoleic (16.73±5.13 vs. 27.08±2.28%, p<0.001), total n-6 PUFA (29.11±4.46% vs. 36.01±1.77%, p=<0.001) and alpha-linolenic (0.18±0.15% vs. 0.40±0.11%, p=<0.001).

    Conclusion: This preliminary investigation suggests that fatty acid profiles of Nigerian and Mexican DM2 subjects may indicate background differences in diet intake or genetics.

  • P022 Defining the functional sensitivity of the Abbott ARCHITECT C-peptide assay for the investigation of endogenous insulin production in autoimmune diabetics

    M McNaughton (Department of Clinical Biochemistry, The Western General Hospital, Edinburgh, United Kingdom), C Clarke (Department of Clinical Biochemistry, The Western General Hospital, Edinburgh, United Kingdom)

    Background: Serum C-peptide levels can be used as a marker of endogenous insulin production and residual β-cell function in patients with autoimmune type 1 diabetes. There is a growing body of evidence which suggests that many patients with type 1 diabetes retain residual β-cell function, and that even very low levels of C-peptide secretion may represent significant preservation of β-cell function. Thus, high sensitivity C-peptide assays can be used to identify patients who may benefit from interventions to preserve β-cell function. The limit of quantification (LOQ) or functional sensitivity of an assay is the lowest concentration of analyte for which results are considered reliable and are therefore reportable. It is recommended that laboratories establish that commercial assays meet the manufacturer’s performance claims. Therefore, the aim of this study was to determine the functional sensitivity of the Abbott ARCHITECT C-peptide assay (Abbott state an LOQ of 0.08 ng/mL which is equal to 26.7 pmol/L).

    Method: Patient samples spanning the lower analytical range of the C-peptide assay were analysed on the Abbott Architect i2000SR to establish the functional sensitivity of the assay. This was determined by assessing the precision profile of the assay, defined as the precision obtained over an analytical range of the assay. The functional sensitivity is defined as the concentration where the CV is less than 20%.

    Results: The Abbott ARCHITECT C-peptide assay was able to measure C-peptide levels of 3.5 pmol/L with a CV of 15%.

    Conclusion: The functional sensitivity of the Abbott C-peptide assay was established as 3.5 pmol/L. This assay is therefore sensitive enough to investigate the extent of endogenous insulin secretion in patients with definite autoimmune diabetes.

  • P023 Unexplained hypoglycaemic episodes in a patient with type 1 diabetes mellitus: the DiaPort pump

    A Nance (Department of Clinical Chemistry and Immunology, Heartlands Hospital, Birmingham, United Kingdom), J Duffy (Department of Clinical Chemistry and Immunology, Heartlands Hospital, Birmingham, United Kingdom), C Webster (Department of Clinical Chemistry and Immunology, Heartlands Hospital, Birmingham, United Kingdom)

    A 24-year-old woman had been diagnosed with type 1 diabetes mellitus in childhood, which was well-controlled throughout her adolescence (HbA1c <50 mmol/mol). In 2014, she began to suffer hypoglycaemic episodes with impaired awareness, as well as occurrences of diabetic ketoacidosis (DKA). Subsequent management has revealed this to be a significantly unusual presentation. Her hypoglycaemic unawareness has had a profound effect on her everyday life, with the loss of her job, her driving licence, and her ability to leave the house on her own, along with over a dozen inpatient stays since 2014.

    Regaining good glycaemic control is a priority for her healthcare team, and she has trialled three different kinds of insulin pumps: Medtronic, Omnipod, and most recently DiaPort. Medtronic and Omnipod both resulted in admissions for DKA, and the Medtronic pump was investigated for dysfunction. During her inpatient stays, the patient was given insulin both intravenously and subcutaneously. IV insulin has both provided good management and been associated with hypoglycaemia; a change to subcutaneous insulin was associated with hyperglycaemia. In May 2017, samples were taken after a period of 48 hours with no insulin administered. Results were: glucose 2.4 mmol/L, insulin 2720 pmol/L, C-peptide <94 pmol/L, glucagon 12 pmol/L, cortisol 197 mmol/L. This indicated hypoglycaemia due to exogenous insulin; the poor lipolytic and ketotic response supports the hyperinsulinaemia.

    Referral to a national specialist provided guidance regarding potential further investigations, including checking antibodies to insulin and the insulin receptor, and considering sub-cutaneous storage of insulin and subsequent uncontrolled release (never before seen by the specialist), and surreptitious self-administration of insulin (statistically much more likely). To address the possible subcutaneous storage issue, the patient has since been fitted with a DiaPort pump, which administers insulin directly into the peritoneal cavity, enabling it to reach the liver more rapidly.

  • P024 Frequency of testing for glycated hemoglobin affects the probability of achieving target levels in patients with sub-optimally controlled diabetes mellitus

    C Duff (Department of Clinical Biochemistry, University Hospitals of North Midlands NHS Trust, Stoke on Trent, United Kingdom), I Solis-Trapala (Institute for Applied Clinical Sciences, Keele University, Stoke on Trent, United Kingdom), O Driskell (Department of Clinical Biochemistry, University Hospitals of North Midlands NHS Trust, Stoke on Trent, United Kingdom), D Holland (Keele University Benchmarking Service, Keele University, Stoke on Trent, United Kingdom), H Wright (Department of Clinical Biochemistry, University Hospitals of North Midlands NHS Trust, Stoke on Trent, United Kingdom), J Waldron (Department of Clinical Biochemistry, Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), C Ford (Department of Clinical Biochemistry, Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), J Scargill (Department of Clinical Biochemistry, Salford Royal NHS Foundation Trust, Salford, United Kingdom), M Tran (Department of Clinical Biochemistry, University Hospitals of North Midlands NHS Trust, Stoke on Trent, United Kingdom), F Hanna (Department of Diabetes and Endocrinology, University Hospitals of North Midlands NHS Trust, Stoke on Trent, United Kingdom), J Pemberton (Diabetes UK North Staffordshire Branch, Diabetes UK, Stoke on Trent, United Kingdom), A Heald (Department of Endocrinology, Salford Royal NHS Foundation Trust, Salford, United Kingdom), A Fryer (Department of Clinical Biochemistry, University Hospitals of North Midlands NHS Trust, Stoke on Trent, United Kingdom)

    Background: The measurement of glycated haemoglobin (HbA1c) is a core component of monitoring disease control in patients with diabetes mellitus and should be measured in line with time intervals recommended by NICE. We have previously shown that >50% of monitoring tests for HbA1c are outside recommended time intervals and that testing frequency is linked to diabetes control.

    Aims: To examine the impact of number of tests/year on the probability of achieving commonly utilised HbA1c targets and on HbA1c changes over time.

    Methods: Data from 20,690 patients aged >16 years with diabetes with a baseline HbA1c test >53 mmol/mol (7%) were extracted from Clinical Biochemistry laboratory records at three UK hospitals during 2007-2011. We examined the impact of number of HbA1c tests/year on: (i) probability of achieving targets of 53 mmol/mol and 48 mmol/mol in a year using a multi-state model and (ii) changes in mean HbA1c over time using a linear mixed-effects model. All models were adjusted for sex, age and centre.

    Results: The overall probabilities of achieving the 53 mmol/mol (7%) and 48 mmol/mol (6.5%) targets within 1 year were 0.20 (95% confidence interval: 0.19-0.21) and 0.10 (0.09-0.10), respectively. Compared with the recommended four tests/year, having only one test or >4 tests/year were associated with lower likelihoods of achieving either target. Undergoing 2-3 tests/year gave similar likelihoods to four tests/year. Mean HbA1c levels were higher in patients who had 1 test/year compared to those with four tests/year (difference: 2.64 mmol/mol, p<0.001).

    Conclusions: Our data support the recommendations on HbA1c monitoring frequency (4 tests/year), suggesting that patients who receive this level of monitoring have better-controlled diabetes. However, given the poor conformity to guidance on testing interval, and the low overall probabilities of patients achieving targets, the importance HbA1c testing frequency on diabetes management may not be being adequately recognised.

  • P025 Questions asked of laboratories about pathways for HbA1c measurement: now required for both diagnosis and management of diabetes

    S Ghosh (Diabetes Centre, University Hospitals Birmingham NHS Foundation Trust, Birmingham, United Kingdom), R Susarla (Diabetes Centre, University Hospitals Birmingham NHS Foundation Trust, Birmingham, United Kingdom), R Round (Diabetes Centre, University Hospitals Birmingham NHS Foundation Trust, Birmingham, United Kingdom), C Chaloner (Department of Clinical Biochemistry, Manchester University NHS Foundation Trust, Manchester, United Kingdom), C Crotty (Haematology Laboratory, University Hospital Waterford, Waterford, Ireland), K Gordon (Department of Clinical Biochemistry, Hampshire Hospitals NHS Foundation Trust, Hampshire, United Kingdom), L Hikin (Leicester Royal Infirmary, University Hospitals Leicester, Leicester, United Kingdom), H Johnstone (Chemical Pathology, Epsom & St Helier University Hospitals NHS Trust, Surrey, United Kingdom), K Stuart (Chemical Pathology, West Hertfordshire Hospitals NHS Trust, Hertfordshire, United Kingdom), S Thomas (Chemical Pathology, Gloucestershire Hospitals NHS Foundation Trust, Cheltenham, United Kingdom), R Webster (Clinical Biochemistry, University Hospitals Birmingham NHS Foundation Trust, Birmingham, United Kingdom), I Stratton (Gloucestershire Retinal Research Group, Gloucestershire Hospitals NHS Foundation Trust, Cheltenham, United Kingdom), J Webber (Diabetes Centre, University Hospitals Birmingham NHS Foundation Trust, Birmingham, United Kingdom), G Roberts (College of Medicine, Swansea University, Swansea, United Kingdom), S Manley (Diabetes Centre, University Hospitals Birmingham NHS Foundation Trust, Birmingham, United Kingdom)

    Background: The HbA1c workload has increased markedly since it was introduced for diagnosis of diabetes in patients in the community. As hyperglycaemia linked to undiagnosed diabetes occurs in hospital, a national diabetes panel is preparing guidance on its use in hospital admissions.

    Aim: To review current practice for HbA1c in NHS trusts and other hospitals using a questionnaire on relevant aspects of HbA1c measurement.

    Methods: Questions were compiled by an experienced, diabetes translational research team and circulated by email to pathology colleagues in 16 laboratories for the pilot study.

    Results: Completed questionnaires were received from 8 laboratories: 3 from university teaching hospitals and 4 from district general hospital trusts in England, plus a regional hospital in Ireland.

    High performance liquid chromatography was used in 7 laboratories (4:3 ion exchange/boronate affinity) and capillary electrophoresis in 1 laboratory. The workload varied from 250-450/analyser per day on either one, two or three analysers. Point of care testing was available in 4 trusts in addition to laboratory testing with devices used for adults rather than paediatric patients in 1 trust only.

    HbA1c was measured during normal working hours (8.30am/9am to 5pm/8pm Monday to Friday). A Saturday service was available in 2 trusts and in 1 over the weekend when possible. No trust had a system in place for urgent requests. Turnaround time was generally 24 hours on weekdays but if received over the weekend could be 72 hours. Only 1 laboratory highlighted values below the reference range or requiring urgent referral to the requester; 4 could detect variant haemoglobin on measurement of HbA1c and 1 did not report HbA1c if haemoglobinopathies present.

    Conclusions: The establishment of optimal pathways for the measurement of HbA1c is required for its use as a diagnostic and monitoring tool in both hospitals and the community.

Methods Instrumentation

  • P026 Fourier transform infrared spectroscopy: moving from chemical spot test for urinary calculi analysis

    P Chincholkar (Department of Clinical Pathology, Singapore General Hospital, Singapore), NH Hassan (Department of Clinical Pathology, Singapore General Hospital, Singapore), MM Yap (Department of Clinical Pathology, Singapore General Hospital, Singapore), WY Ng (Department of Clinical Pathology, Singapore General Hospital, Singapore)

    Aims: Fourier transform infrared spectroscopy (FT-IR) is able to identify compounds in urinary calculi as formed (e.g. calcium oxalates, urates, phosphates) and also provides a semiquantitative measure of the components. These are advantages over the commonly used chemical spot method for calculi component testing. Our study aims to evaluate the FT-IR method for analysis of urinary calculi.

    Methods: A Bruker Alpha FT-IR analyser with ATR (attenuated total reflection) mode was used for comparison of results by Biolabo chemical spot analysis. Specimen is placed directly for analysis with little further processing. Using known compounds, accuracy and precision of FT-IR analysis were studied by daily analysis 3 times over a 5-day period and through identification with the library BLG-2 of over 5000 typed and referenced calculi compounds. Comparison with clinical specimens (n=121) tested with the Biolabo method was performed. Results were scored for agreement between the two methods.

    Results: FT-IR analysis of known compounds scored 100% matched identification (hit quality >900) every time showing good consistency. Hence, identification and quantitation was based on a hit quality score of 800 as definitive presence. Comparison with clinical samples showed that FT-IR gave corroborative results (vs chemical spot) – e.g. oxalates 79.3% vs 71.9%, phosphates 48.8% vs 66.9%, urates 34.7% vs 46.3% and carbonates 45.5% vs 5%. Carbonates by the Biolabo test are much less sensitive by observation of an effervescence reaction. Agreement between methods was 76.0% (oxalates), 75.0% (phosphates) and 80.2% (urates). Result reporting can include component quantity (in percent) thereby providing composition details which are not available with chemical spot analysis.

    Conclusion: The FT-IR method agreed well with chemical spot method. Giving composition details would provide a better picture to the clinician. The FT-IR method is quick, requires little sample preparation and with a large reference library is shown to be accurate.

  • P027 How low can your clinical research method go for the analysis of serum estrogens?

    R Wardle (Scientific Operations, Waters Corporation, Wilmslow, United Kingdom), L Calton (Scientific Operations, Waters Corporation, Wilmslow, United Kingdom)

    Background: The two major biologically active estrogens in non-pregnant humans are 17β-estradiol (E2) and estrone (E1). Here we address three challenges faced when trying to quantify E2 and E1 levels as low as 3.7 pmol/L by LC-MS/MS for clinical research and present the optimised method performance data.

    Methods: Different sample extraction techniques were investigated including LLE and SPE using Oasis µElution. The optimal extraction conditions were achieved when performing LLE with a mixture of hexane and ethyl acetate. Chromatographic separation of E2 from E1 was achieved using a Cortecs Phenyl, 2.1 x 50 mm column with a mobile phase gradient consisting of water with ammonium fluoride (MP A) and methanol with acetonitrile (MP B). The addition of ammonium fluoride assisted in the ionization of estrogens and provided optimal analytical sensitivity using an ACQUITY UPLC I-Class FTN with a Xevo TQ-XS mass spectrometer.

    Results: A linear response across the measuring range of 3.7-7350 pmol/L was achieved for E2 and E1. Within-run and total precision were <10% at concentration levels between 7-3000 pmol/L and <16% at a concentration of 4 pmol/L. The lowest concentration at which precision of <20% CV and S:N (ptp) of >10:1 was achieved was 3.7 pmol/L. Comparison with an independent LC-MS/MS method for E2 showed good agreement with minimal bias.

    Conclusion: Many challenges were overcome in our quest to quantify estrogens at low concentration levels (3.7 pmol/L) for clinical research without the need for derivitisation.

    For Research Use Only, Not for use in diagnostic procedures.

  • P028 Evaluation of a quantitative β-trace protein method for detection of CSF rhinorrhoea

    A Williams (Clinical Pathology, Nottingham University Hospitals, Nottingham, United Kingdom), S Barber (Clinical Pathology, Nottingham University Hospitals, Nottingham, United Kingdom), H Divyateja (Clinical Pathology, Nottingham University Hospitals, Nottingham, United Kingdom)

    Background: CSF rhinorrhoea is a breakdown of the barrier separating the spinal fluid space from the nose, leading to leakage of CSF. This is most commonly caused by trauma or complications from cranial surgery. Rapid detection of CSF rhinorrhoea is vital, as this can lead to bacterial infection and meningitis. The most commonly used biomarker for identification of a CSF leak is β-transferrin, which is detected using immunofixation electrophoresis. An alternative biomarker is β-trace protein (β-TP), which is the second most abundant protein in the CSF. Although also present in serum at low concentrations, the high CSF:serum ratio (33:1) allows β-TP to be used as a specific marker for CSF leaks. A study by Risch et al. (2011) showed that a cut-off of 1.11 mg/L β-TP identified patients with CSF leakage with a sensitivity of 93% and a specificity of 100%.

    Method: Levels of β-TP were measured in 45 patient samples (nasal secretions and other fluids) using a nephelometric method on the Siemens BNII analyser; results were compared to those obtained from qualitative β-transferrin analysis using immunofixation electrophoresis. Results were also correlated against clinical findings.

    Results: The patient comparison showed good agreement between the two methods, with the β-TP method able to give a definitive result in equivocal samples from electrophoresis. Analysis of historical EQA material showed a slight positive bias compared to the method mean. Repeatability was good with a CV of 1.4% at 2.98 mg/L and 2.4% at 0.88 mg/L; intermediate precision was 2.2% and 0.5% respectively.

    Conclusion: This method is to be introduced to the laboratory imminently. The automated β-TP method will enable faster detection of CSF rhinorrhoea, as well as allowing the assay to be performed out of hours if necessary. This will lead to faster clinical intervention, and a better overall patient experience.

  • P029 Increasing the number of biochemistry results that can be reported on jaundiced patients

    S King (Department of Clinical Biochemistry, Mid Cheshire Hospitals NHS Foundation Trust, Crewe, United Kingdom), E Rumsby (Department of Clinical Biochemistry, Alder Hey Children's Hospital, Liverpool, United Kingdom), S Robinson (Department of Clinical Biochemistry, Mid Cheshire Hospitals NHS Foundation Trust, Crewe, United Kingdom)

    Background: Beckman Coulter recommendations were limiting the number of results that could be reported in jaundiced patients. This presented a clinical challenge, particularly in neonates.

    Aim: To investigate the effect of bilirubin interference on affected assays.

    Method: Surplus paediatric and adult blood was collected and pooled according to icteric index (0 to +5). Dilutional linearity experiments were performed. Duplicate measurements of urea were performed on the neat and diluted pools on the Beckman AU5800. Duplicate measurements of TSH and cortisol were performed on the Beckman DxI.

    For free hormones, the known +5 and 0 patient pools were mixed in different proportions. This enabled the controlled production of pools up to and including the upper limit for icterus +3, without significantly altering matrices. Analytical recovery was assessed by calculating expected values and comparing to those observed from duplicate analysis on the DxI.

    Results: Dilutional linearity was achieved for urea, cortisol and TSH up to and including icterus +4. r2 values were all >/= 0.99 (2 dp) and followed y=mx+c. Calculated recoveries for free T4 and free T3 were within the limits of within-batch assay imprecision up to and including icterus +3.

    Conclusions: Icterus doesn’t affect the performance of the TSH, urea and cortisol assays up to and including icterus +4. It wasn’t possible to obtain patient material at the upper limit of the +5 index, so it is not possible to report these tests in extremely jaundiced patients at +5 icteric indices.

    Free T4 and T3 results can be reliably reported up to and including icterus +3, but it was not possible to assess icterus up to the top of +4 due to insufficient patient samples, confounded by the requirement to mix pools.

    Interference rules in the analyser software were adjusted accordingly with a positive impact on our neonatal service.

  • P030 Validation of a liquid chromatography mass spectrometry method for 1,25-dihydroxy vitamin D3

    F Ivison (Clinical Biochemistry, Manchester University NHS Foundation Trust, Manchester, United Kingdom)

    DHVD, 1,25-dihydroxy vitamin D3, is the active metabolite of vitamin D, required to maintain blood calcium levels in response to increases in parathyroid hormone. Measurement has proved challenging as it circulates in picomolar concentrations and must be differentiated from other dihydroxyvitamin D species.

    Clinically it is essential to be able to determine the cause of hypercalcaemia, which may be due to DHVD excess, particularly in potential extra-renal production and vitamin D dependent rickets and there is a role for monitoring DHVD treatment in chronic kidney disease.

    The LCMS assay which has been developed uses immunoextraction of 0.5 mL serum or plasma followed by Amplifex derivatisation of the dried eluent, with the analysis carried out using the Sciex 6500+ instrument taking a run time of 11 minutes.

    The limit of quantitation was determined at 15 pmol/L and is linear up to at least 600 pmol/L. Repeatability ranged from 6.1% at 23 pmol/L to 2.5% at 172 pmol/L and intermediate imprecision was 16.4% at 25 pmol/L to 8.1% at 174 pmol/L. Uncertainty of measurement was determined as 24±3.8 pmol/L, 104±11.8 pmol/L and 174±12.4 pmol/L. The method is not affected by icterus, haemolysis or lipaemia.

    A negative bias was observed across the concentration range found in 78 patient samples in comparison to a commercial radioimmunoassay (mean -45%). This was not unexpected and is likely due to better specificity of the LCMS assay and the lack of a commutable standard reference calibrator.

    Good performance has been achieved with the current vitamin D external quality assessment scheme and using historical distributions, demonstrating a negative bias compared to the all lab trimmed mean (average -15%) and the specific method group (average -8.6%).

  • P031 Effect of haemolysis on measurement of anti-tissue transglutaminase

    S Knowles (Blood Sciences, Pathology Department, Derby Teaching Hospitals NHS Foundation Trust, Derby, United Kingdom), L Hawke (Blood Sciences, Pathology Department, Derby Teaching Hospitals NHS Foundation Trust, Derby, United Kingdom)

    IgA antibodies against tissue transglutaminase (tTG) are highly specific markers of coeliac disease and are the first line serological test of choice. In IgA deficiency however, an IgG antibody assay is required and automated IgG tTG assays are now routinely available.

    To streamline our serological testing for coeliac disease we commenced verification of the IgG-tTG assay on the Phadia Immunocap 250 (Thermofisher). The manufacturer states haemolysed samples are not suitable but provides no cut-offs for sample acceptability. The aim of this study was to determine the effect of haemolysis on IgG tTG and derive a laboratory protocol for interpretation of haemolysed samples.

    Three serum pools with a range of IgG-tTG levels were prepared. These were aliquoted and varying concentrations of lysed red cells were added to give a range of IgG tTG pools with different degrees of haemolysis. IgG-tTG was then re-measured on the Phadia 250 alongside haemolysis index on a Roche cobas 8000 analyser.

    Haemolysis was found to have a significant effect on the assay, with a 50% decrease in IgG-tTG level (initially measured at 8.8 U/mL) at Roche haemolysis index 112 (approximate haemoglobin concentration 112 mg/dL). This decrease would result in a change of interpretation from equivocal to negative (reference range <7 U/mL, positive >10 U/mL). Even a low haemolysis index of 34 resulted in a 14% decrease IgG tTG (initially 8.8 U/mL).

    Following on from this, the effect of haemolysis on the IgA-tTG assay (Phadia 250) was also investigated. The effect on this assay was found to be much less significant with only a 15% decrease in IgA-tTG when the haemolysis index was 248.

    In our laboratory, haemolysis index is performed on all IgG tTG samples and appropriate comments are added when haemolysis index >30 to highlight the effect to users.

  • P032 Verification of the automated Inova-Werfen QUANTA Flash calprotectin chemiluminescent immunoassay

    J Bailey (Clinical Biochemistry, Torbay and South Devon NHS Foundation Trust, Torquay, United Kingdom), A Sharma (Clinical Biochemistry, Torbay and South Devon NHS Foundation Trust, Torquay, United Kingdom)

    Aim: Faecal calprotectin analysis is increasingly requested by primary care to guide referrals for suspected inflammatory bowel disease. We aimed to verify the Inova-Werfen QUANTA Flash Calprotectin chemiluminescent immunoassay, with prior stool extraction, on the BIO-FLASH instrument and assess its suitability for clinical use, with interest in potentially reduced turnaround times.

    Methods: Imprecision of the extraction procedure was assessed by testing two patient samples five times per day on five separate days (with individual extractions). Imprecision of the analytical component was assessed by re-assaying one day of both patients’ extracts within-day and between-day. Patient samples (n=43) routinely referred for analysis on a commercial ELISA were also assayed on the BIO-FLASH; calprotectin concentrations and interpretative conclusions (negative/intermediate/positive) were compared. EQA samples (n=9) were assayed in duplicate, with results compared to all-lab trimmed means and other BIO-FLASH users.

    Results: Repeatability and intermediate imprecision (%CV) of the extraction procedure across 5 days was 14.2% and 9.2% respectively at a mean faecal calprotectin of 1424 mg/kg, and 33.6% and 6.9% respectively at 124 mg/kg. Re-analysis of individual stool extracts on separate days showed imprecision of <4%CV between replicates at both concentrations. Mean bias for the assay against the comparator ELISA was +147 mg/kg (+38%) across the concentration range. However, 98% of compared samples fell into the same diagnostic categories by both assays. EQA results were in diagnostic agreement with the all-lab trimmed means and other enrolled BIO-FLASH users.

    Conclusions: The QUANTA Flash Calprotectin assay and BIO-FLASH instrument are suitable for routine use. They are user-friendly, allowing continuous sample loading/unloading. Imprecision of the full procedure is comparable with other method %CVs shown by UK NEQAS. Diagnostic agreement with established assays is good. The time from extraction to result is approximately 70 minutes, therefore the assay may reduce turnaround times and allow repatriation of calprotectin testing.

  • P033 LC-MS/MS analysis of angiotensin I for assessment of plasma renin activity in clinical research

    D Foley (Scientific Operations, Waters Corporation, Wilmslow, United Kingdom), L Calton (Scientific Operations, Waters Corporation, Wilmslow, United Kingdom)

    Background: We have evaluated a LC-MS/MS method for the measurement of angiotensin I (Ang1) as a marker of plasma renin activity (PRA) for clinical research activities involving biomarkers of hypertension. An analytical method was developed using µElution Solid Phase Extraction (SPE) in 96-well plate format, reducing sample preparation time and optimizing analytical sensitivity.

    Methods: Ang1 was used to create calibrators and QC materials in 1% (w/v) Bovine Serum Albumin (BSA) in Phosphate Buffered Saline (PBS). Method precision was demonstrated in two ways; analysis of the spiked Ang1 BSA/PBS QCs and analysis of commercially available PRA controls. Samples were treated with Ang1 generation buffer and incubated for 3 hours at 37°C, diluted with Ammonium Hydroxide and Ang1-13C515N internal standard was added. Offline automated SPE was carried out with a Waters® Oasis® MAX µElution plate. Using an ACQUITY UPLC I-Class, samples were injected onto a Waters ACQUITY UPLC HSS T3 column with VanGuard T3, using a water/acetronitrile/formic gradient and quantified with a Waters Xevo TQ-S mass spectrometer.

    Results: The method was shown to be linear from 0.15-116 nmol/L for Ang1 (0.05–38.5 nmol/L/h). Coefficients of variation (CV) for total precision and repeatability on five separate occasions for Ang1 BSA/PBS QC samples were all ≤9.5% (n=30) for Ang1 at 0.77, 7.7 and 77.1 nmol/L. PRA controls were ≤7.7% over the five occasions with mean PRA values ranging from 1.52-10.28 nmol/L/h (1.97-13.33 ng/mL/h). Analytical sensitivity investigations demonstrate a CV<20% at 0.15 nmol/L (0.05 nmol/L/h) for Ang1 with S/N>10:1. Both extraction efficiency and matrix effects were demonstrated to be reproducible.

    Conclusions: A clinical research LC-MS/MS method has been developed to quantify Ang1 and evaluate PRA utilizing SPE with LC-MS/MS. This offline automated method demonstrates good linearity, precision, and accuracy, while providing high sample throughput capabilities.

    For Research Use Only, Not for use in diagnostic procedures.

  • P034 Verification of the BioPorto and Abbott NGAL assays on Abbott analysers

    K Chadwick (Clinical Biochemistry, The Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), S Whitehead (Clinical Biochemistry, The Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), C Ford (Clinical Biochemistry, The Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), R Gama (Clinical Biochemistry, The Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom)

    Introduction: Acute kidney injury (AKI) is common, costly and is associated with increased mortality. Serum creatinine (SCr) is a late marker of AKI, rising 24-48 h post-insult, by which time significant loss of renal function may have occurred. Neutrophil gelatinase-associated lipocalin (NGAL) is a biomarker for the early prediction of AKI; levels rise 2-6 h post insult. Commercial CE-marked automatable NGAL assays are marketed by Abbott Diagnostics (Abbott Park, IL, USA) (urine) and BioPorto Diagnostics (Hellerup, Denmark) (urine/plasma). We performed a laboratory verification of the Abbott assay on an Abbott Diagnostics i2000sr analyser and the BioPorto assay on an Abbott Diagnostics c16000 analyser and also validated the performance of the Abbott assay in plasma.

    Methods: Intra-inter- batch imprecision, linearity, recovery and limit of quantitation (LoQ) were assessed for the two assays (urine and plasma). Inter-assay agreement was assessed using 51 urine and plasma samples.

    Results: Both the Abbott and BioPorto assays demonstrated acceptable intra-/inter-batch imprecision using manufacturers supplied QC material; percentage coefficients of variation (%CVs) were 0.7-3.1% (BioPorto) and 2.2-6.8% (Abbott). The two assays displayed mean recoveries of 84-102% (BioPorto) and 111-117% (Abbott). Both assays were linear up to 6000 ng/mL. LoQs for the BioPorto and Abbott assays were 27.0 ng/mL and 2.7 ng/mL respectively using a CV of <20%. In a direct comparison between the two manufacturers kits, the BioPorto assay gave results that were, on average, 9.2% higher than the Abbott assay.

    Conclusion: The performance characteristics of plasma and urine NGAL measurement were similar for both Abbott and BioPorto assays (even though the Abbott assay is only CE marked for urine analysis). The Abbott assay had a much lower LoQ than the Bioporto assay.

  • P035 Comparison of immunoassay of fT4 with LC-MS/MS in abnormal thyroid states

    N Jassam (Department of Blood Sciences, Harrogate Hospital, Harrogate, United Kingdom), SJ Soldin (Department of Laboratory Medicine, NIH Clinical Center, Bethesda, United States), S Spoors (Department of Blood Sciences, Doncaster Royal Infirmary, Doncaster, United Kingdom), H Rana (Department of Blood Sciences, Doncaster Royal Infirmary, Doncaster, United Kingdom), N Lauber (Department of Blood Sciences, Harrogate Hospital, Harrogate, United Kingdom), JH Barth (Department of Blood Sciences, Leeds Teaching Hospitals , Leeds, United Kingdom)

    Introduction: The IFCC Committee for Standardisation of thyroid hormones have recently reported a protocol for the standardisation of free thyroxine (fT4) assays on 14 immunoassay (IA) platforms. The standardisation procedure used reference method procedures and blood samples from euthyroid patients. This study only examined samples from euthyroid subjects. We have sought to explore whether the relationship between IA and Liquid chromatography-tandem mass spectrometry (LC-MS/MS) are valid in abnormal thyroid states.

    Method: fT4 was measured using routine immunoassays from 3 manufacturers and compared with LC-MSMS. This study was performed on two panels of patients with thyroid disease, each panel comprised of 20 subjects.

    Results: In the range of fT4 >20 pmol/L, fT4 values from IAs were significantly lower than LC–MSMS with a bias of -39.9%, -23.5% and -45.3% for Siemens, Roche and Abbott respectively. The difference plot reveals an increasing negative bias with increasing fT4 concentration. Correlation equations of routine immunoassays are IA=0.45 MS+6.3 (Siemens), IA=0.74 MS+0.55 (Roche), IA= 0.43MS+5.14 (Abbott).

    Conclusion: Data from subjects with abnormal thyroid function is in agreement with the finding from the IFCC Harmonisation Committee, in which the comparison between FT4 measurements on the above mentioned IA platforms and ED-MS/MS reference method was performed in euthyroid subjects. For patients with hypothyroidism and hyperthyroidism there is, however, a poor correlation between IA platforms and LCMS/MS. This is clinically important as the LCMS/MS result correlates with the patient’s clinical condition. However, confirmatory study with larger cohort is required.

  • P036 Effect of in vitro haemolysis on plasma cortisol measured on the Beckman Unicel DxI800

    J Shepherd (Clinical Biochemistry, Hull and East Yorkshire Hospitals, Hull, United Kingdom), H Ian (Clinical Biochemistry, Hull and East Yorkshire Hospitals, Hull, United Kingdom), S Douglas (Clinical Biochemistry, Hull and East Yorkshire Hospitals, Hull, United Kingdom)

    To comply with UKAS, laboratories must follow manufacturer’s guidelines so that analytical processes are compliant with CE marking. Interference studies for haemolysis, lipaemia and icterus do not always cover the variation in sample quality commonly seen in laboratories. Particular inconvenience is caused when collecting small volume paediatric samples for dynamic function tests where haemolysis is frequent even in experienced hands.

    This study aims to mimic in vitro haemolysis and establish the effect on the Beckman Unicel DxI800 cortisol assay across the Haemolysis Index (HI) range measured using a Beckman AU system.

    Consent was obtained for using whole blood plasma taken during a selective venous sampling procedure where cortisol was already requested. Six separate whole blood plasma samples were each divided into 4 aliquots. One sample from each set was centrifuged and the plasma used to establish a baseline cortisol. The remaining aliquots were subjected to different periods of ultrasonic shear stress giving a range of haemolysis. Following centrifugation haemolysed aliquots were separated and stored at 40C for 12 h prior to analysis for cortisol. The effect of haemolysis was determined by establishing the difference between the baseline sample and corresponding haemolysed aliquots, thus enabling a correlation between HI and plasma cortisol levels using Spearman correlation methodology.

    Spearman correlation did not reveal a significant trend for cortisol vs haemolysis (p=0.60) over the range of HI tested (0-6 or 0 – >1000 mg/dL). Differences observed between results were within the expected analytical variation of the cortisol assay.

    Beckman’s protocol only gives validity to cortisol in samples with HI of 4 or less (<500 mg/dL) paediatric specimens from dynamic function tests regularly exceed this level. A simple methodology for in vitro haemolysis studies is described giving validity to reporting cortisol at greater levels of haemolysis than previously ascertained.

  • P037 Understanding infliximab drug, total and free antibody testing in therapeutic drug monitoring for patients with inflammatory bowel disease

    R Nice (Department of Clinical Biochemistry, Royal Devon & Exeter Hospital, Exeter, United Kingdom)

    Infliximab is a biologic drug used to treat moderate to severe inflammatory bowel disease (IBD). Infliximab targets the pro-inflammatory cytokine tumour necrosis factor-alpha. The success of infliximab is complicated by issues of primary non-response (PNR) or loss of response (LOR). LOR is in part caused by development of anti-drug antibodies against infliximab.

    Demand for therapeutic drug monitoring for infliximab in IBD has grown in recent years. There is however, no robust algorithm or guidelines for how and when infliximab drug and anti-drug antibody (ADA) levels should be used to inform treatment strategies. There are a variety of assays available for TDM, with particular confusion over the clinical utility of free (unbound) or total antibody measurements.

    This study aims to investigate the effect of ADA levels on infliximab drug measurement and the effect of drug levels on the measurement of free and total ADA and to investigate the clinical utility of free vs total ADA measurement.

    This study involved performing interference experiments using Immundiagnostik IDKmonitor ELISA assays for infliximab drug, free and total antibody levels followed by analysis of a clinical cohort of 81 patients samples on all 3 assays to investigate the clinical utility of free and total antibody testing.

    Interference studies show Infliximab drug measurement is unaffected by ADA concentrations up to 358.8 AU/mL, it is not possible to detect free ADA in the presence of even small amounts of infliximab (1.1 mg/L). The IDK total ADA assay is not completely drug tolerant, however at therapeutic drug levels it provides reliable antibody measurements. Conclusions of this study are that total ADA testing is a better indicator of the development of immunogenicity than free ADA. A combination of free and total antibody measurement may be useful in predicting LOR or determining treatment strategies for patients who have lost response.

  • P038 Evaluation of the re-formulated Abbott Architect free T3 assay reference range

    TG Morris (Clinical Blood Sciences, St George's University Hospitals NHS Foundation Trust, London, United Kingdom), H Johnstone (Chemical Pathology, Epsom and St Helier Hospitals NHS Trust, London, United Kingdom), N Johri (Chemical Pathology, Epsom and St Helier Hospitals NHS Trust, London, United Kingdom)

    Introduction: Free thyroxine (T4) and thyroid-stimulating hormone (TSH) are the first-line tests used by clinical laboratories to investigate thyroid function. The measurement of free 3,5,3’ triiodothyronine (T3) may also be required in cases of suspect free T3 thyrotoxicosis. The Abbott ARCHITECT free T3 assay has recently changed from a 2-point to a 6-point calibration. As a consequence the ‘normal’ reference range for the re-formulated assay has changed from 2.6-5.7 pmol/L to 2.9-4.9 pmol/L. The aim of this current study was to investigate whether this range is suitable for our local population.

    Methods: The levels of free T3 were retrospectively measured in euthyroid patients with normal free T4 (9-22 pmol/L) and TSH (0.35-5 mU/L) levels. Free T4 and TSH measurements were made using Abbott ARCHITECT assays. Free T3 levels were measured in 134 patients in total, 75 from the St Helier Hospital (STH) region and 59 from the Epsom General Hospital (EGH) region. Reference ranges were calculated from the data by determining the mean ±2 standard deviations (SD). All data was normally distributed.

    Results: For all 134 patients in the study the free T3 reference range was calculated to be 2.26-5.14 pmol/L (mean = 3.70, SD = 0.72). The free T3 reference ranges for the STH and EGH regions were 2.47-5.36 pmolL (mean = 3.92, SD = 0.72) and 2.18-4.66 pmol/L (mean = 3.42, SD = 0.62), respectively.

    Conclusion: The re-formulated reference range quoted by Abbott is not appropriate for the STH and EGH populations. Interestingly, the STH and EGH populations had different reference ranges which may reflect differences in ethnicity, population, age and analyser used.

  • P039 The stability of common biochemical analytes in modern serum tubes

    S Choudhury (Department of Clinical Biochemistry, Imperial College Healthcare NHS Trust, London, United Kingdom), S Talluri (Department of Clinical Biochemistry, North West London Pathology, London, United Kingdom), V Kripesh (Department of Clinical Biochemistry, North West London Pathology, London, United Kingdom), O Yasar (Department of Clinical Biochemistry, North West London Pathology, London, United Kingdom), J Alaghband-Zadeh (Department of Clinical Biochemistry, Imperial College Healthcare NHS Trust, London, United Kingdom), J Cegla (Department of Clinical Biochemistry, Imperial College Healthcare NHS Trust, London, United Kingdom), T Tan (Department of Clinical Biochemistry, Imperial College Healthcare NHS Trust, London, United Kingdom)

    Introduction: Current practice in our laboratory has been to cancel samples that have not been separated by 8 hours. Divergent practices across GP surgeries produces a significant number of samples that are received over eight hours post-collection, leading to the cancelling of potassium and phosphate results. To reduce the number of samples affected, we investigated the stability of common analytes in modern serum collection tubes up to 24 hours.

    Method: Eleven healthy participants provided blood samples collected by two experienced physicians, taken in a single blood draw within a 30 minute period. Six females and five males participated. Four BD Vacutainer® SST™ Tubes (yellow-top gel tubes) and four BD Vacutainer® Plus Plastic Serum Tubes (red-top tubes), were taken. One yellow-top and red-top tube was separated and analysed immediately (T=0), with a further yellow-top and red-top tube separated at fixed timepoints post-collection: T=8, T=12 and T=24. Routine chemistry analytes were analysed on the Abbott Architect platform. The percentage deviation of each sample from the T=0 value was calculated and compared to the Total Allowable Error (TAE) specified on the ‘Westguard Desirable Biological Variation Database’.

    Results:

    Potassium: The mean percentage deviation from T=0 for yellow-top tubes was 3.56%, 4.86% and 7.91% for T=8, T=12 and T=24, respectively. In the red-top tubes, the percentage deviation was 4.57%, 5.55% and 9.20% respectively. Westguard rules permit a TAE of 5.61%.

    Phosphate: The mean percentage deviation from T=0 for yellow-top tubes was 6.83%, 9.35% and 35.43% for T=8, T=12 and T=24, respectively. In the red-top tubes, the percentage deviation was 4.99%, 6.28% and 26.96% respectively. Westguard rules permit a TAE of 10.11%.

    Conclusion: The acceptable delay to separation of samples should be extended from 8 hours to 12 hours.

  • P040 Integrating immunoassay and HPLC methods: a two method strategy for HbA1c analysis

    A Sanders (Department of Clinical Biochemistry, Dudley Group NHS Foundation Trust, Dudley, United Kingdom), Z Esmail Zadeh (Department of Clinical Biochemistry, Dudley Group NHS Foundation Trust, Dudley, United Kingdom), A Chadburn (Department of Clinical Biochemistry, Dudley Group NHS Foundation Trust, Dudley, United Kingdom), H Ashby (Department of Clinical Biochemistry, Dudley Group NHS Foundation Trust, Dudley, United Kingdom), P Mohammed (Department of Clinical Biochemistry, Dudley Group NHS Foundation Trust, Dudley, United Kingdom)

    An increasing incidence of diabetes along with the use of HbA1c for diagnosis as well as monitoring has led to a huge increase in HbA1c requests in recent years. Most UK laboratories use liquid chromatographic methods, often employing multiple analysers to manage workload. Immunoassay methods have the potential to increase throughput but have remained unpopular due to their inability to detect variant haemoglobins and the potential to report false HbA1c results. In our laboratory we use Tosoh G8, indicating when it cannot be used for diagnosis due to the presence of probable Hb variants. As part of an equipment refresh we have installed a Roche c513 immunoturbidometric HbA1c analyser and undertaken an extensive comparison of results from patients with and without the common haemoglobin variants (HbS/C/D/E/F). Our aim was to devise a strategy utilising one or both methods to deliver a safe and efficient HbA1c service.

    Comparison of HbA1c results from patients without a variant Hb (n=175), as well as the variants: HbS (n=103), HbC (n=21), HbE (n=25), HbD (n=20) and HbF (n=18) gave the following: Roche (HbA1c) = 1.04 (TosohHbA1c)-3.37, Roche (HbA1c)=1.02 (TosohHbA1c)-2.58, Roche (HbA1c) = 1.03 (TosohHbA1c) - 4.39, Roche (HbA1c) = 1.14 (TosohHbA1c) - 7.23, Roche (HbA1c) = 1.03 (TosohHbA1c) - 4.82, Roche (HbA1c) = 0.7 (TosohHbA1c) + 5.05, respectively. Our findings of HbF interference is in agreement with the contents of the kit IFU. To ensure we have the opportunity to detect probable Hb variants we created a logic rule in our LIMS (LabCentre, CliniSys) whereby all first-time HbA1c would have a Tosoh G8 HbA1c. Any subsequent requests would have a Roche c513 HbA1c and a comment indicating that the result is for trend analysis only.

    We believe that this 2 test approach offers a low risk solution that could work for other laboratories, and may provide the answer as to how to deliver a robust and future-proof service.

  • P041 Measurement of aldosterone concentration in serum

    R Tibbs (Department of Clinical Biochemistry (Viapath), King's College Hospital, London, United Kingdom), DR Taylor (Department of Clinical Biochemistry (Viapath), King's College Hospital, London, United Kingdom), RP Vincent (Department of Clinical Biochemistry (Viapath), King's College Hospital, London, United Kingdom)

    Aims: Quantitative determination of aldosterone concentrations in EDTA-plasma is routinely performed by competitive immunoassay on the DiaSorin Liaison XL platform in our laboratory. Commonly, aldosterone is requested in combination with renin in order to calculate an aldosterone:renin ratio as part of the investigation of suspected primary hyperaldosteronism (PA). EDTA-plasma is an absolute requirement for the renin Liaison XL assay. Alternatively, aldosterone may be requested in combination with cortisol as part of our adrenal venous sampling protocol (a radiological technique used for the lateralisation of autonomous aldosterone secretion in cases of confirmed PA). Serum cortisol is measured using the Siemens Centaur immunoassay. In this study, we investigated whether serum was also suitable for aldosterone analysis, as this would allow a single serum sample to be taken from each sampling site for cortisol and aldosterone analysis.

    Methods: Aldosterone concentrations were measured on fifty sets of paired serum and EDTA-plasma. Samples were chosen to cover a wide range of aldosterone concentrations (130-126,000 pmol/L), including the very high values expected in adrenal venous sampling. Results were compared using Deming regression and Bland-Altmann analysis (Analyse IT v2.2). Linearity of the serum assay and serum aldosterone stability were also assessed.

    Results: Serum and EDTA-plasma aldosterone showed good correlation (Deming fit 56.18+0.98x) and a bias of 0.9%. Serum aldosterone was found to be linear up to 2770 pmol/L, and samples were stable for up to four freeze-thaw cycles.

    Conclusions: Serum is a suitable sample type for aldosterone analysis on the DiaSorin Liaison XL platform. Results show good correlation to those obtained using EDTA-plasma and therefore separate reference ranges for each sample type are not required. Measurement of aldosterone in serum will simplify sample collection procedures for adrenal venous sampling.

  • P043 Hyperphosphataemia: bringing phosphate down a PEG or two in a case of multiple myeloma

    L Murfitt (Clinical Biochemistry, Salford Royal Foundation Trust, Salford, United Kingdom)

    Background: A 68-year-old female with a history of IgG kappa monoclonal gammopathy displayed persistent hyperphosphataemia in the absence of renal dysfunction or hypoparathyroidism. The patient’s serum phosphate concentration progressively increased over the previous 4 years, mirroring the elevations of monoclonal protein. Previously, there have been reports of pseudohyperphosphataemia in the presence of paraproteinaemia. It has been suggested that paraproteins may react with ammonium molybdate causing increased turbidity or alternatively, forming a macrocomplex with phosphate. Therefore, it was hypothesised that the elevated phosphate concentrations may be the result of pseudohyperphosphataemia.

    Aim: The aim of this case study was to monitor the progression of hyperphosphataemia following treatment for multiple myeloma and investigate the presence of a potential pseudohyperphosphataemia.

    Method: The patient’s blood results were monitored via the laboratory computer system. Serum samples were diluted 1:2 with PEG 6000 and mixed before centrifugation at 2800 rpm for 30 minutes at 4°C. Phosphate levels were analysed pre- and post-PEG precipitation on the Siemens Advia 2400 and % free phosphate calculated.

    Results: PEG precipitation of a pre-chemotherapy serum sample (phosphate = 4.59 mmol/L) demonstrated a phosphate recovery level of 32%, producing a post-precipitation phosphate concentration of 1.42 mmol/L (reference range: 0.8-1.5 mmol/L) which indicated the presence of a pseudohyperphosphataemia. Subsequent samples displayed similar patterns with serum phosphate concentrations normalising following PEG precipitation. As treatment progressed, the levels of monoclonal protein and immunoglobulins decreased which coincided with a reduction in serum phosphate levels to within the reference range. Likewise, there was a greater recovery of phosphate following PEG precipitation further suggesting that the previous hyperphosphataemia was most likely due to phosphate complexed with IgG.

    Conclusions: This case study revealed the presence of pseudohyperphosphataemia in the setting of multiple myeloma. This data highlights the importance of thoroughly investigating spurious results to ensure appropriate patient care and management.

  • P044 The Roche Elecsys Prolactin II assay: a single cut-off for detection of macroprolactinaemia

    AJ Chadburn (Department of Clinical Chemistry, The Dudley Group of Hospitals NHS Foundation Trust, Dudley, United Kingdom), P Mohammed (Department of Clinical Chemistry, The Dudley Group of Hospitals NHS Foundation Trust, Dudley, United Kingdom), H Ashby (Department of Clinical Chemistry, The Dudley Group of Hospitals NHS Foundation Trust, Dudley, United Kingdom)

    Macroprolactin is known to interfere in all routine prolactin immunoassays to some degree. Precipitation of macroprolactin with polyethylene glycol (PEG) is a simple technique to remove this interference. However, PEG precipitation itself has an effect on prolactin assays due to unwanted precipitation of monomeric prolactin; the extent of this is assay dependent. The aims of this study were to investigate the effect of PEG precipitation on the Roche Elecsys Prolactin II assay; to establish cut-offs for the detection of macroprolactin; and to assess the need for a correction factor in the estimation of monomeric prolactin.

    PEG precipitation was carried out on 156 sera with total prolactin >700 (range 709-9742) mU/L. Mean post-PEG recovery was 89.7±14.5% (± standard deviation). Thirty six percent of patients had previously undergone PEG precipitation and analysis using the Tosoh ST AIA-PACK PRL assay. When compared to the Tosoh results, a cut-off of 70% recovery in the Roche assay gave 100% sensitivity for ruling out macroprolactin; a cut-off of 70% recovery also correctly identified 100% of the patients with macroprolactin (three patients). However, since the Tosoh assay had an equivocal range, using a single cut-off “misclassified” five patients.

    Roche post-PEG recovery >87% was 88% sensitive for excluding macroprolactin but included no equivocal results by the Tosoh method. In samples with recovery >87% the percentage difference between the mean post-PEG and saline-diluted prolactin was 4% i.e. the correction factor for estimation of monomeric prolactin was 1.04x post-PEG prolactin.

    In conclusion, despite limited precipitation of monomeric prolactin with PEG in the Roche Elecsys Prolactin II assay, PEG precipitation of sera with raised prolactin is still essential. A cut-off of 70% post-PEG recovery can be applied to rule in/out macroprolactinaemia. A correction factor of 1.04x post-PEG prolactin gives an estimate of monomeric prolactin.

  • P045 Evaluation of the CEDIA® buprenorphine II assay as part of a drugs of abuse screening service

    R Cooper (Department of Clinical Biochemistry, Barts Health NHS Trust, London, United Kingdom), A Dawnay (Department of Clinical Biochemistry, Barts Health NHS Trust, London, United Kingdom)

    Background: ThermoScientific have recently released the CEDIA® buprenorphine II assay; which claims improved cross-reactivity with norbuprenorphine and norbuprenorphine-3-glucuronide, and reduced cross-reactivity with opiates, such as morphine, dihydrocodeine and codeine, compared with the CEDIA® buprenorphine I assay. The second-generation assay has the potential to reduce the number buprenorphine false-positive results requiring confirmation using analytically superior methods, such as LC-MS/MS. The aim of this study was to evaluate and compare the CEDIA® buprenorphine II to the buprenorphine I assay.

    Methods: To evaluate the CEDIA® buprenorphine II assay a three-way comparison, using 57 anonymised patient urine specimens, was performed on the CEDIA® buprenorphine I and II assays, and an in-house LC-MS/MS confirmatory method. Imprecision was assessed using two levels of quality control material (7.5 ng/mL and 12.5 ng/mL buprenorphine). We assessed bias using five EQA samples and linearity by dilution of calibration material in human urine.

    Results: Imprecision of the CEDIA® buprenorphine II assay revealed a repeatability coefficient of variation (CV) of 4.9% and 4.5% and intra-laboratory CV of 6.7% and 5.5%, at 7.5 ng/mL and 12.5 ng/mL buprenorphine, respectively. Linearity studies were acceptable with an r2=0.998 and no bias was observed using EQA material (p=>0.05). The patient comparison revealed a greatly improved sensitivity and specificity of 100% and 95.65%, respectively, for CEDIA® buprenorphine II; compared to a sensitivity and specificity of 94.12% and 26.09%, respectively, for CEDIA® buprenorphine I.

    Conclusion: We have demonstrated that the CEDIA® buprenorphine II screening assay has an improved sensitivity and specificity compared to CEDIA® buprenorphine I and discovered the positive predictive value increased by 32% using CEDIA® buprenorphine II. Incorporation of this assay has the potential to reduce the number of buprenorphine confirmations on either GC-MS/MS or LC-MS/MS by approximately 32%, as well as reduce the risk of false-negatives due to the increased cross-reactivity with buprenorphine’s metabolites.

  • P046 Evaluating a new method using polyethylene glycol precipitation to screen for macroprolactinaemia

    C Agius (Clinical Chemistry, Mater Dei Hospital, Msida, Malta), I Brincat (Clinical Chemistry, Mater Dei Hospital, Msida, Malta), C Thomas (Clinical Chemistry, Mater Dei Hospital, Msida, Malta), J Grima (Clinical Chemistry, Mater Dei Hospital, Msida, Malta), G Buhagiar (Clinical Chemistry, Mater Dei Hospital, Msida, Malta)

    Aim: To evaluate the polyethylene glycol (PEG) precipitation procedure based on a recently published method (Chen Y et al. A new method of using polyethylene glycol precipitation of macroprolactin to detect genuine hyperprolactinemia. J Clin Lab Anal 2016; 30: 1169-1174).

    Methodology: Patient samples (n=20) with prolactin (PRL) concentrations ranging from 421 to 5999 mU/L were examined. The percentage prolactin recovery (% PRL R) for each sample aliquot was established using PEG treatment. Patient sera were either undiluted or else subjected to 5-fold dilution using either saline or the manufacturer PRL kit diluent. Patient sera were also subjected to 5-fold dilution with saline or the manufacturer PRL kit diluent to establish baseline %PRL R. All analytical runs were performed on a Siemens ADVIA Centaur XP platform.

    Results: Non-PEG treated samples having 5-fold diluted serum and different diluents demonstrated mean % PRL R of 104% for saline and of 99% for the kit diluent. Undiluted serum with PEG treatment resulted in a mean %PRL R of 76%. PEG-treated 5-fold diluted serum with saline resulted in a mean %PRL of 140%. PEG-treated 5-fold diluted serum with kit diluent resulted in %PRL R of 99%. Two samples had a %PRL R of 38% in non-diluted sera treated with PEG. These would have been considered positive for the presence of macroPRL (local established cut-off of less than 45% indicative of macroPRL), however, the % PRL R of PEG treated 5-fold diluted sera did not indicate the presence of macroPRL, with %PRL R greater than 60% for both saline and kit diluents.

    Conclusion: Five-fold diluted serum with 25%-PEG 6000 may possibly be a superior method to detect monomeric PRL by allowing for a more selective precipitation. Best %PRL R were obtained using the manufacturers PRL kit diluent as opposed to saline.

  • P047 Problems with validating the 51Cr-EDTA clearance method for the glomerular filtration rate determination

    M Jankute (Department of Clinical Chemistry, Birmingham Children's Hospital, Birmingham, United Kingdom), S Heap (Department of Clinical Chemistry, Birmingham Children's Hospital, Birmingham, United Kingdom), M Wilson (Medical Physics Department, Queen Elizabeth Hospital Birmingham, Birmingham, United Kingdom)

    Glomerular filtration rate (GFR) is a commonly accepted indicator of kidney function. Estimated GFR is routinely calculated using plasma creatinine concentration, height and factor, but accurate GFR can only be measured using tracers that have no significant tubular secretion or reabsorption. The 51chromium-ethylenediaminetetraacetic acid (51Cr-EDTA) clearance method remains one of the most commonly used techniques to accurately determine GFR. However, it is not commercially available and lacks an external quality assessment (EQA) scheme. As a result, the 51Cr-EDTA clearance test remains developed “in-house” and its accuracy is assessed at the discretion of the laboratory. Due to the nature of the radioactive tracer, variation in calculations and gamma counters used in different laboratories, it is an extremely difficult method to validate. The aim of our study was to validate the 51Cr-EDTA clearance method for the GFR determination to UKAS Accreditation 15189 standards. The routine parameters were all acceptable when using Wallac 1470 Wizard Gamma Counter with an assay repeatability of <2.0%, intermediate imprecision of <2.5% and measurement uncertainty of <3.5% for low counts and <4.0% for high counts. Effects of haemolysis on the patients’ samples proved to be insignificant in this type of assay. Finally, we assessed the performance of our method and comparability of the GFR reported values with other providers by sending samples to the Medical Physics at the University Hospital Birmingham (UHB). Both re-count and re-calculation were performed by UHB and calculated GFR values were comparable to those generated by our laboratory. We have successfully validated 51Cr-EDTA clearance method for GFR determination and continue to monitor our performance with the UHB collaboration.

  • P048 Less is more: are dried blood spots the answer for therapeutic drug monitoring of lithium?

    AJM Clark (Department of Clinical Biochemistry, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), NL Barlow (Department of Clinical Biochemistry, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), JD Berg (Department of Clinical Biochemistry, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom)

    Lithium salts are therapeutic drugs used extensively as mood stabilisers in the treatment of bipolar affective disorder (BPAD) in patients under the care of mental health services, however a narrow therapeutic range, variable metabolism between individuals, and propensity for renal and thyroid toxicity necessitate regular monitoring of blood lithium levels to direct dosage. Dried blood spot (DBS) sampling dominates in the field of neonatal screening, being regarded as less invasive and requiring much smaller volumes of blood than standard venepuncture, but why has the use of capillary blood microsamples not been extended to include routine blood testing in adults? Monitoring of blood lithium presents an interesting opportunity to utilise capillary blood microsampling to improve the patient experience of therapeutic drug monitoring (TDM).

    Method development and validation for a capillary blood microsample assay using inductively-coupled plasma mass spectrometry (ICP-MS) was carried out at SWBH NHS Trust, including comparison between two capillary blood collection devices – dried blood spots (DBS) on filter paper and the Mitra® Microsampler.

    Extraction optimisation built upon existing laboratory operating procedures, with addition of a centrifugation step. A 3 mm punch from a DBS, estimated to contain approximately 3-6 μL of blood, produced adequate counts per second (cps) for detection of lithium on the ICP-MS (top calibrator:blank device ratio of >1000 cps). The assay performed well with laboratory-produced DBS calibrators created from spiked whole blood (r2>0.997) covering a concentration range from 0.10-3.00 mmol/L, and whole blood EDTA samples from patients on lithium therapy gave detectable results with potential to correct results to produce a “serum equivalent” result (consistent bias of -20-30% from original serum result). Assay development was deemed successful with %CVs of ≤10%, and next steps will test the assay using paired serum and capillary bloodspot samples collected from patients on lithium therapy and gauging user opinion.

  • P049 Validation of a liquid chromatography tandem mass spectrometry method to assess anti-hypercholesterolaemia drug non-adherence in urine

    A Cross (Department of Chemical Pathology and Metabolic Diseases, University Hospitals of Leicester NHS Trust, Leicester, United Kingdom), A Burns (Department of Chemical Pathology and Metabolic Diseases, University Hospitals of Leicester NHS Trust, Leicester, United Kingdom), R Cole (Department of Chemical Pathology and Metabolic Diseases, University Hospitals of Leicester NHS Trust, Leicester, United Kingdom), P Gupta (Department of Chemical Pathology and Metabolic Diseases, University Hospitals of Leicester NHS Trust, Leicester, United Kingdom)

    Introduction: Hypercholesterolaemia is an established risk factor for CVD. Several classes of lipid-lowering therapies exist including HMG-CoA reductase inhibitors (statins), the fibrates, and cholesterol-absorption inhibitors such as ezetimibe. Non-adherence to medication is a critical barrier in the management of chronic disease. LC-MS/MS has been used as a powerful tool to assess non-adherence to antihypertensive therapy and using this test a third of patients were non-adherent to prescribed antihypertensive medications.

    A qualitative LC-MS/MS method was developed and validated for assessing non-adherence to atorvastatin, rosuvastatin, ezetimibe and fenofibrate in urine.

    Method: Urine samples were extracted using a solid phase extraction procedure and analysed on an Agilent 6410 LC-MS/MS system using a Zorbax SB-C18 column. Mobile phases used were 0.1% glacial acetic acid in deionised water and 0.1% glacial acetic acid in methanol. Two chromatographic methods were developed.

    The validation parameters investigated included lower limit of detection (LLOD), matrix and analyte interference, dilution integrity, carryover, and ion suppression and enhancement. Stability was assessed at time points over a month using patient samples attained from the University Hospital of Leicester’s lipid clinic who were on the drugs of interest at a wide-range of doses. Stability was also assessed using blank urine spiked with reference material.

    Results: All analytes had LLODs ≤0 ng/mL. No matrix or analyte interference was found. Pre- and post-extraction stability of analytes was acceptable. All analytes showed no significant carryover up to high levels concentrations. Ion suppression and enhancement was observed however this had no clinical significance.

    Conclusion: The method has been successfully validated, with two alternative chromatographic programs. We anticipate that this method will be useful in assessing non-adherence in patients with hypercholesterolaemia.

  • P050 Evaluation of a new calibrator for the Sentinel FOB Gold faecal immunochemical test

    C John (NHS Bowel Cancer Screening Programme - Southern Hub, Berkshire and Surrey Pathology Services, Royal Surrey County Hospital, Guildford, United Kingdom), C Piggott (NHS Bowel Cancer Screening Programme - Southern Hub, Berkshire and Surrey Pathology Services, Royal Surrey County Hospital, Guildford, United Kingdom)

    The faecal immunochemical test (FIT) is used worldwide in many bowel cancer screening programmes, and is recommended in NICE guidelines as a suitable test for ruling out colorectal cancers in patients with low risk symptoms. FIT uses antibodies specific to the globin moiety of haemoglobin (Hb). Several FIT assays are available but there is no international standardisation so each method has the potential to give different results. The FOB Gold FIT assay (Sentinel Diagnostics) on the SENTiFIT 270 analyser recovered differently compared with three other quantitative FIT systems. It showed a concentration-related positive bias with results up to 3 times higher at the top of the analytical range. Sentinel Diagnostics have developed a new calibrator traced to a reference material (ERM-467).

    The aim of this study was to investigate whether Hb measured with the new Sentinel calibrator gave results more closely aligned to three other FIT methods.

    Ten Hb-lysate and 18 EQA samples were analysed on the SENTiFIT 270 (0-912 µg Hb/g faeces) using the original calibrator and new calibrator. The same samples were run on 3 other quantitative FIT systems, the HM-JACKarc (Kyowa Medex Co. Ltd), OC-SENSOR DIANA/PLEDIA (Eiken Chemical Co.) and NS-Prime (Alfresa Pharma Corp.).

    The mean relative percentage bias of the results obtained with the 3 systems and the original SENTiFIT calibrator for the lysate samples were: HM-JACKarc -75.9%, OC-SENSOR -98.4% and NS-Prime -130%. With the new calibrator the percentage bias for the lysate comparisons were 10.1%, -12.1% and -56.1% on the respective analysers. EQA sample results showed a similar trend.

    This study shows that Hb concentrations using the new Sentinel calibrator are more closely aligned to the other 3 FIT systems using lysate and EQA samples. Further work is needed to assess whether a similar trend is seen for faecal samples.

  • P146 Serum EDTA analysis on an Abbott Architect: development and validation of the assay plus an assessment of its clinical value

    SJ Whitehead (Department of Clinical Chemistry, New Cross Hospital, Wolverhampton, United Kingdom), K Chadwick (Department of Clinical Chemistry, New Cross Hospital, Wolverhampton, United Kingdom), C Ford (Department of Clinical Chemistry, New Cross Hospital, Wolverhampton, United Kingdom), R Gama (Department of Clinical Chemistry, New Cross Hospital, Wolverhampton, United Kingdom)

    Background: Potassium ethylenediaminetetraacetic acid (kEDTA) is a widely used blood anticoagulant in laboratory sample tubes. EDTA contamination of clinical chemistry serum samples may result in spurious hyperkalaemia and hypocalcaemia. If unrecognised, this could adversely affect patient care. Strategies for identifying serum kEDTA contamination are typically based upon a combination of raised potassium and low alkaline phosphatase and calcium results, but this approach is subjective. We performed a validation of an assay for EDTA on an Abbott Architect c16000 analyser with the aim of introducing it into routine use.

    Methods: Intra-/inter-batch imprecision, linearity, recovery, interference and carryover were assessed.  Serum was spiked kEDTA plasma to mimic sample contamination and the effect on EDTA, potassium, calcium, magnesium, ALP and zinc concentration was established.

    Results: The assay displayed acceptable imprecision, recovery and linearity. No significant carryover was detected but the assay was subject to positive interference from haemolysis. EDTA contamination was detectable when serum was contaminated with 1% (v:v) kEDTA plasma. A change in serum potassium of 0.54 mmol/L (11.9% was observed) at a measured EDTA concentration of 0.19 mmol/L; equivalent to a contamination of 3.2% (v:v). At the same level of contamination, reductions in measured levels of calcium (4.7%), zinc (-22.2%) and ALP (4.7%) were observed; no change was observed for magnesium.

    Conclusions: The EDTA assay displayed acceptable performance.  Even small levels of kEDTA contamination were detectable. Routine measurement of serum EDTA concentrations has the potential to identify low levels of contamination which may not be detected using more subjective measures.

Miscellaneous

  • P052 Striking biochemical abnormalities in three young, adult male patients with drug/mental health issues

    L Bernstone (Department of Clinical Biochemistry, Wythenshawe Hospital, Manchester University NHS Foundation Trust, Manchester, United Kingdom), T Morris (Department of Clinical Biochemistry, Wythenshawe Hospital, Manchester University NHS Foundation Trust, Manchester, United Kingdom)

    Over a three week period at our hospital, three young, adult males presented to A&E with markedly abnormal electrolytes and/or severe AKI.

    Patient 1 had been found unconscious at home following a fall; he had severe pressure sores and hadn’t been seen for 16 days. His past history included epilepsy, psychosis, and cannabis use. Initial bloods showed hypernatraemia, hyperkalaemia, raised CK and AKI stage 3 with a urea of 92.8 mmol/L and a creatinine of 652 µmol/L. The AKI and hypernatraemia were thought to be due to severe dehydration and he was admitted to ITU for fluid resuscitation. On discharge his renal function had returned to normal.

    Patient 2 presented with episodes of confusion, hallucinations and panic attacks. He reported recent vomiting, decreased mood and poor appetite. There was a past history of alcohol dependence and heavy cannabis use. He was found to have normal renal function; however serum sodium was markedly low at 114 mmol/L and hypokalaemia was also present. Subsequent blood gas analysis revealed metabolic alkalosis. His confusion was felt to be due to the combined effects of alcohol, cannabis and hyponatraemia. Serum sodium normalised following administration of 0.9% saline.

    Patient 3 presented with jaundice, muscle pain and recent lack of fluid and dietary intake. He was currently receiving treatment for paranoid schizophrenia but had no other medical history of note. Initial bloods showed AKI stage 3 with significantly elevated ALT and bilirubin. It emerged that prior to his admission he had been found in a river with rodent bites to his body. Subsequent testing for leptospirosis proved positive and this was determined to be the cause of his hepatorenal failure.

    These unusual cases illustrate how a combination of mental health issues, drug/alcohol misuse and underlying medical conditions can result in very extreme biochemical abnormalities in young patients.

  • P053 Infliximab and adalimumab drug and free antibody monitoring: comparison of three ELISA assays

    M Perry (Blood Sciences, Royal Devon and Exeter NHS Foundation Trust, Exeter, United Kingdom), T McDonald (Blood Sciences, Royal Devon and Exeter NHS Foundation Trust, Exeter, United Kingdom)

    Introduction: The monoclonal antiTNF alpha therapies infliximab and adalimumab are expensive treatments that have revolutionised management of inflammatory bowel disease. Patient response is variable and a poor response may be due to development of endogenous anti-drug antibodies to the therapies. Measuring both the drug levels and the anti-drug antibodies may help to guide patient management. There are several different commercially available ELISA assays which measure antiTNF drug and anti-drug antibodies. The National Institute for Health and Care Excellence (NICE) have recognised that it is a priority to understand how results from different commercial ELISA compare if antiTNF monitoring is going to become established in clinical practice.

    Aim: The aim of this work is to assess the comparability of infliximab drug levels, free infliximab antibodies, adalimumab drug levels and free adalimumab antibody levels measured using three different commercially available ELISA kits manufactured by Theradiag, Promonitor and Immundiagnostik.

    Methods: Surplus serum samples received in the Royal Devon & Exeter Blood Sciences laboratory for antiTNF monitoring were analysed for infliximab or adalimumab drug and anti-drug antibody levels using three different ELISA methods.

    Results:

    There is agreement of infliximab drug levels and adalimumab drug levels between the different manufacturers. Comparison of the anti-drug antibody assays demonstrated variable agreement between the ELISA assays (all 3 assays gave the same result in 79% of samples for free anti-infliximab antibodies; and all 3 assays gave the same result in 93% of the samples analysed for free adalimumab-antibodies) although this was limited by a small proportion of anti-drug antibody positive samples.

    Conclusion: Laboratory staff and clinicians should be aware of the differences between the antiTNF drug methods and antibody methods. Reference standard material should be made available to standardise the methods for antiTNF monitoring purposes.

  • P054 Application of NPEx at HEFT

    R Henderson (Department of Biochemistry and Immunology, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), D Husband (Directorate of Laboratory Medicine, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), C Carter (Department of Biochemistry and Immunology, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), C Webster (Department of Biochemistry and Immunology, Heart of England NHS Foundation Trust, Birmingham, United Kingdom)

    NPEx (National Pathology Exchange) is a national service to connect UK pathology laboratories together through a single exchange hub so that requests and results can be communicated electronically between LIMS in a matter of seconds. The use of NPEx improves turnaround times, reduces error rates, reduces paperwork, allows tracking of specimens between laboratories, and generally improves the robustness of the system for sending work between laboratories.

    At Heart of England NHS Foundation Trust (HEFT) the LIMS is DXC iLab TP v2.1. ARES interface is used for incoming NPEx orders and outgoing results, and an instrument interface (Lab2Lab) is used for outgoing NPEx orders and incoming results.

    At HEFT, NPEx has been introduced for handling routine work received from other NHS laboratories, and returning the results directly to their LIMS. The use of NPEx to process work sent out to referral laboratories is currently being implemented.

    Additionally, NPEx is being used to process external quality assurance (EQA) requests and results, in order to reduce error rates and reduce staff time spent processing requests and returning results to the EQA provider. Currently general chemistry, eGFR and steroid EQA requests and results are handled via NPEx at HEFT, with plans to extend to other analytes imminently.

    Another application of NPEx at HEFT is a partnership with Werlabs, a private company that offers blood tests with results going directly back to the patient following a review by the Werlabs medical team. This is currently in the testing phase, with plans to introduce it for patients in 2018.

    This work demonstrates that there are several applications for NPEx, each coming with its own challenges, and will discuss the experiences that HEFT has had in the initial phase of NPEx use, as well as demonstrating the benefits of its use in the laboratory.

  • P055 An audit of reducing substances in a university hospital chemical pathology laboratory

    M Mirzazadeh (Department of Chemical Pathology, Epsom and St Helier Hospitals NHS Trust, Carshalton, United Kingdom), H Johnstone (Department of Chemical Pathology, Epsom and St Helier Hospitals NHS Trust, Carshalton, United Kingdom)

    Aim: To review the requesting for urine and faecal reducing substances.

    Results: Requests for reducing substances were studied from April 16 to March 17. Of 362 requests 76% (275) were from primary care and the rest 24% (87) from hospital, 258 ( 71%) in children and 104 (29%) in adults; 89% (321) of requests were for faecal reducing substances and the rest were for urine 11% (41). No sample (0%) was reported as positive for faecal reducing substances, whereas 1 sample was reported as trace positive (0.3%). From 41 urine samples, 1 (2.5%) was reported positive. One hundred requests were randomly selected to study the clinical information on the request card. The most common reason for faecal reducing substances was lactose intolerance and for the urine it was prolonged jaundice.

    Discussion: Considering the incidence of lactose intolerance in Northern Europe population of 5% and higher rate in mixed ethnic population covered by the trust, the positive rate is expected to be more than 40 in 362 samples. The reason for the low detection rate observed is several fold: the stool sample is unstable and should be in the laboratory within 2 hours of production, the test is qualitative and subject to some personal interpretation and there is no EQA scheme. For the urine reducing substances clinical suspicion of galactosaemia is the main reason for requesting, however, it is neither sensitive nor specific and more sensitive test is available for the diagnosis.

    Conclusion: Provision of urine and faecal reducing substances by the laboratories should be reviewed due to lack of analytical and clinical sensitivity and specificity. We suggest training for clinicians on alternative tests available for diagnosing target diseases.

  • P056 Concentration-dependent cut-offs for haemolysis indices

    AJ Chadburn (Department of Clinical Chemistry, The Dudley Group of Hospitals NHS Foundation Trust, Dudley, United Kingdom), P Mohammed (Department of Clinical Chemistry, The Dudley Group of Hospitals NHS Foundation Trust, Dudley, United Kingdom), H Ashby (Department of Clinical Chemistry, The Dudley Group of Hospitals NHS Foundation Trust, Dudley, United Kingdom)

    Haemolysis can interfere with chemistry methods due to colorimetric interference e.g. most assays when haemolysis is severe; or release of an interfering substance from red cells e.g. alanine aminotransferase (ALT) and creatine kinase (CK). Some of the assays affected by low level haemolysis can be potentially urgent and require immediate action from clinicians. However, if even modest haemolysis is present these results are unavailable. The aim of this study was to investigate the possibility of using concentration dependent cut-offs for haemolysis such that grossly elevated, and therefore urgent, results can still be reported.

    Serum pools with elevated ALT (~700, ~1200 and ~2500 IU/L), CK (~1000, ~1500, ~3000 and ~5000 IU/L) and amylase (~500 and ~1200 IU/L) were spiked with haemolysate to produce gross, high, medium and low haemolysis with H indices of ~1800, ~900, ~475 and ~190, respectively. Analysis was carried out on a Roche c702 chemistry analyser.

    Amylase, which suffers moderate interference from haemolysis (H index 500), showed no interference in both pools with low and medium haemolysis but >10% interference at high and gross haemolysis. CK showed a concentration-dependent interference from haemolysis with the highest concentration pool unaffected by haemolysis (<10% interference) even at gross haemolysis but lower concentration pools affected by haemolysis to varying extents. Surprisingly, ALT was unaffected by gross haemolysis in all pools.

    In conclusion, we found that it is possible to use concentration-dependent cut-offs for haemolysis for CK and ALT.

  • P057 Process mapping of emergency department samples arriving in automated biochemistry and root causes of delays in reporting of results

    P Crook (Department of Blood Sciences, Viapath, St Thomas' Hospital, London, United Kingdom), S Brady (Department of Blood Sciences, Viapath, St Thomas' Hospital, London, United Kingdom)

    Achieving a rapid turnaround time (TAT) when processing emergency department (ED) biochemistry samples is important to facilitate prompt clinical decision making, appropriate to the urgency for diagnosis and treatment.

    The target TAT for ED samples is 60 minutes; the laboratory key performance indicator (KPI) states that 80% of samples should meet this TAT. If the TAT is exceeded, there is potential for the patient’s quality of care, health, and healthcare experience to be negatively impacted.

    An audit was conducted to map the processes involved in the sample journey, with an aim to innovate ways in which these could be improved, to contribute to a more efficient service.

    Samples arriving in the biochemistry laboratory via pneumatic tube from ED were monitored for one week during daytime working hours. The journey of each sample from receipt to results release was chronicled. During the study period, 301 biochemistry samples were recorded; of which 29 breached the 60 minute TAT. Deviations in the sample journey were identified, including samples not being loaded immediately post-centrifugation, samples being incorrectly sorted onto the pre-analytics rather than onto the analyser, and barcode read errors. The largest area for improvement was identified to be the risk of delay in sample loading after the point of centrifugation. Thus, a simple change was made by utilising stop clock timers, to give a more audible alert of completion of centrifugation to laboratory staff. Additionally, a flow diagram was displayed in the laboratory illustrating the correct ED sample receipt process as a visual aid, and findings of the audit communicated with all staff involved in the sample journey.

    This audit demonstrates that making small improvements in the pre-analytical processes positively impacts service delivery. Clarity of processes and communication of clinical urgency encourages a patient-first culture and meeting of KPI targets.

Paediatrics and IEM

  • P058 The role of the laboratory in the early identification of hyperammonaemia and the opportunities missed

    H Aitkenhead (Department of Chemical Pathology, Great Ormond Street Hospital for Children, London, United Kingdom)

    Laboratories have a critical role in identifying patients with hyperammonaemia. Urgent recognition and treatment of such patients, especially in the neonatal period, is vital as if left untreated morbidity and mortality is high. Unfortunately blood ammonia is unstable in vitro and ammonia may be mildly raised as a result of delayed separation, haemolysis and contamination. There is a concern that samples that do not comply with stringently applied acceptance criteria are being rejected by laboratories. A questionnaire on the ammonia analysis was piloted on MetBio.Net stakeholder laboratories in order to review the current practice and inform best practice. The 18 laboratories were asked whether they require samples to be unhaemolysed, sent on ice, within a defined time interval and whether they test vacutainers for contamination.

    Results: All 18 laboratories responded. Out of the three laboratories that test for contamination, only one laboratory has found contamination. All but 1 laboratory recommends that samples are sent within a specific time interval; 6 laboratories reject samples that they consider too old for analysis. Thirteen laboratories recommend that samples are sent on ice; 2 laboratories reject samples that they are not received on ice. All laboratories assess the haemolysis status of sample sent for ammonia analysis. Five laboratories reject samples which they consider to be haemolysed whilst 12 laboratories reject samples which they consider to be grossly haemolysed. Thirteen laboratories who reject samples request repeat samples.

    Conclusion: The data shows that a proportion of samples sent to laboratories for ammonia analysis are rejected. As a result, opportunities for identifying patients with significant hyperammonaemia may be missed which may lead to significant morbidity and even death. The MetBio.Net guidelines for the investigation of hyperammonaemia are being revised and will recommend more lenient sample acceptance criteria.

  • P059 An audit of bloodspot phenylalanine monitoring in phenylketonuria patients across South Wales

    J Flatt (Department of Medical Biochemistry and Immunology, University Hospital of Wales, Cardiff, United Kingdom), R George (Department of Medical Biochemistry and Immunology, University Hospital of Wales, Cardiff, United Kingdom), S Moat (Department of Medical Biochemistry and Immunology, University Hospital of Wales, Cardiff, United Kingdom)

    New European guidance for blood phenylalanine monitoring of phenylketonuria (PKU) was published last year (Lancet Diabetes Endocrinol. 2017; 5: 743), replacing previous guidance (NSPKU 2014. ISBN: 0 9535550 0 3). Recommended monitoring frequencies remain the same, but the upper limits of the phenylalanine target ranges have changed for certain age groups. These are more relaxed in 12-18 year olds (600 instead of 480 μmol/L), but more stringent in 6-11 year olds (360 instead of 480 μmol/L) and adults (600 instead of 700 μmol/L). Bloodspot phenylalanine (BSPhe) monitoring was audited against the old and new guidelines to evaluate the impact on the numbers of patients meeting monitoring recommendations.

    The laboratory information system was used to collect BSPhe concentrations and frequency of monitoring data for PKU patients in South Wales over a one year period (Jan-Dec 2016).

    Patients up to one year of age should send weekly monitoring samples, then fortnightly until the age of 12 and monthly thereafter. The adherence to these recommendations is highly variable between individuals: 25% of paediatric patients sent >80% of the expected number of samples, versus only 10% of adult patients. In patients under 12, higher monitoring frequency correlated with better phenylalanine control. The data revealed that the majority of young adult and adolescent females have poor phenylalanine control. The median BSPhe was 790 μmol/L in this population, with a range of 500-1452 μmol/L. This is a concern because of the risk of maternal PKU.

    The only major difference between the guidelines was in 6-11 year olds, where 55% of patients were within the previous target range, compared to 9% using the new guidelines. It was concluded that the new PKU guidelines could be implemented provided that this age group is targeted for improved phenylalanine control.

  • P060 Fabry disease screening: experiences of referrals from a laboratory attached to a tertiary metabolic unit

    E Bate (Department of Clinical Biochemistry, Pathology at Wigan and Salford, Salford Royal NHS Foundation Trust, Salford, United Kingdom), J Borzomato (Department of Clinical Biochemistry, Pathology at Wigan and Salford, Salford Royal NHS Foundation Trust, Salford, United Kingdom), D Darby (Department of Clinical Biochemistry, Pathology at Wigan and Salford, Salford Royal NHS Foundation Trust, Salford, United Kingdom)

    Background: Fabry disease (FD) is an X-linked condition resulting in reduced alpha-galactosidase activity. This causes globotriaosylceramide accumulation resulting in a range of pathology; renal and cardiac failure, angiokeratomas, hypohydrosis and tinnitus. The prevalence of FD is quoted as 1:40,000, however genetic screening studies suggest an FD mutation incidence of 1:1368 in males.

    The Royal College of Physicians ‘National clinical guideline for stroke’ recommends that young stroke patients with signs of FD should be offered FD screening. Similarly, the European Society of Cardiology recommend screening in patients with hypertrophic cardiomyopathy (HCM) over the age of 30 years, as 0.5-1% of patients with HCM aged over 35 years have FD. Suggestions have been made that due to the higher prevalence of FD in patients with left ventricular hypertrophy (LVH), end-stage kidney disease or cryptogenic stroke, screening may be justified.

    Method: A service delivery audit was conducted to understand how Fabry screening at one hospital with a tertiary metabolic unit is being used. Data was gathered on FD requests with results over a 7 month period (03.07.17 to 03.01.18).

    Results: Fifty-four patients had FD screening results; two thirds of the patients were male and the modal age group was 50-59 years. The requests were from neurology (50%), endocrinology/metabolic medicine (26%), renal (17%), cardiology (4%) and dermatology (2%). All 12 positive results for FD were from metabolic medicine; 4 of these patients were identified with FD following cardiology investigation for HCM or LVH, 6 had relatives with FD, 1 was referred from the renal team with chronic kidney disease and 1 patient was diagnosed in a paediatric hospital.

    Conclusion: It is important to identify patients with FD, both for the patients and their relatives. Given enzyme replacement and secondary prevention it is important to ensure that we are detecting patients with FD in high-risk groups.

  • P061 Two cases of severe neonatal primary hyperoxaluria

    N Unsworth (Department of Clinical Biochemistry, Imperial College Healthcare, London, United Kingdom), O Clifford-Mobley (Department of Clinical Biochemistry, University Hospitals Bristol, Bristol, United Kingdom), EL Williams (Department of Clinical Biochemistry, Imperial College Healthcare, London, United Kingdom)

    Background: Primary hyperoxaluria type 1 (PH1) is a rare, inherited disorder of glyoxylate metabolism caused by deficiency of the liver-specific, peroxisomal enzyme alanine:glyoxylate aminotransferase (AGT). Here we detail two unrelated cases of severe infantile PH1 that presented within months of each other at the same centre.

    Case 1: A male infant with a history of failure to thrive and vomiting from 1 month of age presented at 4 months with acute renal failure (creatinine 666 µmol/L), echogenic kidneys, hyponatraemia and severe hypertension. Trio whole exome sequencing revealed a homozygous pathogenic splice site variant in the AGXT gene (c.846+1G>T) which occurred in the index case as a result of maternal uniparental disomy of chromosome 2. This finding was confirmed by Sanger sequencing. Urine oxalate was raised at 635 µmol/mmol Cr (RR<291) as was plasma oxalate at 90 µmol/L (RR<10). Oxalate and glycolate were also present on urine organic acid profiling which was later confirmed by PH metabolite analysis, supporting a diagnosis of PH1.

    Case 2: A male infant presented to A&E at 12 weeks old with haematemesis, renal failure (creatinine 445 µmol/L) and echogenic kidneys. Biochemical testing showed elevated oxalate and glycolate (organic acid profile), hyperoxaluria (531 µmol/mmol Cr) and raised plasma oxalate (171 µmol/L). Sanger sequencing of the whole AGXT gene showed this patient to be a compound heterozygote for two known pathological variants: c.33dupC and c.847-3C>G, thus confirming the diagnosis of PH1.

    Conclusion: These two cases highlight the importance of both genetic and biochemical testing in the diagnosis of rare inherited metabolic disorders. In particular, detection of oxalate and glycolate in urine organic acid profiling should trigger further investigations for primary hyperoxaluria type 1. Case 1 also raises the point that in unusual cases, rare autosomal recessive disorders can be unmasked by chromosomal abnormalities.

  • P062 Development and validation of the Waters Acquity QDa system for the quantitation of plasma amino acids

    YI Leung (Biochemical Sciences, St Thomas' Hospital, London, United Kingdom), R Carling (Biochemical Sciences, St Thomas' Hospital, London, United Kingdom), D Lewis (Biochemical Sciences, St Thomas' Hospital, London, United Kingdom), R Churchus (Biochemical Sciences, St Thomas' Hospital, London, United Kingdom)

    Plasma amino acid analysis has an important role in the investigation and monitoring of inherited metabolic disorders. Traditionally, ion exchange chromatography (IEC) with post column ninhydrin derivatisation and UV detection has been considered the gold standard method of analysis. However, the method lacks specificity and the long analysis time (180 mins per sample) is a limiting factor. Recent advances in liquid chromatography tandem mass spectrometry (LCMSMS) mean this technology is now a viable alternative to IEC and this is reflected in the ERNDIM Quantitative Amino Acid External Quality Assessment (EQA) scheme: in 2007, 2% (4/176) of participating laboratories used LCMSMS, by 2017 this had increased to 24% (62/258).

    Compared with IEC, LCMSMS offers superior specificity and a rapid analysis time (20 minutes per sample). However, significant analytical expertise is required and it is more costly. The Waters Acquity QDa is an LCMS system that could provide a more cost effective solution, whilst still delivering improved specificity and rapid throughput in comparison to IEC.

    We validated plasma amino acid analyses on a Waters Aqcuity QDa. Sample preparation involved addition of stable isotope internal standard mix, protein precipitation and derivitisation with the Waters AccQTag kit. Chromatography was performed on a CORTECS C18 column (2.1X150 mm) with gradient elution and MS detection. Twenty-eight amino acids were quantitated, including sulphocysteine, homocystine and homocitrulline. Chromatographic separation of isoleucine, leucine, and alloisoleucine was achieved. All analytes were linear to 50 µM, LOQs were between 2 and 5 µM and intra and inter batch CVs <10%. Accuracy of the method was assessed by comparison of samples with an established IEC method (n=50 patients, n=12 EQA).

    In conclusion, the QDa enables quantitation of the clinically important plasma amino acids and could provide clinical laboratories with a simple, robust and cost effective alternative to IEC or LCMSMS.

Point of Care Testing

  • P063 Validation of the Roche cobas b221 point of care blood gas analyser for fluid pH analysis

    J Raju (Clinical Chemistry, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), R Williams (Clinical Chemistry, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), J Duffy (Clinical Chemistry, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), C Webster (Clinical Chemistry, Heart of England NHS Foundation Trust, Birmingham, United Kingdom)

    The measurement of pleural fluid (PF) pH has diagnostic, therapeutic and prognostic implications in exudative pleural effusions. PF pH is especially valuable in the management of malignant and parapneumonic effusions, which displays a spectrum of disease severity ranging from free-flowing effusions to those which require surgical drainage. The most accurate method for measuring PF pH is by blood gas analyser under anaerobic conditions, with point of care testing (POCT) providing faster analysis. Previously, at Heart of England NHS Trust, PF pH analysis was performed using the Radiometer ABL 800 analyser. Here we describe the validation of the Roche b221 analyser against the previous method. The use of the Roche POCT device provides a faster, more robust analysis, with greater accuracy and precision.

    Anonymised patient samples were used to calculate the total analytical imprecision and 95% predictive uncertainty for five different pH values; 7.4, 7.2, 7.0, 6.8 and 6.7, alongside three levels of IQC (7.2, 7.4, and 7.5). All results were below the target value of ±1%. Accuracy was assessed by analysing 12 EQA samples against their reported target values. Data analysis was performed using a Bland-Altman plot, which calculated an undetectable bias of -0.0% (target of 1%). Sample stability was assessed using capped and uncapped syringes, with and without air inside. The sample pH was measured at the following time points: 0, 15, 30, 60, 90, 120 and 240 minutes. Capped and uncapped, without air revealed no clinically significant change in results between 0-240 minutes (±0.1pH), whereas sample with air revealed a clinically significant change after 15 minutes. Interference studies revealed that 200 µL of 2% lidocaine resulted in a significant change in pH (p<0.05), whereas this was not observed with lithium heparin.

    These results demonstrate that the Roche b221 analyser is fit for purpose when analysing pleural fluid pH.

  • P064 Validation of the Roche cobas b101 point of care system for glycated haemoglobin testing

    J Raju (Department of Clinical Chemistry, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), R Williams (Department of Clinical Chemistry, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), J Duffy (Department of Clinical Chemistry, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), C Webster (Department of Clinical Chemistry, Heart of England NHS Foundation Trust, Birmingham, United Kingdom)

    Near patient glycated haemoglobin (HbA1c) analysis in outpatients and the community is important to monitor the glycaemic control of diabetic patients and to aid diagnosis of diabetes mellitus. HbA1c analysis using POC system helps in quick analysis, real-time reporting and immediate decision making potentially reducing follow up appointments. It also offers convenience to patients, single finger prick test rather than blood test requiring larger volume sample, phlebotomists and transportation to laboratory.

    Previously, at Heart of England NHS Trust, POC HbA1c testing was performed using the Siemens DCA Vantage HbA1c analysers. Here we describe the validation of nine Roche cobas b101 analysers against the previous method. The use of the Roche POCT device provides a simple, faster and more robust HbA1c testing.

    Internal quality control materials were measured on all nine of the b101 devices for imprecision test. A coefficient of variation (%CV) was calculated for the within-batch measurements (CVW) and for the between-batch measurements (CVB). These values were used to calculate the total analytical imprecision (CVA) and a 95% predictive uncertainty. All results were within the target level of ±3% (NGSP) and WEQAS desirable ±4.8% and minimum ±6.7%.

    Accuracy was assessed by analysing anonymised patient samples. The primary b101 HbA1c method was compared to the previous DCA and the laboratory Tosoh G8 methods, with a Bland Altman and Passing Bablok fit statistical test performed for each method comparison. Differences in means were expressed as a percentage bias. Data analysis calculated a bias of -2.9% for b101 vs DCA and -0.5% for b101 vs Tosoh G8 (target of 6%). Other eight b101 analysers were then compared with primary b101 analyser. Results were within the acceptable limit of ±6%. These results demonstrate that the b101 analyser is fit for purpose when analysing HbA1c in outpatient and community settings.

  • P065 Measuring whole blood cortisol using venous and finger prick sampling on the i-CHROMA™ cortisol method: a proof of concept study

    J Bolodeoku (Research Laboratory, JB Consulting (MDP) Limited, Oxfordshire, United Kingdom), J Honour (University College London Hospitals, London, United Kingdom), S Bethancourt-Watson (University College London Hospitals, London, United Kingdom), K Geertsma (CAH Is Us Support Group, London, United Kingdom), P Hindmarsh (University College London Hospitals, London, United Kingdom)

    Blood cortisol estimation is extremely useful in diagnosis and management of adrenal insufficiency. A number of methods for cortisol estimations have been developed using blood and saliva samples, these measurements can be quite challenging.

    The purpose of this study was to compare 24 hour cortisol profile measurements using venous whole blood and finger prick (capillary) samples on an immunoassay (POCT) cortisol system.

    A patient was admitted and had two 24 hour cortisol profile measurements in July and October 2017. Whole blood venous samples and corresponding finger prick samples were taken over 24 hours at hourly intervals from 10:00h to 10.00h. Cortisol measurements were made using the i-CHROMA™ Cortisol POCT method.

    The July 24 hour cortisol profile revealed a circadian pattern with 216.36 nmol/lL(venous) and 230.69 nmol/L (capillary) at 10:00 then was undetectable (<80 nmol/L) in both samples at 20:00 and rose again to 243.89 nmol/L (venous) and 240.36 nmol/L (capillary) at 10:00. There was very good correlation (r=0.91) and no difference between the venous and capillary blood samples (n=21).

    The October 24 hour cortisol profile revealed a circadian pattern starting with 246 nmol/L (venous) and 257.3 nmol/L (capillary) at 10:00 then was undetectable (<80 nmol/L) in both samples at 20:00 and rose to 304.31 nmol/L (venous) and 264.13 nmol/L (capillary) at 08.00. There was very good correlation (r=0.90) and no difference between the venous and capillary blood samples (n=23). Mean venous cortisol was 204.5 (SD 53.8) nmol/L and capillary 209.9 (SD 57.0) nmol/L. Slight differences compared to July profile were due to medication changes. Bland Altman plot showed a bias of -5.4 nmol/L with no effect of concentration on the bias observed.

    The i-CHROMA™ Cortisol method was able to detect the patient’s circadian rhythm using both venous and capillary blood samples.

Quality Assurance

  • P066 The implementation of daily Internal quality control review booklets in a clinical biochemistry laboratory

    RD Williams (Department of Clinical Biochemistry, Birmingham Heartlands Hospital, Birmingham, United Kingdom), EL Evans (Department of Clinical Biochemistry, Birmingham Heartlands Hospital, Birmingham, United Kingdom), C Webster (Department of Clinical Biochemistry, Birmingham Heartlands Hospital, Birmingham, United Kingdom)

    Internal quality control (IQC) is an integral part of any pathology laboratory, providing real-time assessment of assay precision, in order to maintain compliance with established Westgard rules. A poor approach to the implementation, record keeping and review of IQC data can lead to a range of issues, including the release of incorrect patient results, unnecessary investigations into rejected analytical runs and incorrect patient management.

    The automated chemistry department at Heart of England Trust ran in excess of 10 million tests for the year 2016-17, and facilitates over 100 different analytical tests. Previously within our laboratory, daily IQC data was recorded on simple box-ticking sheets, with separate documents used to record IQC failures and follow-up procedures. This presented with problems of duplicate and incoherent records, staff confusion, wasteful printing and poor management of IQC issues within the laboratory. Following a UKAS finding, the automated chemistry laboratory implemented the use of a daily IQC review booklet. These were to allow consolidation of monthly IQC data, current or resolved issues and relevant comments for each analyser within the laboratory. Furthermore, within each booklet, a simple step-by-step process into the investigation of IQC failure was also included with an aim to help biomedical scientists when troubleshooting.

    In order to fully benefit from the use and compliance of the IQC booklets, staff members from across the chemistry department were consulted for their opinions on what structure and format would work best. An audit was performed post-implementation to assess their effectiveness within the laboratory. This revealed initial teething problems of missing staff signatures when recording an IQC issue, although this was rectified through further training. However, the primary issues of duplicate records, staff confusion and wasteful printing were all resolved.

  • P067 The HILs are alive with the sound of innovation

    R Marrington (Birmingham Quality - UK NEQAS, University Hospitals Birmingham, Birmingham, United Kingdom), F MacKenzie (Birmingham Quality - UK NEQAS, University Hospitals Birmingham, Birmingham, United Kingdom), J French (Birmingham Quality - UK NEQAS, University Hospitals Birmingham, Birmingham, United Kingdom)

    The presence of haemoglobin (haemolysis), bilirubin (icterus) and lipids (lipaemia) in serum can affect the reporting of many laboratory test results, not only in the validity of the results themselves but also even whether a result should be reported. This can impact on whether a patient may need to be re-bled or not.

    Birmingham Quality has established the UK NEQAS Serum Indices (HIL) EQA scheme which not only looks at haemolysis, icterus and lipaemia as individual analytes but also systematically looks at the impact these serum indices have on a particular specific analyte – analyte X – which changes from month to month. Laboratories are asked whether they would report the result for the measured analyte based on their serum indices. Laboratories are also asked to provide information on the cut-offs that they use for determining whether or not a result would be reported for that particular analyte.

    The results have shown, for a number of analytes, that there are significant differences in practice for the interpretation of serum indices both within-, and between-manufacturers. For example, three specimens were distributed with varying degrees of haemolysis for the analysis of total protein. The specimen which was grossly haemolysed would have had its total protein result reported by 57% of participants, but not by the other 43%. This was not wholly due to between manufacturer differences and Birmingham Quality used graphical representation to demonstrate these differences to all participants. Haemolysis did adversely affect the total protein result for one manufacturer but there were many participants using this method that did report the result when most other laboratories/systems would not have. There are clinical consequences since total protein may be used in the calculation of globulins or calculations of amount of monoclonal proteins which may trigger subsequent additional unnecessary testing.

  • P068 EQA sample handling via NPEx for Birmingham Quality UK NEQAS services

    F MacKenzie (Birmingham Quality / UK NEQAS, University Hospitals Birmingham NHS Foundation Trust, Birmingham, United Kingdom), A Robins (Birmingham Quality / UK NEQAS, University Hospitals Birmingham NHS Foundation Trust, Birmingham, United Kingdom), S Box (X-Lab Ltd, Leeds, United Kingdom), A Marsay (X-Lab Ltd, Leeds, United Kingdom)

    Birmingham Quality (BQ) now routinely makes requests and receives results for their UK NEQAS Services using X-Lab’s lab-to-lab communications vehicle, NPEx.

    This new innovative time saving approach removes the drudgery of the sample handling and the booking-in process.

    Working together, BQ and NPEx developed an approach that would simplify the booking-in and reporting elements of the EQA process.

    Although BQ had successfully requested and received results for their EQA schemes directly from laboratory LIMS via dial-up modem links in the mid 1990s the use of firewalls effectively ended this approach. The NPEx model of lab-to-lab communication, which has been running for 10 years and which has 50 sites with multiple users, addresses and resolves the security issues.

    Two pilot sites, Birmingham Heartlands Hospital and Manchester Royal Infirmary were enrolled and with their much appreciated enthusiasm and effort, the system went live on 4 September 2017.

    There is less scope for error and when your UK NEQAS samples arrive they are already booked-in and with a few swipes the BQ bar coded results document/manifest are read. The process itself is simple and intuitive but relies on a sophisticated message handling and the encryption at the BQ end which feeds into the established NPEx model. Both parties had to write bespoke software to enable the system to work as successfully as it does.

    It has been calculated that a large DGH/teaching hospital with multiple analysers/UK NEQAS registrations might save up to one man-year in time across the booking in process and in the transcription and data entry process for all the schemes it belongs to. Nationally this has the potential to save the NHS thousands of pounds per annum.

    This approach may well become the norm in future years.

  • P069 Establishing a sample exchange scheme for inter-laboratory comparison of results for plasma biotinidase

    V Powers (Department of Clinical Biochemistry, University Hospitals Bristol NHS Foundation Trust, Bristol, United Kingdom), A Bowron (Department of Blood Sciences, Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle, United Kingdom), J McNeilly (Biochemistry Department, Queen Elizabeth University Hospital, Glasgow, United Kingdom)

    Background: Participation in inter-laboratory comparisons of examination results is included in the UKAS ISO15189:2012 medical laboratories requirements for quality and competence. For most assays, external quality assurance (EQA) schemes fulfil this requirement. However in the absence of an EQA scheme other approaches are required to provide objective evidence that assays are fit for purpose.

    Plasma biotinidase is an assay used to diagnose the inherited metabolic disorder biotinidase deficiency. There is no UK or European EQA scheme for plasma biotinidase.

    Aim: To establish a pilot sample exchange scheme to assess biotinidase assay performance and to fulfil requirements of ISO18189:2012

    Methods: Three specialist metabolic laboratories agreed to take turns to circulate 8-10 samples with a range of biotinidase activities plus their own IQC material over a 12 month period. Samples were transported frozen on dry ice. Enzyme analysis was performed using an in-house fluorimetric rate assay using biotinyl-6-aminoquinoline as substrate. Results were returned to the laboratory sending the samples with reference ranges and interpretive comments. Results were collated and returned to each laboratory.

    Results: So far three sample exchanges have taken place. Plasma biotinidase activities reported ranged from 0.2 to 7.5 nmol/min/mL. There was mostly good agreement between the three laboratories. Importantly interpretation of the result in the context of the laboratories own reference range was consistent. All three laboratories highlighted the requirements to confirm low results by repeat analysis on a fresh sample to exclude sample degradation as a cause.

    Each of the three laboratories has undergone UKAS inspection and the pilot scheme has been accepted by UKAS as meeting the requirements for inter-laboratory comparison of examination results.

    Conclusion: A pilot sample exchange scheme for biotinidase analysis has been established between three laboratories. Method comparison is good, providing validation of the in-house assays and meeting the requirements of UKAS.

  • P070 Development of a three site ALP isoenzyme sample exchange scheme

    E Middling (Department of Blood Sciences, Royal Victoria Infirmary, Newcastle-upon-Tyne, United Kingdom), I Jabbar (Department of Biochemistry, Royal Hallamshire Hospital, Sheffield, United Kingdom), E Hanon (Department of Blood Sciences, Royal Victoria Infirmary, Newcastle-upon-Tyne, United Kingdom), K Smith (Department of Clinical Biochemistry, Queen Elizabeth Hospital, Birmingham, United Kingdom)

    ALP isoenzyme electrophoresis is a qualitative assay for which there is currently no commercial EQA scheme available. Comparison with other laboratories is a crucial requirement to demonstrate quality for ISO 15189 accreditation, therefore a sample exchange scheme was established between 3 laboratories.

    Two samples are distributed on a bi-monthly basis, with responsibility for distributing samples and collating reports rotated between participating laboratories. The results are reported according to each laboratories own protocols, and the comments compared to look for consensus or to highlight areas of discrepancy.

    Eight distributions have been completed (16 samples); 14 of these samples had a raised total ALP, with the predominant isoforms being reported as bone in 5 samples, intestinal in 3 samples and liver and placental in 2 samples each; 1 sample had a variant bone isoform present, and 1 sample was from a patient with benign transient hyperphosphatasaemia.

    Generally good agreement was seen between the laboratories in the reporting of liver and bone isoforms. The bone variant, placental and benign transient hyperphosphatasaemia samples gave more discrepancy in the way the samples were reported and proved the most useful for discussion and educational purposes. One limitation of the scheme however is the available sample volume for repeat or further confirmatory analysis, which may have increased the consistency in reporting of these unusual samples. Greater discrepancy in interpretation was also observed when the samples had a normal ALP, however only 1 laboratory would have routinely analysed these samples.

    In conclusion, a qualitative sample exchange scheme has been established for ALP isoenzyme analysis between three laboratories, which has assisted in the laboratories obtaining ISO 15189 accreditation. A 4th laboratory has recently joined, and we are currently undergoing discussions on further expansion of the scheme, however sample volume requirements will be the key limiting factor.

  • P071 Evaluation of interference caused by lipaemia, icterus and haemolysis on assays performed on Beckman Coulter AU5800 and AU680 platforms

    L Kelly (Department of Clinical Biochemistry, Betsi Cadwaladr University Health Board, Wrexham, United Kingdom), YP Teoh (Department of Clinical Biochemistry, Betsi Cadwaladr University Health Board, Wrexham, United Kingdom)

    Background: Many commonly used clinical biochemistry assays suffer from interference from lipaemia, icterus and haemolysis (LIH). Traditionally samples with observable LIH were identified by visual inspection. In the era of high throughput automated systems, this is not possible. LIH assessment has been incorporated into automated analysers to identify and (semi)-quantitate LIH presence in samples. The cut-off values for determining when LIH cause ‘significant’ interference is usually determined by the manufacturers with little in-house verification. In addition, there is currently no widely accepted criteria to define acceptable levels of interference and those used by manufacturers can appear arbitrary.

    Methods: The level of interference caused by LIH was assessed on a total of 33 chemistry assays performed on Beckman AU5800 and AU680 platforms using pooled serum and plasma spiked with intralipid, bilirubin or haemolysate to simulate LIH interference. We also examine the number of samples likely to be affected if an acceptable interference criteria based on biological variation was implemented in our laboratory.

    Results: The majority of the assays tested were in agreement with the manufacturers stated interference levels. However, a number of these assays showed significantly less interference than stated, raising the possibility that the stated interference maybe too conservative. Several assays showed greater than stated interference for example, alanine aminotransferase (ALT), cholesterol and creatinine kinase (CK) showed greater interference from haemolysis than stated by the manufacturer and gamma-glutamyl transferase (GGT) was more affected by icterus. Implementing interference limits based on biological variation would cause very tight acceptable limits form some analytes such as sodium (0.3%) and wide limits for other such as C-reactive protein (34%).

    Conclusion: Our data was in agreement with that stated by the manufacturer for many but not all of the assays tested.

  • P072 Investigation of excessive haemolysis in emergency department samples

    P Mcbride (Clinical Biochemistry, Antrim Hospital, Northern Trust, Antrim, United Kingdom)

    The aim of the study was to determine the cause of excessive rates of in vitro haemolysis in blood samples from emergency department (ED).

    During July to September 2017, haemolysis rate in ED blood samples was observed to increase from baseline 4.8% to 9.7%. Haemolysis rates remained constant from inpatient wards at 2.5%. Investigations were undertaken to identify factors contributing to increased ED haemolysis. 1) Phlebotomy procedure and staff training was reviewed by the ED; rates of haemolysis did not improve. 2) The effect of two blood collection tubes, lithium heparin (LiH) and serum separator tube (SST), on haemolysis rates was assessed during two consecutive periods of alternate use; in successive 2 week periods haemolysis rate was identical at 8.1% for both LiH and SST. 3) Pre analytical conditions were examined; all samples were kept at room temperature in the ED before dispatch to the laboratory. Haemolysis rates were found to be independent of time to centrifugation. 4) The effect of the pneumatic tube system (PTS) was initially dismissed because haemolysis rate for samples from inpatient wards, travelling similar distances via PTS, showed no increase. After other factors were excluded, the specific ED PTS pathway was investigated. Patient samples (n=9) were collected in duplicate, one sample was sent via PTS and a second, paired sample was hand delivered to the laboratory. PTS mean haemolysis was 0.53 g/L (0.15-1.91) compared to 0.03 g/L (0.0-0.09) for hand delivered samples. Further investigation by the PTS supplier revealed a replacement component resulted in rapid acceleration and deceleration of PTS pods, localised to the ED PTS pathway. When this was modified, reducing transport velocity, haemolysis rates returned to normal.

    Conclusion: This case shows that changes in pneumatic tube system performance may alter rates of in vitro haemolysis and the PTS should undergo reverification.

  • P073 Electronic management of referral laboratories information

    J Duffy (Department of Clinical Chemistry, Heart of England NHS Foundation Trust, Birmingham, United Kingdom), C Webster (Department of Clinical Chemistry, Heart of England NHS Foundation Trust, Birmingham, United Kingdom)

    Introduction: A major resource issue for pathology laboratories is managing information on their referral laboratories. It is a UKAS ISO 15189 requirement (standard 4.5) that laboratories should have a documented procedure for selecting and evaluating referral laboratories. A procedure for regularly monitoring the quality of performance of their referral laboratories should also be in place. Historically, at the Heart of England NHS Trust (HEFT), this has been managed on Q-pulse using Word and Excel documents. This system is time consuming for both our laboratory and the referral labs involved, inefficient, subject to error and hard to maintain in real time.

    Aim: To develop an electronic application for managing referral laboratory information, which will reduce time spent maintaining this information and ultimately produce a solution that can be extended for use by other laboratories.

    Methods/Specifications: A Windows-based application was developed internally at HEFT to meet the following specifications: 1) a clear and easily editable database with full audit trail; 2) a colour coded, sortable table of laboratories and tests to enable direct visualisation of referral labs and the status of their information; 3) automated generation of letters for monitoring quality of performance with auto-population of existing data held on the laboratories; 4) two access levels, one for staff involved in sending samples to view information and another for administration purposes; 5) electronic storage for returned information; 6) ability to record acceptance of information.

    Results and conclusion: A referral lab application has been produced that is in routine use at HEFT. Audit data shows the application has significantly reduced workload associated with collecting and maintaining information on referral labs. This application has potential for use on a wider national scale and could look to compete with assayfinder.com in the future.

Toxicology TDM

  • P074 Utility of carbamazepine-10,11-epoxide

    C Harborow (Department of Clinical Biochemistry and Metabolic Medicine, Liverpool Clinical Laboratories, Liverpool, United Kingdom), K Birch (The Neuroscience Laboratories, The Walton Centre NHS Foundation Trust, Liverpool, United Kingdom), C Chadwick (The Neuroscience Laboratories, The Walton Centre NHS Foundation Trust, Liverpool, United Kingdom)

    Introduction: Carbamazepine (CBZ) is used primarily to treat epilepsy and neuropathic pain; serum levels are used to monitor therapy or assess toxicity. The therapeutic range is considered to be 4-12 mg/L as determined by the Pathology Harmonisation Group. Serum CBZ does not take into account the major active metabolite carbamazepine-10,11-epoxide (CBZ-E) which shares roughly equal anticonvulsant and neurotoxic effects. In the Neuroscience Laboratories, results for CBZ-E concentration are generated alongside every CBZ result but only reported if requested. This investigation seeks to examine the utility of CBZ-E results in patients on CBZ therapy.

    Methods: Samples analysed for anti-epileptic drugs (AEDs) by LC-MS/MS between February-October 2017 were examined for the presence of CBZ and CBZ-E (n=94). Presence of other AEDs including lamotrigine, phenobarbital, phenytoin and valproate was also recorded.

    Results: Data was split into 5 groups depending on presence of other AEDs: CBZ only; CBZ with lamotrigine; phenobarbital; phenytoin; valproate. Samples with more than two AEDs present were excluded (n=4). Several patients had a CBZ result within the therapeutic range but the additive effect of raised CBZ-E could potentially have caused symptoms of toxicity. CBZ-E/CBZ ratio was calculated to examine the effect of polytherapy on CBZ-E metabolism. 93% of samples had a CBZ-E/CBZ ratio under 0.30; 6 out of the 7 samples with a ratio greater than 0.30 were from patients receiving polytherapy, thus highlighting the potential role for CBZ-E in monitoring treatment.

    Conclusion: If clinical effects are in excess of what would normally be expected at a given CBZ concentration, CBZ-E quantification may be valuable. CBZ-E quantification may be useful in patients where CBZ is used in polytherapy alongside other AEDs which may cause increased CBZ-E. Awareness of an active metabolite is important and so reporting of CBZ-E results alongside CBZ may be useful for clinicians prescribing this drug.

  • P075 Addressing prescription medicine misuse with new screening services: early evidence suggests significant abuse will be detected

    O Roberts (Derriford Combined Laboratory, Plymouth Hospitals NHS Trust, Plymouth, United Kingdom), R Griffiths (Department of Clinical Biochemistry & Toxicology, Sandwell and West Birmingham Hospital NHS Trust, Birmingham, United Kingdom), L Ford (Department of Clinical Biochemistry & Toxicology, Sandwell and West Birmingham Hospital NHS Trust, Birmingham, United Kingdom), J Berg (Department of Clinical Biochemistry & Toxicology, Sandwell and West Birmingham Hospital NHS Trust, Birmingham, United Kingdom)

    Introduction: There is growing awareness and concern at the abuse of prescription medicines. We have recently introduced a new LC-MS/MS drugs of abuse screening panel, including prescription medication identified in a 2016 report from the Advisory Council for the Misuse of Drugs (ACMD). Medication includes pregabalin, gabapentin, mirtazapine, quetiapine, promethazine, clonazepam, alprazolam, zopiclone and fentanyl. In an attempt to record the prevalence of diversion, we look at the number of undeclared prescription medications identified in urine samples as part of a routine medico-legal drugs of abuse screening service.

    Methods: Data was collected between 4th December 2017 and 21st January 2018. Results of the screens were recorded and compared to the accompanying request form declaring any medications taken in the last 30 days, whether it be prescription or over-the-counter (OTC).

    Results: A total of 693 medico-legal samples were received during the audit period; 56 (8.1%) samples tested positive for one or more of the new panel prescription medications, with 143 positive results reported in total.

    Of these, 111 had been declared (77.6%) and 32 had not been declared (22.4%). The most common medication identified was mirtazapine (46 positive), followed by quetiapine (23) and pregabalin (22). The three medications most likely to be undeclared were promethazine (71.4% not declared), gabapentin (33.3%) and quetiapine (17.4%).

    Discussion: The prevalence of undeclared prescription medication amongst this cohort is surprisingly high. This may be due to high levels of polypharmacy, lack of awareness of active ingredients in OTC medication such as Night Nurse®, which contains promethazine, or true diversion of prescription medications. Medications with neurological effects such as sedation or anti-depressant/psychotic effects are the most likely to be undeclared amongst this cohort. This data may provide a good starting point to thoroughly investigate the prevalence of diversion of prescription medication in the UK.

  • P076 What is killing drug users in Scotland? Specific drug analysis by tandem mass spectrometry is the way forward

    P Cawood (Clinical Biochemistry, Royal Infirmary of Edinburgh, Edinburgh, United Kingdom), JA McCauley (Clinical Biochemistry, Royal Infirmary of Edinburgh, Edinburgh, United Kingdom), PMJ Gracie (Clinical Biochemistry, Royal Infirmary of Edinburgh, Edinburgh, United Kingdom)

    Scotland leads Europe in drug-related-deaths (DRD) and 2016 saw a large increase in DRD compared to previous years. Increasing use of gabapentinoids (22% of DRD) and non-prescription benzodiazepines (27% of DRD mainly etizolam and diclazepam) were the main changes. Increasing use of alprazolam (Xanax) use has been reported recently. Current immunoassay screening methods in urine and oral fluid are unable to detect these.

    We report on a tandem mass spectrometry method that includes 24 drugs, measured in urine and oral fluid. The method requires 50 µL, takes 5.2 minutes and is accredited to ISO15189.

    We report on the findings of 11,049 oral fluid samples collected from substance misuse clients between January 2016 and June 2017.

    Gabapentinoids were present in 21.2% of samples; 92% of these were also positive for methadone (80%) or buprenorphine (12%). Gabapentinoids are taken to enhance the intoxicating effects of methadone. As gabapentinoids do not bind to the opioid receptor the mechanism is likely to be by increasing the level of methadone. This may account for some of the increase in DRDs in Scotland.

    Other findings were 6-mono-acetyl-morphine (6MAM) in 39% of samples, cocaine (24%), methadone (72%), buprenorphine (14%), dihydrocodeine (9%), codeine only (2.3%), oxycodone (0.3%), amphetamine (3%), tramadol (3%), MDMA (0.7%), prescription benzodiazepines (47%). Methamphetamine was found in only one of the 11,049 samples.

    Measuring parent drug and metabolites in urine facilitates the detection of samples spiked with methadone, buprenorphine or diazepam. Oral fluid has the advantage of witnessed collection. 6MAM and cocaine in oral fluid have the advantage of a longer duration of detectability than urine. Measuring gabapentin and pregabalin identifies non-compliance with prescribed drugs and identifies those who source these on the streets.

    Etizolam, diclazepam, delorazepam and alprazolam are now included in the Lothian drug panel for future assessment of the use of these drugs.

  • P077 Validation of Waters Xevo G2-XS QTof high resolution accurate mass-mass spectrometer for toxicological screening in post-mortem femoral blood

    L Hikin (Department of Chemical Pathology and Metabolic Medicine, University Hospitals Leicester, Leicester, United Kingdom), R Oakes (Department of Chemical Pathology and Metabolic Medicine, University Hospitals Leicester, Leicester, United Kingdom), D Lane (Department of Chemical Pathology and Metabolic Medicine, University Hospitals Leicester, Leicester, United Kingdom), J Dave (Department of Chemical Pathology and Metabolic Medicine, University Hospitals Leicester, Leicester, United Kingdom), A Burns (Department of Chemical Pathology and Metabolic Medicine, University Hospitals Leicester, Leicester, United Kingdom), S Morley (Department of Chemical Pathology and Metabolic Medicine, University Hospitals Leicester, Leicester, United Kingdom), P Smith (Department of Chemical Pathology and Metabolic Medicine, University Hospitals Leicester, Leicester, United Kingdom)

    The University Hospitals Leicester (UHL) performs toxicological screening in post-mortem femoral blood for ~25% of UK coroners. Our previous method employed LC-MS/MS with MRM and detected ~400 drugs and metabolites. Introduction of the high resolution accurate mass-mass spectrometer (HRAM-MS) enabled screening for ~1200 drugs and metabolites, with the added ability to detect new/unknown compounds. Validation was required, in accordance with Scientific Working Group for Forensic Toxicology (SWGTOX) recommendations.

    Reference materials of drugs at high/lethal concentrations (n=42) were injected followed by water blanks to check for drug interferences/carryover. To investigate matrix interference, blank samples (n=10) were extracted and analysed in duplicate. Lower limits of detection (LLOD) were established for 40 frequently encountered drugs; three separate blank whole blood matrices were spiked with drugs at concentrations close to their anticipated LLOD and analysed in duplicate for three days. LLOD was defined as the lowest concentration with: a reproducible response, ≥3x signal: noise, retention time (±0.3 min), acceptable fragment match (observed ≥50% expected), mass error (±10 ppm) and acceptable peak shape. A comparison of post mortem samples (n=138) and external quality assurance (EQA) was performed. On-board sample extract stability was determined by re-injecting extracts stored on-board for 3 days.

    The HRAM-MS had equal or superior sensitivity compared to our existing method; for the majority of drugs (n=26) the LLOD was ≤1 ng/mL or ≤0.1 mg/L. Comparison of post-mortem and EQA samples gave acceptable results with excellent agreement between methods. Forensically significant drug interference was not observed. Carryover was apparent owing to instrument sensitivity; wash steps were adjusted accordingly. Illicit drugs were not detected in any blank blood matrices analysed.

    Transition from our current method to HRAM-MS offered significant quality improvements, enabling comprehensive screening of a much larger repertoire of drugs, reduced labour-requirements, enhanced sensitivity and reduction in the sample volume required.

  • P078 Clinical Toxic Metal Screen: what do we find? A review of 3 years' data

    NL Barlow (Department of Clinical Biochemistry, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), A Ahmed (Department of Clinical Biochemistry, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), J Murray (Department of Clinical Biochemistry, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), JD Berg (Department of Clinical Biochemistry, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom)

    The clinical Toxic Metal Screen (TMS) offered by the trace elements laboratory at SWBH NHS Trust is a semi-quantitative inductively-coupled plasma mass spectrometry assay used to screen for 52 elements simultaneously using a single aliquot of patient blood/urine. It is usually requested when no single element has been implicated in toxicity. The TMS is a useful indicator of any significantly elevated metal concentrations but as it does not generate accurate, reportable concentrations, these must be confirmed by subsequent quantitative analysis.

    Indications for testing and TMS findings were reviewed over a three year period (01/04/2014 – 31/03/2017). In total there were 66 TMS requests for 51 patients from 31 locations. Neuropathy (19.6%), encephalopathy (11.8%) and ingestion of a known harmful entity (5.9%) were amongst the common indications for screening. For 47.1% of patients the indication for screening was not provided. Toxicologically significant findings were reported in 7 (10.6%) screens: the elements involved were lead, copper, zinc, selenium, silver, chromium, cadmium, nickel, tin and barium. No toxicologically significant findings were reported in 56 (84.8%) screens. Over 40% of the latter were found to be contaminated with elements including nickel, chromium, manganese and zinc. Raised gadolinium, barium and iodine levels observed were most likely attributable to recent imaging procedures. The most commonly elevated element detected was arsenic, which is often attributable to dietary seafood intake containing non-toxic arsenobetaine. Quantitative analyses were undertaken for 27 patients and subsequent follow-up testing in 4 patients.

    The TMS is an essential tool for the exclusion of heavy metal toxicity as a cause of presenting signs and symptoms. This audit highlights the complexity of result interpretation with non-toxicologically significant elevations being common as a result of sample contamination or dietary factors. Provision of relevant clinical and exposure history is vital for meaningful assessment of the results.

  • P079 Bridging the gap in quality assurance for UKAS accreditation in clinical toxicology

    J Pethick (Department of Clinical Biochemistry and Toxicology, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), L Ford (Department of Clinical Biochemistry and Toxicology, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), J Berg (Department of Clinical Biochemistry and Toxicology, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom)

    Quality assurance is an essential requirement for the operation of a clinical laboratory (ISO:15189 clause 5.6.3.2). Participation in a recognised proficiency testing scheme or sample sharing is an affordable and reliable way to achieve this. However, participation in such schemes can be problematic for state-of-the-art in-house toxicology tests, for example screening for drugs of abuse by UPLC-MS/MS. Of note, the frequency with which drugs are positively identified in a proficiency testing scheme are often not sufficient to meet the requirements of the laboratory, for example compounds may only appear once in a two year cycle. The use of Time of Flight (TOF) for unknown toxicology screening poses another conundrum as many schemes are not able to regularly test a library of 1300 drugs detectable by this method in suitable sample types. In order to meet the requirement of UKAS accreditation for these tests we developed a solution using in-house proficiency testing.

    Methods: Using the excess urine or oral fluid of patient samples previously analysed by our laboratory a bank of proficiency testing samples was created. Samples were selected based on the drugs that were reported and cover the range of drugs in our 32 drugs of abuse panel and compounds that are not included in any third party proficiency testing scheme, but detectable by ToF. Samples are selected once monthly, analysed and compared to previously reported results. A report is generated allowing comparison between different analysers and competent staff in terms of sample extraction and interpretation.

    Results: Proficiency testing in this way has demonstrated excellent consistency in analysis and interpretation of LC-MS/MS and ToF results between staff and analysers and meets the requirements for quality assurance.

    Conclusion: We have increased the frequency with which drugs are tested for quality purposes and extended the range of drugs tested for ToF.

  • P080 Validation of a method to screen for drugs in blood and urine using liquid chromatography quadrupole time-of-flight mass spectrometry

    A Lawson (Department of Clinical Chemistry, Immunology and Toxicology, Heart of England Foundation Trust, Birmingham, United Kingdom), S George (Department of Clinical Chemistry, Immunology and Toxicology, Heart of England Foundation Trust, Birmingham, United Kingdom), B Johnson (Department of Clinical Chemistry, Immunology and Toxicology, Heart of England Foundation Trust, Birmingham, United Kingdom), D Vincente (Department of Clinical Chemistry, Immunology and Toxicology, Heart of England Foundation Trust, Birmingham, United Kingdom)

    Introduction: The rise of novel psychoactive substance use over the last 10 years has meant that toxicology laboratories are asked to detect an ever changing catalogue of substances using highly sensitive techniques. High resolution mass spectrometry is seen as the solution to this problem and has been implemented in a number of laboratories, replacing more traditional screening technologies. Here we present the validation of a liquid chromatography quadrupole time-of-flight mass spectrometry (LC-QT of MS) method to screen for drugs in blood and urine.

    Methods: Blank blood or urine was spiked with 240 over-the-counter, prescription and illicit drugs at a number of concentrations. Urine samples (200 µL) were extracted by solid phase extraction and blood samples (50 µL) extracted by protein/phospholipid removal. Extracts were analysed using a Waters G2-XS system running UNIFI in both positive and negative ionisation modes. Drugs were identified by retention time, mass and characteristic fragment ions when compared to a library. To validate the use of the method, the existing screening approach in use at Heartlands Hospital was compared to the LC-QTof based methodology by analysis of 70 blood and 42 urine samples from 70 cases.

    Results: The vast majority of drugs (95%) screened were detectable at levels of 10 µg/L in blood and urine with intensities recorded suggesting that many would be detected at much lower levels. The LC-QTof approach detected 231 more drugs than the current screen and did not miss any drugs previously detected. No carry over was observed when a blood sample containing 10 mg/L of clozapine, cocaine, codeine, diazepam and tramadol was extracted and analysed.

    Conclusion: The screening protocol outlined here allows for rapid, sensitive and untargeted analysis of a broad range of compounds in blood and urine. Implementation will allow for a simplification of the screening protocol in use, leading to faster turnaround times and improved patient care.

  • P081 An audit into infliximab requests: could we improve our demand management?

    R Bray (Biochemistry, Northwick Park Hospital, London, United Kingdom)

    Introduction: The current IBD protocol for biologics monitoring at Northwick Park Hospital states that drug levels are measured after induction; if there is a loss of response; after treatment change; if the patient has a reaction; or when the drug is stopped. An audit was carried out to investigate how many infliximab requests have been received and the reason for the requests. The aim of the audit was to determine if inappropriate requests for infliximab are being received.

    Method: Infliximab request data was collected from the LIMS between 1/4/16 and 31/3/17, and analysed to observe trends.

    Results: Three hundred and forty three infliximab requests were received from 223 different patients during this period. Of these requests, 104 (30%) had no clinical details stated. Of the 239 requests with clinical details stated, only 17 had reasons within the current protocol for testing; 193 requests (81%) stated the diagnosis of the patient as the clinical details. Sixty eight requests resulted in a positive antibody response. Of these 68 results, 22 had a repeat request with the majority of samples (14) continuing to be positive on the repeat. Eight samples returned negative after the initial positive results demonstrating a response to the change in treatment. Three patients were positive for antibodies on 3 or more occasions during this time period.

    Conclusions: The majority of requests received for infliximab do not state a reason that falls within the current protocol recommendations. Since the vast majority of requests are returning within the therapeutic range, it is possible that these are inappropriate requests if the patient is in clinical remission.

    Future Work: In order to ensure appropriate requests for infliximab are received, a clinical scientist will be attending future IBD MDT meetings to discuss an update to the current protocol and to ensure there is a biochemistry input going forward.

Endocrinology

  • P098 Observational studies on macroprolactin in a routine clinical laboratory

    S Gibbons (Specialist Laboratory Medicine, LTHT, Leeds, United Kingdom), J Barth (Specialist Laboratory Medicine, LTHT, Leeds, United Kingdom), C Lippiatt (Specialist Laboratory Medicine, LTHT, Leeds, United Kingdom), R Desborough (Specialist Laboratory Medicine, LTHT, Leeds, United Kingdom)

    Introduction: It is now recommended that all samples with raised prolactin should be examined for the presence of macroprolactin. We have performed a retrospective review of our experience of macroprolactin to determine the incidence and the natural history of macroprolactin.

    Methods: A retrospective study of macroprolactin was made in a large clinical laboratory. Samples for macroprolactin were measured on those samples where it is requested and where the total prolactin is >1000 mIU/L. Prolactin was measured using the Siemens Centaur and macroprolactin was measured following PEG-precipitation.

    Results: The incidence of macroprolactin in samples where the total prolactin was >1000 mIU/L was 36/670 (5.4%). During this period, 12064 samples were received for prolactin analysis.

    Over the period since 2006, 22 subjects had a sample with an isolated macroprolactin measurement followed by another sample without macroprolactin after a median period of 0.46 yrs. Twenty-five subjects had multiple consecutive measurements of macroprolactin lasting a median period of 2.1 years.

    Fourteen subjects had more than 6 samples which had been subjected to PEG precipitation. In these subjects, the reproducibility of PEG precipitation over a median of 6 years was 1.1 %CV (recovery 75% (26%-110%) (median [range]).

    Conclusion: The presence of macroprolactin can change over time and we cannot advise that once a test for macroprolactinaemia has been performed that it is not necessary to repeat the investigation if a subsequent sample is hyperprolactinaemic; nor can one assume that macroprolactin will not develop even if it has been excluded previously.

  • P099 A cautionary tale of using immunoassays as screening tests

    S Gibbons (Specialist Laboratory Medicine, LTHT, Leeds, United Kingdom), J Barth (Specialist Laboratory Medicine, LTHT, Leeds, United Kingdom), C Ford (Specialist Laboratory Medicine, LTHT, Leeds, United Kingdom), A Fleming (Specialist Laboratory Medicine, LTHT, Leeds, United Kingdom), A Fairhurst (Specialist Laboratory Medicine, LTHT, Leeds, United Kingdom), G Mamta (Specialist Laboratory Medicine, LTHT, Leeds, United Kingdom)

    Introduction: Immunoassay techniques are subject to a number of limitations which have been described extensively in the literature. Interference, both positive and negative can significantly impact on the validity of results obtained, and while this is widely accepted and understood by our laboratory colleagues, the same is not always true for our service users. At LTHT frontline testosterone analysis for men is carried out by immunoassay, any result falling below 8.0 nmol/L is reflexed for an LC-MS/MS testosterone.

    Case: A 53 year old male presented to his GP with a 2 month history of erectile dysfunction. Routine bloods revealed 9am testosterone of 6.9 nmol/L (immunoassay), reflex LC-MS/MS testosterone was reported as 0.8 nmol/L. The GP referred the patient to a nurse-led andrology clinic. Repeat bloods showed 9am testosterone of 9.3 nmol/L (immunoassay), LC-MS/MS testosterone was not reflexed due to our screening algorithm. It was concluded testosterone was in the normal range and hence no further action was required. Due to on-going symptoms, the patient returned to his GP the following month, a third blood sample revealed a testosterone of 6.7 nmol/L (immunoassay), the LC-MS/MS result was 2.0 nmol/L. Hypogonadism was confirmed by LH 0.8 iu/L and FSH 1.5 iu/L. The mean immunoassay testosterone was 8 nmol/L and mean LC-MS/MS was 1.4 nmol/L. These findings were relayed to the andrology nurse, suggesting probable positive interference in all immunoassay testosterone results and likely diagnosis of hypogonadism. The patient has subsequently been referred to the endocrinology clinic.

    Conclusion: This case reinforces our role in ensuring analytical problems such as these are considered when interpreting laboratory data. Never has this been truer with the increasing number of nurse-led clinics and super-specialised clinicians. Moreover, it highlights a potential pitfall in using an assay (immunoassay) which lacks sensitivity as a screening test.

Immunology

  • P130 Case report: a challenging sample matrix

    S King (Department of Clinical Biochemistry, Mid Cheshire Hospitals NHS Foundation Trust, Crewe, United Kingdom)

    Case History: An 86 year old female with known chronic kidney disease, hypertension and arthralgia was referred urgently on the suspected cancer pathway with symptoms of increased bowel frequency, diarrhoea, anorexia and anaemia. An incidental finding of asymptomatic, unprovoked pulmonary embolism (PE) was found on a computerised tomography scan of her chest and abdomen, so she was admitted and commenced on warfarin treatment.

    Blood was sent for routine biochemical analyses on admission, but results could not be produced due to sample probe obstruction. No clots were visible on initial inspection of both serum and lithium heparin samples. The serum solidified after several hours’ storage at 4°C, so the laboratory arranged further investigation for suspected cryoglobulinaemia. Following re-solubilisation of the washed cryoprecipitate in saline at 37°C, immunofixation showed a monoclonal IgM kappa paraprotein and polyclonal IgG. Total cryoglobulin load was estimated to be 60 g/L and results were phoned to the requesting team.

    After 4 days of unsuccessful warfarin treatment she was switched to Clexane and discharged, but subsequently re-admitted 2 weeks later having collapsed at home. She was largely unresponsive during the second admission and died after 5 days. The cause of death was recorded as intra-abdominal sepsis and small bowel obstruction.

    Discussion: The likely diagnosis was Type II cryoglobulinaemia due to monoclonal IgM with polyclonal IgG. The clinical syndrome of this condition varies, but is mediated by vasculitis leading to purpura, arthralgias, nephropathy, neuropathy and intestinal ischaemia. Death is most commonly caused by PE or intestinal ischaemia. Multiple factors affect protein solubility, including temperature. Treatment depends on the type, severity and underlying causes, but rituximab is effective in many cases.

    Conclusion: Cryoglobulinaemia is a potentially life-threatening pathology which is under-diagnosed and under-recognised, so good communication between the laboratory and clinicians is essential.

Oncology

  • P138 Could prostate-specific antigen velocities and alkaline phosphatase velocities be an early marker of prostatic malignancy?

    F Rose (Department of Clinical Biochemistry, Imperial College Healthcare NHS Trust, London, United Kingdom), E Bodenham (Department of Clinical Biochemistry, Imperial College Healthcare NHS Trust, London, United Kingdom), S Choudhury (Department of Clinical Biochemistry, Imperial College Healthcare NHS Trust, London, United Kingdom), P Nacmanson (Division of Women's, Children's and Clinical Support, Imperial College Healthcare NHS Trust, London, United Kingdom), J Cegla (Department of Clinical Biochemistry, Imperial College Healthcare NHS Trust, London, United Kingdom), J Alaghband-Zadeh (Department of Clinical Biochemistry, Imperial College Healthcare NHS Trust, London, United Kingdom)

    Introduction: Prostate cancer accounts for 26% of new male cancer cases in the UK. There remains no reliable method for distinguishing high-risk tumours at an early “curable” stage. Single prostate-specific antigen (PSA) measurements lack specificity, and currently there is insufficient evidence to incorporate PSA velocity into clinical decision-making. Alkaline phosphatase (ALP) velocity has been shown to predict overall and metastasis-free survival, but only in small studies for castrate-resistant prostate cancer. This study further examined the role of PSA and ALP velocities in identifying high-risk tumours at an early stage.

    Method: A retrospective audit was performed identifying 89 cases where patients had undergone investigation for prostate cancer. Five groups were created, initially stratified by prostate histology 1) no abnormality, 2) inflammation/atrophy, 3) hyperplasia, then 4) malignancy-stable outcome and 5) malignancy-poor outcome based on a clinician’s assessment of the patient record). PSA and ALP velocities were calculated prior to biopsy or surgical intervention using a velocity calculation (1000 x ((([Latest Value]/[Previous Value])-1)/[Days between results])). Mean PSA and ALP velocities for each group were calculated, and statistical analysis was performed using the Kruskal-Wallis test to evaluate significance.

    Results: PSA velocity (%change per day*10) by group: no abnormality -0.49 (±2.75); inflammation -2.24 (±7.04); hyperplasia 0.16 (±1.31); stable-outcome malignancy 0.18 (±1.56), poor-outcome malignancy 7.57 (±11.66). The PSA velocity in the poor-outcome malignancy group was statistically significantly higher than for all other groups (p<0.05).

    ALP velocity (%change per day*10) by group: no abnormality 2.59 (±7.96); inflammation 0.16 (±1.94); hyperplasia -1.4 (±9.43); stable-outcome malignancy -8.14 (±20.90), poor-outcome malignancy -1.88 (±10.36). There was no significant difference in ALP velocities between the groups.

    Conclusion: PSA velocity could be used to detect poor-outcome prostate malignancies at an early clinical stage, potentially prompting early use of more aggressive therapies. ALP velocity did not discriminate between clinical trajectories.

FRIDAY

Bone Disease Calcium Metabolism

  • P082 Variation in calculating adjusted serum calcium and its impact on clinical practice: need for standardisation

    L Ward (Department of Clinical Biochemistry, Viapath LLP, Bedford Hospital NHS Trust, Bedford, United Kingdom), J Sheffield (Department of Clinical Biochemistry, Viapath LLP, Bedford Hospital NHS Trust, Bedford, United Kingdom), N Wassef (Department of Clinical Biochemistry, Viapath LLP, Bedford Hospital NHS Trust, Bedford, United Kingdom), WS Wassif (Department of Clinical Biochemistry, Bedford Hospital NHS Trust, Bedford, United Kingdom)

    What impact does an alternative formula for serum adjusted calcium have on the biochemical calcaemic status of patients and demand for PTH and vitamin D assays?

    Our departmental protocols for processing vitamin D and PTH are primarily based on adjusted calcium concentration. Results during 2017 for albumin concentrations of >40 g/L, total and adjusted calcium concentrations, were examined. The laboratory formula for calculating serum adjusted calcium, when the albumin is <51 g/L, is based on an albumin of 40 g/L: Adjusted Calcium = [Ca] + 0.02 {40-[alb]} mmol/L. Results were re-calculated using another formula used in other laboratories: Adjusted calcium = [Ca]-0.02 {[alb]-45} mmol/L, when the albumin concentration is > 45 g/L.

    Using the current laboratory adjusted calcium formula, 28253 (40.9%) of 69078 results were below 2.25 mmol/L; our cut off for processing vitamin D requests. The alternative adjusted calcium formula identified only 7534 (10.9%) of 69078 results below 2.25 mmol/L. In contrast, using the laboratory adjusted calcium formula, 566 (0.8%) of 69078 had an adjusted calcium above 2.6 mmol/L; our cut off for processing PTH requests. The alternative adjusted calcium formula identified 1147 (1.7%) of 69078.

    Using a formula based on an albumin of 45 g/L, 11163 patients (82%) were re-classified from hypocalcaemia to normocalcaemia. In contrast, 581 patients (51%) re-classified from normocalcaemia to hypercalcaemia.

    If this alternative formula for adjusted calcium was introduced the laboratory would be processing considerably less vitamin D requests, although slightly more PTH requests than previously. Given the high cost of vitamin D immunoassay and large number of requests, using this alternative formula would result in considerable savings of approximately £111,000 annually.

    There is a need to standardise the calculation of adjusted serum calcium and the use of the ACB recommendation of March 2015 is endorsed.

  • P083 The utility of three bone markers for monitoring anti-resorptive therapy

    AP Courtney (Clinical Biochemistry, North West London Pathology, London, United Kingdom)

    Introduction: Bone markers are useful, non-invasive, repeatable, cheap tests that are sensitive to early changes in bone turnover. They are commonly used to monitor the response to therapy for osteoporosis. Our osteoporosis clinic routinely utilises three routine markers that our bone marker service offers: urine N-telopeptide (NTx), serum bone specific ALP (BALP) and serum total procollagen type-1 N-terminal peptide (P1NP). Here we review the utility of these three markers in monitoring bisphosphonate therapy in the osteoporosis clinic.

    Methods: Five years of bone marker data from the osteoporosis clinic at Imperial College Healthcare NHS Trust, was extracted from our Clinical Biochemistry data repository. Data was anonymised, and the following inclusion criteria were used: females who were on, or had been on bisphosphonates (excluding those on anabolic therapy where known) and those with multiple bone marker measurements for at least two out of the three of urine NTx, serum BALP and serum P1NP. Significant changes in these markers were compared.

    Results: The data suggest that measurement of NTx and P1NP are sufficient to identify those with significant changes in bone turnover. Compared to P1NP, NTx may potentially be superior for identifying earlier increases in turnover. Both NTx and P1NP appear to detect decreases in turnover equally well. Both markers occasional out-perform one another in some individuals suggesting that measurement of both a resorption and formation marker may be advantageous. The measurement of BALP does not appear to add value in the management of patients on anti-resorptive therapy.

    Conclusions: There appears to be merit in measurement of both serum P1NP and urine NTx for serial monitoring of patients receiving anti-resorptive therapy. Although BALP does not appear to provide any additional benefit in these patients, this marker has been shown to be useful in other cases of metabolic bone disease, such as Paget’s.

  • P084 Total and adjusted calcium measurement in critically ill patients: as bad as each other?

    J Scargill (Department of Clinical Biochemistry, Pathology at Wigan and Salford, Salford Royal NHS Foundation Trust, Manchester, United Kingdom), M Guy (Department of Clinical Biochemistry, Pathology at Wigan and Salford, Salford Royal NHS Foundation Trust, Manchester, United Kingdom)

    Introduction: Hypocalcaemia is common in critically ill patients. Despite evidence that albumin adjusted calcium may not reflect ionised calcium (iCa) in this group, it remains the practice of our laboratory to report adjusted calcium (ACa) on patients in ICU. A retrospective comparison of iCa to total calcium (TCa) and ACa values, using both a standard and locally derived equation, was undertaken in patients on the intensive care unit.

    Method: All iCa and ACa results received from ICU from 1st January-30th June 2016 were retrieved. Patients with a whole blood sample for iCa and a serum sample collected for ACa measurement within 30 minutes of each other were included. Patients on CVVH treatment were excluded. iCa was measured on a Radiometer ABL800, serum total calcium and albumin on Siemens ADVIA 2400 (o-CPC and BCG methods respectively).

    Results: One thousand three hundred and fifty three paired samples were identified from 474 adult patients (median age 61 years, 55.5% male). Prevalence of hypocalcaemia using iCa (<1.12 mmol/L) was 33%. Prevalence of hypocalcaemia (<2.20 mmol/L) using TCa was 91%, ACa using a standard equation 62%, and using ACa using a local equation was 16%.

    ROC curve analysis for all 3 surrogate measures of iCa for hypocalcaemia were similar, with AUC for TCa of 0.84 (95%CI 0.82-0.87, optimum cutoff 1.94 mmol/L), 0.86 for ACa determined by a standard equation (95%CI 0.83-0.88, optimum cut-off 2.12 mmol/L), and for ACa determined by a local equation 0.85 (85%CI 0.83-0.87, optimum cut-off 2.30 mmol/L).

    Discussion: Whilst prevalence of hypocalcaemia using iCa, TCa and ACa (as defined by standard cut-offs) varied, when compared to iCa, total and adjusted calcium methods had similar diagnostic test performance. This study supports the findings of smaller prospective studies that use TCa and ACa is not suitable for diagnosing hypocalcaemia in critically ill patients and should not be reported.

  • P085 Dent's disease: a genetic disorder with late presenting severe renal and bone disease

    MOR Hajjawi (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), J Simmonds (Clinical Genetics, Leeds Teaching Hospitals NHS Trust, Leeds, United Kingdom), R Jewell (Clinical Genetics, Leeds Teaching Hospitals NHS Trust, Leeds, United Kingdom), M Chanayireh (Renal Department, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), E Moulder (Orthopaedics Department, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), RL Wilmot (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), D Narayanan (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom)

    A 16 year old male presented with bilateral genu valgum (knock-knees) due to severely deformed femurs. He suffered from chronic pain in his legs, was unable to exercise, and his BMI reached 30.4 kg/m2. This affected his mental health, social engagement and his education; corrective surgery was scheduled. Pre-operative blood tests found: 25OH-vitamin D, 22.6 nmol/L; low phosphate, 0.42 mmol/L (1.0-1.8); raised alkaline phosphatase, 822 iU/L (75-400); adjusted calcium, 2.40 mmol/L (2.2-2.6); parathyroid hormone, 6.2 pmol/L (1.39-9.3). His overt vitamin D deficiency was corrected, but this did not resolve his hypophosphatemia or hyperphosphatasaemia. Unexpectedly, his renal function started to deteriorate: creatinine, 204 µmol/L (65-114) vs baseline creatinine 99 µmol/L; potassium, 3.4 mmol/L (3.5-5.3). Urine dipstick was positive for blood and protein. Imaging showed abnormal asymmetry of the kidneys with one being dominant. Kidney biopsy was inconclusive but calcium deposits were detected in the interstitium. Further testing found: high urine calcium excretion, 10.4 mmol/24 hrs (2.5-7.5); urine phosphate excretion, 39 mmol/24 hrs (16-48); protein creatinine ratio, 225 mg/mol. A review of his history found that at age 8 he presented with joint pain and was dipstick positive for blood and protein. At age 13 he presented with bilateral pes planovalgus (flat feet) causing patellofemoral irritability. Family history prompted referral to clinical genetics. Genetic analysis identified a mutation in the CLCN5 gene associated with a form of hypophosphataemic rickets (Dent’s disease type 1). This X-linked condition, detected in only 250 families, is typically associated with proximal renal tubular dysfunction characterised by proteinuria, hypercalciuria, nephrocalcinosis, nephrolithiasis and chronic kidney disease. This case highlights how rare genetic disorders can present relatively late, when they are not readily considered in the differential diagnosis.

  • P086 Under recognition of benign transient hyperphosphatasaemia complicates the diagnosis of walking difficulty

    MOR Hajjawi (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), V Mogford (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), V Johnson (Orthopaedics Department, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), RL Wilmot (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), D Narayanan (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom)

    The duty biochemist in the pathology laboratory incidentally came across a raised alkaline phosphatase (ALP) result and requested that ALP isoenzyme analysis be performed. The sample was from a one year old female infant brought to the A&E department by her parents, who noticed that she was no longer able to bear her own weight and was reluctant to walk This had been ongoing for approximately two weeks and did not appear to be resolving. The child had a viral infection during the previous week. There was no history of trauma, plain X-rays did not reveal evidence of a fracture. Under manual manipulation, the child’s ankles and knees moved freely and without pain, as did the subtalar, midtarsal and tarsometatarsal joints in her foot. Blood tests showed the increased ALP, 8484 iu/L (75-300); adjusted calcium, 2.62 mmol/L (2.2-2.6); phosphate, 2.20 mmol/L (1.0-2.0); PTH, 1.0 pmol/L (1.3-9.3); 25-hydroxy vitamin D, 76.6 nmol/L.The extremely elevated ALP result triggered an unnecessary urgent referral to the orthopaedic surgery department, which in turn led to urgent discussions with the endocrinology department. Electrophoresis showed that the patient had benign transient hyperphosphatasaemia; this test halted a spiralling series of investigations. In this case, unrecognised benign transient hyperphosphatasemia resulted in a misleading distraction in the diagnosis of the patient’s walking difficulty. Transient hyperphosphatasemia is characterised by a marked rise, and then fall in ALP. This rise in ALP usually normalises within 2-4 months, with peak activity at around 6 weeks. The cause of transient hyperphosphatasemia is not fully understood; it has been linked to viral infections, liver development and impaired excretion. However, there is no evidence linking this condition to any physical symptoms or adverse outcomes. Benign transient hyperphosphatasemia is a common yet under recognised condition.

  • P087 Changes in bone turnover markers in advanced prostate cancer patients treated with LHRH agonists and transdermal oestradiol patches

    A Courtney (Department of Blood Sciences, Imperial College NHS Healthcare Trust, London, United Kingdom), R Harvey (Department of Blood Sciences, Imperial College NHS Healthcare Trust, London, United Kingdom), R Abel (Department of Surgery & Cancer, Imperial College London, London, United Kingdom), P Abel (Department of Surgery & Cancer, Imperial College London, London, United Kingdom)

    Introduction: A recent study by Langley et al (Eur Urol 2016, 69(6): 1016-25) has shown that prostate cancer (PC) patients treated with oestradiol patch (OP) instead of luteinising hormone-releasing hormone agonist (LHRHa) therapy do not suffer loss of bone mineral density (BMD) but in fact have improved BMD. This is consistent with previous studies that have shown that oestradiol supresses bone resorption and promotes bone deposition. In this study we measured biochemical bone turnover markers (BTM) to further investigate the mechanism of the observed BMD changes.

    Method: The resorption marker urine N-telopeptide (NTx) and formation marker serum total procollagen type-1 N-terminal peptide (P1NP) were measured using immunoassays in 9 OP and 9 LHRHa treated patients. Baseline, 1 and 2 year post-therapy samples were analysed and compared. Slope-values for changes in BTM were derived by Pearson regression and compared using Mann-Whitney U-test.

    Results: From baseline measurements there was an increase in P1NP (mean slope 26.2 µg/L year-1, SD = 31.3; n=9) and NTX (mean slope +12.0 nM/mM Cr year-1, SD = 17.8; n=7) in OP treated PC patients. In LHRHa patients there was a decrease in both P1NP (mean slope -39.6 µg/L year-1; SD = 76.0; n=9) and NTx (mean slope -24.5 µg/L year-1; SD = 36.6; n=9). Comparison of slope-values between the groups showed that the differences were statistically significant for both P1NP (p<0.005) and NTx (p<0.01).

    Conclusions: BTM were increased in OP treated patients, with a relatively higher increase in the formation marker suggesting a net increase in BMD. In LHRHa treated patients both markers decreased over the two-year study period. These findings are consistent with previously reported BMD differences between the two treatment groups. Assessment of bone markers could provide a sensitive and convenient method for monitoring hormone treated PC patients for changes in BMD and management of fracture-risk.

  • P088 Chance findings: choose the next steps wisely to avoid delay

    MOR Hajjawi (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), V Mogford (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), D Narayanan, (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom), RL Wilmot (Chemical Pathology, Hull and East Yorkshire Hospitals NHS Trust, Hull, United Kingdom)

    In June 2017, an 89 year old lady was seen by her GP for a health check, as part of which he requested routine biochemistry. All results were within reference limits except for an e-GFR of 33 (stable for the previous 4 years) and an alkaline phosphatase which was elevated at 772 iU/L (30-125). Alkaline phosphatase isoenzymes were requested by the GP as a follow up test in July when her activity had risen to 842; gamma GT 14 U/L (9-65). The isoenzyme pattern was unusual and whilst not typical it was thought that it could be consistent with benign transient hyperphosphatasaemia and monitoring of activity was advised to confirm this. The alkaline phosphatase remained elevated and repeat isoenzyme analysis in September showed a similar pattern. The case was discussed with her GP and further bone investigation, inclusive of PTH and vitamin D was advised. Results showed the alkaline phosphatase remained elevated at 888 iU/L, an adjusted calcium of 2.19 mmol/L (2.20-2.60) which reduced from 2.25 previously, and phosphate 1.05 mmol/L (0.8-1.5). The most obvious abnormality was however a PTH 97.8 pmol/L (1.3-9.3). Vitamin D was subsequently shown to be <5 nmol/L. Unusually, given the number of vitamin D requests we receive from this age group, she had not had a vitamin D level checked since 2012 when it was 49.5 nmol/L (at that time her alkaline phosphatase and calcium were well within reference limits). High dose vitamin D with monitoring of her calcium and PTH was advised. This case is a reminder that vitamin D deficiency can present with some startling biochemical abnormalities.

  • P089 Calcium >4.0 mmol/L and PTH >200 pmol/L: is it parathyroid carcinoma? Two cases which may suggest yes and no

    JC Clayton (Department of Clinical Biochemistry, Salford Royal NHS Foundation Trust, Salford, United Kingdom), TM Kearney (Department of Diabetes and Endocrinology, Salford Royal NHS Foundation Trust, Salford, United Kingdom), A Robinson(Department of Diabetes and Endocrinology, Salford Royal NHS Foundation Trust, Salford, United Kingdom), R Brindle (Department of General Surgery, Salford Royal NHS Foundation Trust, Salford, United Kingdom), H Doran (Department of General Surgery, Salford Royal NHS Foundation Trust, Salford, United Kingdom), JK Borzomato (Department of Clinical Biochemistry, Salford Royal NHS Foundation Trust, Salford, United Kingdom)

    We present two cases of primary hyperparathyroidism with significant hypercalcaemia and hyperPTH.

    A 74 year old female with known hyperparathyroidism and osteoporosis (right forearm T score = -3.0) on alendronic acid 70mg weekly and cinacalcet 60 mg OD presented to A&E with a 1 week history of confusion and lethargy. Bloods showed adjusted calcium (ACa) of 3.86 mmol/L with PTH of 101.9 pmol/L and she was admitted for treatment. ACa and PTH continued to rise, peaking at 4.53 and 277.6 respectively. Four weeks previously ACa and PTH were 3.29 and 24.0. Assay interference was excluded. She was treated with zoledronic acid and calcitonin, and her ACa and PTH began to normalise. Surgery was performed during which the right and left superior and left inferior parathyroid glands were dissected. Histology suggested parathyroid hyperplasia rather than adenoma. Immediately following surgery ACa and PTH were 2.34 and <0.5. Post-surgery she exhibited a mild degree of hungry bone syndrome on Calcichew (6g calcium), with hypocalcaemia (nadir 1.79) and elevated ALP (pre-surgery 30 U/L, rising to 100 U/L post-surgery). This recovered and she was discharged home.

    A 61 year old male presented to his GP with a 1 month history of weight loss, thirst, dry mouth and painful knees and ankles. Previous medical history includes gastric bypass and cholecystectomy. His bloods showed ACa of 4.23 with PTH of 245.8. He was admitted for fluids and bisphosphonate. He was noted to have a parathyroid mass and iliac brown tumours on CT, suggestive of longstanding hyperparathyroidism-related osteoporosis. Given the red-flag symptoms for parathyroid cancer (significant hypercalcaemia, hyperPTH and palpable parathyroid mass) he had left hemithyroidectomy with level 6 node clearance after which he became profoundly hypocalcaemic with hungry bone syndrome, potentially prolonged due to bariatric surgery.

    These cases show how hyperparathyroidism can present in similar ways but with different outcomes and causes.

  • P090 A gut feeling: an unusual presentation of an isolated raised alkaline phosphatase

    L Seyani (North West London Pathology, Imperial College Healthcare NHS Trust, London, United Kingdom), A Courtney (North West London Pathology, Imperial College Healthcare NHS Trust, London, United Kingdom)

    A 26-year-old man, presented to A&E with one week of feeling unwell. He had a progressive lower limb vasculitic rash, arthritis involving the ankles and knees, and MCP joint synovitis. He represented to A&E a week later with abdominal pain and bloody diarrhoea. His acute phase inflammatory markers were raised with an elevated IgA. A probable diagnosis of Henoch-Schonlein purpura was made and was treated with steroids to which he responded well.

    An incidental finding of raised alkaline phosphatase (ALP) levels was observed after the initiation of steroid treatment. The ALP has remained persistently elevated between 300 and 500 (reference range 30-130 IU/L). ALP isoenzyme electrophoresis shows a marked increased intestinal fraction of ALP with marginally increased bone and normal liver fractions. His renal and other liver profiles, including gamma GT were normal.

    Measurement of the bone turnover markers, serum bone specific ALP and total procollagen type-1 N-terminal peptide showed a high bone turnover. However, the patient remains asymptomatic, denies any joint pains or swelling and has normal bone densitometry and whole body bone scans. Abdominal imaging showed some thickening of the terminal ilium with normal oesophago-gastro-duodenoscopy and colonoscopy. Biopsy of the terminal ilium was unremarkable.

    Can vasculitis in the gut cause an increase in ALP? Or is it a benign biochemical finding? In our patient, no obvious or significant pathology has been directly linked to his marked elevated intestinal ALP, with the underlying cause remaining elusive. Here we illustrate and discuss the importance of measurements of ALP, both total and isoenzymes, to aid in diagnosis of liver, bone, intestinal and parathyroid diseases, their importance in determining the source of an increased ALP and the pitfalls of investigation.

Cardiovascular

  • P091 High troponin I levels and multiple cardiac interventions: is the assay to blame?

    A Gogokhia (Department of Endocrinology, Aneurin Bevan University Health Board, Caerphilly, United Kingdom), P Nalla (Department of Endocrinology, Aneurin Bevan University Health Board, Caerphilly, United Kingdom), N El-Farhan (Department of Clinical Biochemistry, Aneurin Bevan University Health Board, United Kingdom), M Adlan (Department of Endocrinology, Aneurin Bevan University Health Board, Caerphilly, United Kingdom), L Premawardhana (Department of Endocrinology, Aneurin Bevan University Health Board, Caerphilly, United Kingdom), F Stratford (Department of Clinical Biochemistry, University Hospital of Wales, Cardiff, United Kingdom), R Brixey (Department of Clinical Biochemistry, University Hospital of Wales, Cardiff, United Kingdom)

    Cardiac troponin I (cTnI) and cardiac troponin T (cTnT) are sensitive and specific markers of cardiac injury, and useful for diagnosing acute myocardial infarction. However, non-infarction related causes (e.g. tachycardia) and analytical errors (e.g. heterophilic antibodies) may also cause spurious troponin elevation.

    A 82 year old woman, who had previous “cardiac disease”, was admitted with a cough and back pain. Her cTnI was high at 6.06 µg/L (normal range <0.05) and she was treated for an acute coronary syndrome (ACS). Previously, she had three episodes of “ACS” with elevated cardiac troponins but was treated conservatively as her ECGs, echocardiogram and coronary angiograms showed no evidence of coronary artery disease. Despite ACS treatment on this admission, cardiac troponins remained persistently elevated without typical modulation. We then checked troponins in different assay platforms and found that spurious elevation of cardiac troponins occurred only in one assay (Standard Abbott ARCHITECT cTnI). Serial dilutions of her serum showed a linear pattern inconsistent with heterophilic antibody interference. Furthermore, recovery after polyethylene glycol precipitation of >50% excluded macrotroponin interference. We discuss potential causes for this type of assay interference, and highlight its potential to cause incorrect diagnoses resulting in unwarranted and potentially dangerous interventions. Clinicians should be aware of this phenomenon when troponin results are inconsistent with clinical presentation and cardiac investigations.

  • P092 Implementation of a new 1-hour chest pain pathway in the emergency department leading to improved hs-cTnT requesting

    G McKeeman (Dept of Clinical Biochemistry, Belfast Health & Social Care Trust, Belfast, United Kingdom), L Swales (Emergency Department, Belfast Health & Social Care Trust, Belfast, United Kingdom), J Collins (Emergency Department, Belfast Health & Social Care Trust, Belfast, United Kingdom), S Langtry (Emergency Department, Belfast Health & Social Care Trust, Belfast, United Kingdom), P Shortt (Emergency Department, Belfast Health & Social Care Trust, Belfast, United Kingdom), N Johnston (Cardiology, Belfast Health & Social Care Trust, Belfast, United Kingdom)

    Background /Aim: New guidelines were published in 2015 by the European Society of Cardiology (ESC) for management of suspected myocardial infarction (MI) in patients presenting without persistent ST-segment elevation. These note a 1-hour pathway using high sensitivity troponin T (hs-cTnT) measurements (collected at presentation and after 1 hour) alongside clinical assessment for assisting with diagnosing MI. A multidisciplinary team was established to review and implement these guidelines into a new trust chest pain pathway and to audit pathway performance.

    Methods: A pilot study was performed to assess feasibility of implementing the new protocol in the emergency department (ED), which demonstrated that the new pathway was safe, feasible and applicable to the local population. A new trust pathway was finalized to incorporate the new protocol and successfully implemented in the ED after a targeted educational week. An audit was carried out 3 months after introduction to assess safety, impact on clinical decision making and patient flow and to review influence on hs-cTnT requesting.

    Results: Analysis demonstrated that AMI could be safely ruled out for 71% of patients (total cohort n=1454) within 1 hour, with a negative predictive value of 99.2%. There were no deaths at 90 days and no new cardiac diagnoses in the rule-out group (n=1041) and 85% of patients in the rule-in group (n=87) had a diagnosis of NSTEMI confirmed. Only 23% of the cohort required further hs-cTnT testing (at 3 h) and ED hs-cTnT requests have decreased by approximately 200 per month since pathway introduction (Jul-16 to Sept-17).

    Conclusions: Adoption of the new ESC guidelines into clinical practice has improved cost-effectiveness of hs-cTnT requesting in ED (highlighting previous over-requesting) and improved patient flow, allowing for quicker rule-in or rule-out of MI. Performance has been comparable to data published in other multi-centre trials using such a pathway.

  • P093 Analytical evaluation of the Singulex Clarity® cTnl assay using human serum samples

    TG Morris (Clinical Blood Sciences, St George's University Hospitals NHS Foundation Trust, London, United Kingdom), D Gaze (Department of Biomedical Science, University of Westminster, London, United Kingdom), P Collinson (Clinical Blood Sciences, St George's University Hospitals NHS Foundation Trust, London, United Kingdom)

    Introduction: The universal definition of acute myocardial infarction (AMI) puts cardiac troponin at the forefront of diagnosis. With the advent of high-sensitivity assay for cardiac troponins the diagnosis of AMI can be made sooner with a higher degree of confidence. The aim of this study was to provide an analytical evaluation of the Singulex Clarity® cTnl assay using serum samples.

    Methods: Imprecision profile, limit of blank (LoB) and limit of detection (LoD), linearity and sample stability were determined using pooled human serum samples. A sample type comparison was performed with paired serum and EDTA plasma. The 99th percentile upper reference limit (URL) was calculated with serum from 638 apparently healthy individuals. Receiver operating characteristic (ROC) curves were used to compare the clinical performance of the Singulex Clarity® cTnl assay with the Abbott Architect hs-cTnI and Roche hs-cTnT assays in a limited study using serum; the Abbott and Roche 99th percentile URLs were used as the diagnostic cut-offs.

    Results: The LoB was 0.01 ng/L, LoD 0.04 ng/L, 10% CV 0.49 ng/L and linearity 0.5-20,000 ng/L. Serum samples had a consistent positive bias compared to EDTA plasma. Samples were stable for five days and for five freeze-thaw cycles at -20oC, four days at 4oC and for two days at room temperature. The overall 99th percentile URL was 4.33 ng/L (2.53 ng/L in females and 5.53 ng/L in males). Area under the ROC curves for the Singulex Clarity® cTnl assay when compared to the Abbott and Roche assays were 0.900 (95% CI 0.805-0.995) and 0.927 (95% CI 0.860-0.993), respectively.

    Conclusion: The Singulex Clarity® cTnl assay meets the criteria for a high-sensitivity assay. There was an expected matrix effect in serum compared to EDTA plasma. There was an expected male/female difference in 99th percentile URL. Clinical diagnostic performance appeared better than predicate assays.

  • P094 A change of heart: impact of a new 'acute chest pain' high sensitivity troponin T pathway algorithm

    KB Rice (Department of Clinical Biochemistry, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), NL Barlow (Department of Clinical Biochemistry, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), JD Berg (Department of Clinical Biochemistry, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom)

    In 2016, the Sandwell and West Birmingham Hospitals NHS Trust ‘Acute Chest Pain’ algorithm was updated to improve patient flow through A&E by incorporating a new protocol for high-sensitivity Troponin T (hsTnT) into a robust clinical pathway. The new protocol allows rule-out of acute coronary syndrome (ACS) with hsTnT ≤5 ng/L and rule-in with hsTnT ≥52 ng/L. Patients with baseline hsTnT 5-51 ng/L require a 2nd sample after 3 hours, with a concentration change of ≥5 ng/L indicating ACS. In the old protocol, all patients required a 2nd sample, 3 hours post-baseline, and a change of >20% with one hsTnT result ≥14 ng/L, was consistent with ACS.

    The aim of this retrospective audit was to assess the impact of, and adherence to, the new pathway algorithm. The hs-TnT results from a 2 week period prior to, and after, implementation were reviewed.

    Comparing results with baseline hsTnT of 5-51 ng/L, the new criteria led to a slight increase in the number of “positive” results (10% vs 6%). Of patients with hsTnT <5 ng/L, 27% still had a 2nd sample taken at 3 hours. Adherence to the 3 hour time frame for the 2nd sample was poorer following implementation of the new guidance, with only 38% of samples being taken at 2.75 to 3.25 hours after baseline, compared with 46% in 2016.

    The introduction of this new hsTnT protocol has allowed for more rapid discharge of low risk patients on the basis of hsTnT results but yielded a slight increase in positive ACS results with baseline hsTnT in the intermediate range. Second samples being taken when hsTnT <5 ng/L may be justified by clinical context or reflect lack of familiarity with the new guidance. The apparent decrease in adherence to 2nd sample timing guidelines post-implementation raises concerns as action limits are based on expected trajectory of troponin T release following myocardial damage.

  • P095 A case of severe hyperhomocysteinaemia and stroke: secondary prevention following laboratory intervention

    SA Bowles (Department of Blood Sciences, Countess of Chester Hospital NHS Foundation Trust, Chester, United Kingdom)

    A 49-year old male presented with right facial weakness and dysarthria. An MRI examination of his head identified a focal area of restricted diffusion, thought to represent acute ischaemia, but a number of chronic lacunar infarcts were also seen. The only cardiovascular disease risk factor evident was cigarette smoking.

    The patient was started on the usual therapy for secondary prevention, including atorvastatin 80 mg (pre-treatment cholesterol 3.8 mmol/L), but he was unable to tolerate either this or a lower dose, so statin therapy was discontinued and he was referred to the Lipid Clinic.

    Routine pre-clinic bloods included a full blood count and, as a raised MCV (121 fL) was noted by laboratory personnel, a serum B12 and folate were added to the request. These tests indicated deficiency of both vitamins: B12 114 ng/L (150-750); folate 2.6 mg/L (>4.0). In light of these results, and the clinical history, a plasma homocysteine was requested: this was markedly elevated at 235 µmol/L (5-14).

    The patient was prescribed B12 injections and folate supplements and, by the time of his next clinic review, his B12 and folate levels had improved (B12 784 ng/L; folate >23.0 mg/L), although he still had a raised MCV (105 fL). His plasma homocysteine had decreased significantly: 36 µmol/L.

    Homocysteine has both atherogenic and prothrombotic properties, and evidence suggests that hyperhomocysteinaemia is an independent risk factor for cardiovascular disease, including stroke. Clinically significant hyperhomocysteinaemia may be a consequence of a genetic defect in any of the three interrelated pathways of methionine metabolism and/or deficiency of the vitamin cofactors (B12, folate, pyridoxine). The response of this patient’s hyperhomocysteinaemia to folate and B12, without pyridoxine, would tend to exclude classic homocystinuria (cystathione beta-synthase deficiency), but he may have the common methylene tetrahydrofolate reductase polymorphism, with reduced enzyme activity.

    • Endocrinology

      • P096 A clinical evaluation of the value of adding FT4 to mildly abnormal TSH results

        A Jones (Clinical Biochemistry, York Teaching Hospitals NHS Foundation Trust, York, United Kingdom)

        Aim: Current laboratory protocol is to add FT4 to all first-time abnormal TSH results. This evaluation was carried out to see if adding FT4 to mildly elevated TSH results below 10 mU/L, or mildly suppressed TSH results greater than 0.1 mU/L, adds any clinical benefit to interpretation of TFT results.

        Methods: Results from samples with both TSH and FT4 requested were obtained from the laboratory information management system over a four-month period. TSH results within normal limits (0.54-4.8 mU/L), <0.1 mU/L or >10 mU/L, and patients on T4 or T3 therapy were excluded from the data.

        Results: Raised TSH (n=1979): FT4 was outside the reference range in 9 samples (0.45%). One sample was above the reference range: this was a two day old baby, so the results were normal for age. Eight results were below the reference range: two patients were treated hyperthyroid; two patients died within 30 days of the result; four patients had no obvious cause.

        Low TSH (n=1887): FT4 was outside the reference range in 17 samples (0.9%). Five results were below the reference range: three of these were patients with a history of thyrotoxicosis; two were acutely unwell patients. Twelve results were above the reference range: four patients had a previous history of thyroid cancer; one patient had a previous history of thyrotoxicosis; four patients were acutely unwell; two patients were GP patients, no obvious cause; one patient was a new diagnosis of mild thyrotoxicosis.

        Conclusion: There was a low frequency of abnormal FT4 results when associated with first time mildly abnormal TSH results. It is therefore appropriate to advise repeat TSH in 4-6 weeks without adding FT4 to these samples.

      • P097 A case of abnormal thyroid function test induced by cabozantinib

        L Kayali (Chemical Pathology, Nottingham University Hospital, Nottingham, United Kingdom), R Lopez-Real (Chemical Pathology, Nottingham University Hospital, Nottingham, United Kingdom), H Divyateja (Chemical Pathology, Nottingham University Hospital, Nottingham, United Kingdom)

        Presentation: A case of a 14 year old child diagnosed with metastatic medullary thyroid carcinoma (on the background of MEN 2B), treated with cabozantinib 60 mg for almost 2 years. Cabozantinib is a tyrosine kinase inhibitor (TKI) drug; it is approved for the treatment of metastatic medullary thyroid carcinoma and advanced renal cell carcinoma.

        Since he has been treated with cabozantinib, serial measurement of TFT were undertaken in this patient, which revealed abnormal thyroid results after 12 months of starting this medication (TSH>100 mu/L, FT4 5.4 pmol/L, FT3 2.3 pmol/L) suggestive of hypothyroidism.

        Investigations and findings: Clinical examination did not reveal any signs of hypothyroidism. In view of TFTs, he was commenced on levothyroxine initially at 50 mcg/day increased to 75 mcg after few weeks. Although FT4 and FT3 results returned to normal levels, TSH remained persistently elevated (ranged between 16.5 - 95 mu/L).

        In view of persistently elevated TSH despite patient being administered the thyroxine by his parents regularly and that FT4 and FT3 were normal, further biochemical tests to rule out interference were undertaken. TSH was repeated after treatment with blocking tube and on different immunoassay platforms; using one and two step assays but TSH still remained elevated. No antibodies were found against FT4 and FT3 hormones.

        Conclusion: The likely explanation for the abnormal TFT is to be TKI induced. TKI may cause significant thyroid dysfunction in children and adults varying from subclinical hypothyroidism to symptomatic thyrotoxicosis. Biochemists are to be made aware of thyroid dysfunction in patients on TKI so that appropriate biochemical tests are organised to help the clinicians in the management of these patients.

      • P100 Development and validation of a single LC-MS/MS assay for the analysis of serum, urine and salivary cortisol and salivary cortisone

        H Beeston (Department of Clinical Biochemistry, Manchester Royal Infirmary, Manchester, United Kingdom), E Hinchliffe (Department of Clinical Biochemistry, Manchester Royal Infirmary, Manchester, United Kingdom)

        Aim & Introduction : Compared to immunoassays, the increased specificity of LC-MS/MS analysis for cortisol produces more accurate results. It is recommended that LC-MS/MS should be used for quantification of urine and salivary cortisol samples, as well as serum cortisol samples from patients taking metyrapone. The aim of this project was to develop an in-house LC-MS/MS assay for the quantification of serum, urine and salivary cortisol/cortisone.

        Methods: 100 µL sample was subject to a Transcend TLX TurboFlow sample preparation system (with Cyclone-P Turboflow column) integrated with the Dionex UltiMate 3000 UHPLC and TSQ Endura triple quadrupole mass spectrometer (ThermoScientific) utilising a Luna Omega C18 analytical column for quantification of cortisol and cortisone. Calibrators were made from PBS 0.1% BSA spiked with cortisol and cortisone certified reference material. Internal standards used were D4-cortisol and D7-cortisone. Multiple reaction monitoring was used to select and quantify the ions of interest.

        Results: Ion suppression, linearity, recovery, carryover, specificity, extract stability, lower limit of quantification (LLOQ) and precision of the assay were deemed to be acceptable according to published validation criteria. Method comparisons were performed to compare cortisol results produced by LC-MS, immunoassay and EQA material. The assay LLOQ for serum = 5 nmol/L, urine = 5 nmol/L, salivary cortisol = 0.5 nmol/L and salivary cortisone = 0.25 nmol/L; 1 µmol/L of prednisolone had a cross-reactivity of 0.5 % observed in the cortisol transition. In addition, serum samples collected from congenital adrenal hyperplasia patients with significantly elevated 17-hydroxyprogesterone level were assessed for cortisol. The results were in good agreement with measurements by Roche gen II immunoassay and LC-MS/MS (Passing-Bablok regression analysis y=0.982x – 5.6.)

        Conclusions: Overall, this work demonstrates that an accurate and precise LC-MS/MS method has been achieved, suitable for the routine clinical quantification of serum cortisol, urine free cortisol and salivary cortisol/cortisone.

      • P101 A retrospective review of the use of growth hormone status testing at a district general hospital

        K Fenna (Department of Clinical Biochemistry, Royal Surrey County Hospital, Guildford, United Kingdom)

        Background: Dynamic function tests (DFTs) of growth hormone (GH) status aid investigation of possible deficiency/excess. At the Royal Surrey County Hospital (RSCH) the following DFTs are used; exercise test, glucagon test, insulin stress test (IST) and GH suppression test. Insulin-like growth factor-1 (IGF-1) mediates the growth-promoting effects of GH and its serum concentrations are thought to reflect GH status. Measuring baseline IGF-1 baseline is therefore thought to be a useful adjunct to these tests.

        Aim: The aim is to review the use of GH DFTs. The data will provide useful information for interpretation and a baseline to consider service changes. The RSCH is uniquely placed to conduct this project and the results will make a valuable contribution to the clinical service.

        Methods: Results from Jan 2010-Jan 2017 have been accessed, extracted and analysed. For each DFT baseline IGF-1, GH response, time of response and GH response at 180 min and 240 min in the glucagon stimulation test was recorded. From this data, subsequent analysis was performed.

        Results: Two hundred and seventy two GH DFTs were performed. Baseline IGF-1 and GH responses were analysed. The results showed poor correlation and therefore limited predictive value of baseline IGF-1 in both suppression and stimulation testing in all cohorts. Review of the GH peak in the glucagon stimulation test showed 20/22 patients had a peak response on or before 180 minutes. The two patients with a delayed peak both had inadequate responses, so protocol shortening would not have changed the outcome. Therefore shortening the protocol from 240 to 180 minutes could be considered.

        Conclusions: Based on the findings, the glucagon stimulation test in adults to could be shortened to 180 min, as has been done in children. Analysis of the correlation between baseline IGF-1 and GH response in both stimulation and suppression tests showed poor correlation in all cohorts and therefore suggests limited predictive value.

      • P102 Case report: Isolated raised 3-methoxytyramine due to analytical interference from midodrine in plasma metadrenaline analysis by LC-MS/MS

        V Treasure (Department of Clinical Biochemistry (Viapath), King's College Hospital NHS Foundation Trust, London, United Kingdom), R Vincent (Department of Clinical Biochemistry (Viapath), King's College Hospital NHS Foundation Trust, London, United Kingdom), K Neff (Department of Endocrinology, King's College Hospital NHS Foundation Trust, London, United Kingdom), D Taylor (Department of Clinical Biochemistry (Viapath), King's College Hospital NHS Foundation Trust, London, United Kingdom)

        Endocrine Society guidelines recommend biochemical investigation of suspected phaeochromocytoma and paragangliomas (PPGL) using O-methylated catecholamine metabolite measurements; these are metadrenaline (adrenaline metabolite), normetadrenaline (noradrenaline) and 3-methoxytyramine (dopamine). In our laboratory, we measure plasma metadrenalines using TurboFlow liquid chromatography tandem mass spectrometry (TurboFlow-LC-MS/MS). LC-MS/MS methods typically offer excellent sensitivity and specificity, although isobaric interferences remain a possibility.

        We report a case of a 31 year old female who was found to have normal metadrenaline concentration of 87 pmol/L (reference range 80-510) and normetadrenaline concentration of 129 pmol/L (120-1180), but markedly raised 3-methoxytyramine at 3404 pmol/L (<120). A repeat sample confirmed these findings. Whilst these results could be due to head and neck PPGL, SDH-related pathologies or DOPA-related medications, a search for clinical information revealed that the patient had postural orthostatic tachycardia syndrome (POTS) and was receiving midodrine therapy (α1 adrenergic receptor agonist). Due to the structural similarities between midodrine and metadrenalines, we investigated whether midodrine or its metabolite desglymidodrine may be causing 3-methoxytyramine interference.

        Midodrine gave main m/z transitions of 237>180 and 180>133, which correspond to the intact drug midodrine and the hypothetical m/z of desglymidodrine (resulting from in source fragmentation). We also identified ionic cross-talk with metadrenaline (m/z 180>148, 180>121) and 3-methoxytyramine (m/z 151>91, 151>119) transitions. Chromatography revealed that intact midodrine did not co-elute with metadrenaline, normetadrenaline or 3-methoxytyramine. However, when the patient sample was reanalysed, a midodrine-related fragment, inferred to be desglymidodrine, co-eluted with 3-methoxytyramine resulting in a falsely raised result.

        In conclusion, whilst usually specific, LC-MS/MS users must remain alert to the potential for isobaric interference. In this case, careful chromatogram examination and clinical interaction facilitated the identification of midodrine as a novel interference in our method.

      • P103 Anti-Müllerian hormone reference intervals on two automated immunoassay platforms: a study conducted in a Singapore female population

        P Chincholkar (Department of Clinical Pathology, Singapore General Hospital, Singapore), H Rajesh (Department of Obstetrics and Gynaecology, Singapore General Hospital, Singapore), CP Yeo (Department of Clinical Pathology, Singapore General Hospital, Singapore)

        Aims: Anti-Müllerian hormone (AMH) is considered a useful marker for assessing primordial follicular number. The commonly used test format has been a manual enzyme-linked immunosorbent assay (ELISA) but AMH can also now be readily measured on automated immunoassay platforms. Recognising that reference intervals provided in commercial test kits insert (based on Caucasian population) may not be applicable to our local Singapore population of Asians, this study aims to survey AMH levels in a local female population with 2 candidate automated immunoassay platforms.

        Methods: Two test platforms, Roche Cobas e602 and Beckman Coulter DxI800, were used for the study. Healthy females between 21 to 45 years were recruited based on inclusion and exclusion criteria (e.g. FSH >10 U/L), given a lifestyle questionnaire and consent sought for blood collection. Serum was analysed for AMH on both test platforms along with FSH. Data was analysed and reference intervals computed using ‘Analyse it’ software.

        Results: The study recruited 204 females (89 Chinese, 50 Indian, 48 Malay, 17 others). AMH levels on Roche platform are lower than that Beckman Coulter platform, consistent with literature on the 2 analysers’ performance; the computed reference intervals (2.5th-97.5th percentile) for 21-45 y (n=168) are 0.2-9.2 ng/mL and 0.3-11.3 ng/mL respectively. With a lower age cutoff at 40 y, the 21-40 y group (n=157) reference interval shifted to slightly higher values at 0.4-9.5 ng/mL and 0.5-11.5 ng/mL respectively. These data showed higher values at 97.5th percentile compared to the insert values (Roche 7.49 ng/mL, Beckman 7.15 ng/mL).

        Conclusion: Ethnic differences in characteristics of oocyte donors and infertility patients have been reported; with Asian Chinese women having higher mean AMH levels. Our study provided valuable reference intervals for an age group who may benefit from fertility assistance programmes.

      • P104 Method development and reference interval validation for male testosterone using supported-liquid extraction and tandem mass spectrometry

        L Lewis (Department of Clinical Biochemistry, Heartlands Hospital, Birmingham, United Kingdom), C Webster (Department of Clinical Biochemistry, Heartlands Hospital, Birmingham, United Kingdom)

        In males, testosterone has many biological actions, including development of sexual organs, muscle mass, libido, mood and memory. Low or absent testosterone can therefore result in a multitude of effects from decreased strength and endurance to depression, decreased sperm production and more. In order to correctly and accurately diagnose deficiency it is important to know what constitutes a “normal” level of this hormone.

        This project involved the collection of blood samples from willing male volunteers who were deemed likely to have normal levels of testosterone. This involved gaining informed consent and having potential participants complete a questionnaire based on the quantitative Androgen Deficiency in the Aging Male (qADAM) questionnaire widely used in the literature. Concurrently, the development and validation of an LC-MS/MS method for total testosterone measurement using supported-liquid extraction took place. This method was then used to derive a testosterone reference interval from the blood samples collected from volunteers (and therefore specific to the population served by Heart of England NHS Foundation Trust) and also included sub-group analysis of different age and ethnic populations.

        The developed method has acceptable accuracy (EQA bias -1.1%; recovery 97.9%), precision (5.6 and 6.2%, at two levels) and linear range (1-80 nmol/L) as well as good stability and limit of quantitation (0.95 nmol/L). It suffers from very little carry-over, interference or ion suppression. Serum samples from 145 participants were eligible for inclusion in the reference interval, which was determined to be 8.1-29.9 nmol/L. Age-specific and ethnicity-specific intervals were derived but low numbers in some groups and wide confidence intervals make them less reliable than the full reference interval.

      • P105 Interference in immunoassays

        WS Wassif (Department of Clinical Biochemistry, Bedford Hospital NHS Trust, Bedford, United Kingdom), L Ward (Department of Clinical Biochemistry, Viapath LLP, Bedford Hospital NHS Trust, Bedford, United Kingdom), J Sheffield (Department of Clinical Biochemistry, Viapath LLP, Bedford Hospital NHS Trust, Bedford, United Kingdom), N Wassef (Department of Clinical Biochemistry, Viapath LLP, Bedford Hospital NHS Trust, Bedford, United Kingdom)

        Immunoassays are intrinsically susceptible to interference. The frequency of such interference, its extent and the potential consequence of erroneous results, that may adversely affect clinical management, is rather difficult to gauge.

        As part of clinical authorisation, anomalous results which were not consistent with the clinical details (e.g. high FT4 in the thyrotoxic range associated with a normal or high TSH) were further investigated. Thyroid function tests (TFTs) are performed on the Roche Cobas 8000, using immunoassay that utilises biotin-streptavidin interaction, a potential target for interference. Fifty-four samples from 48 patients (35 female, 13 male, age range 4-91 years, mean±SD 49.6±25.9) with peculiar results that were not consistent with the clinical details, were referred to another laboratory for analysis using Siemens Advia Centaur XP to exclude potential interference. Method-related interference was noted more frequently with FT4 assay; 19 (35%) of 54 samples compared with TSH assay, six (12%) of 51 samples examined. Furthermore, interference was more pronounced with FT4 assay, with five (26%) of the 19 samples with grossly elevated FT4 (56, 59, 93, 97, >100 pmol/L) demonstrating essentially normal concentrations.

        Although interference is observed more frequently in TFTs, as anomalous results are picked up readily since TSH and FT4 are interpreted together, this issue is not restricted to these analytes. We reported a case of interference in testosterone assay when an initial result of >100 nmol/L was proved to be erroneous; true concentration of 7.9 nmol/L confirmed by testosterone extraction.

        In order to reduce misdiagnosis and inappropriate treatment, an interactive communication between physician and the laboratory is crucial in order to correlate hormonal assay results with patient clinical status. The laboratory should have a clear protocol to identify and investigate possible interference in TFTs and other endocrine assays.

      • P106 Saliva as an alternative medium to serum for measurement of cortisol after overnight dexamethasone suppression test

        G De Costa (Department of Clinical Biochemistry, King's College Hospital, London, United Kingdom), S Omar (King's College Hospital, London, United Kingdom), GF Cross (Department of Clinical Biochemistry, King's College Hospital, London, United Kingdom), R Ranasinghe (Department of Clinical Biochemistry, King's College Hospital, London, United Kingdom), H Abraha (Department of Clinical Biochemistry, King's College Hospital, London, United Kingdom), DR Taylor (Department of Clinical Biochemistry, King's College Hospital, London, United Kingdom), S Aylwin (Department of Endocrinology, King's College Hospital, London, United Kingdom), RP Vincent (Department of Clinical Biochemistry, King's College Hospital, London, United Kingdom)

        Introduction: The overnight dexamethasone suppression test (ONDST) is an initial screening test recommended by the Endocrine Society for suspected Cushing’s syndrome (CS). The test entails measurement of 9 am serum cortisol after 1 mg dexamethasone taken the previous evening. Midnight salivary cortisol is already established in the diagnostic algorithm and is a non-invasive sample. In this study we compared post-ONDST cortisol in saliva and serum to assess the diagnostic utility of saliva for this application.

        Method: Saliva samples (Salivette®, Sarstedt) were collected from adult patients who underwent ONDST between June 2013 and December 2015 for suspected CS. Pathology and hospital IT systems were used to obtain information on investigations, multi-disciplinary meeting (MDM) outcomes, clinical management and histology (where available). Samples were stored in Salivette® at -20°C, prior to analysis. Salivary cortisol was assayed by an ELISA method (Salimetrics® USA, LoQ 0.3 nmol/L, CV 7% at 1.7 nmol/L, linear up-to 82.77 nmol/L)

        Results: In total, there were 40 patients (17M) aged 50 (40-61) [median (IQR)] years. Biochemical cortisol excess was diagnosed in eight (3 ACTH-dependent) by the MDM. Histological diagnosis was obtained in three cases (one adrenal adenoma and two ACTH-positive pituitary adenoma). Salivary cortisol post-ONDST was higher 7.3 (3.4-10.0) nmol/L in those with cortisol excess vs. normal 1.2 (0.5-2.1) nmol/L (p<0.001). Cortisol excess (including low grade) was excluded if salivary cortisol was ≤1.7 nmol/L giving a sensitivity of 100% and a specificity of 69%.

        Conclusion: In our cohort at the proposed cut-off ≤1.7 nmol/L, the test ruled out cortisol excess. Thus, salivary cortisol post-ONDST is a suitable alternative to serum cortisol in the diagnostic work-up for suspected CS.

      • P107 The frequency of thyroid function testing in patients with type 2 diabetes mellitus

        C Duff (Department of Clinical Biochemistry, University Hospitals of North Midlands NHS Trust, Stoke on Trent, United Kingdom), T Palit (Department of Clinical Biochemistry, University Hospitals of North Midlands NHS Trust, Stoke on Trent, United Kingdom), A Heald (Department of Endocrinology, Salford Royal NHS Foundation Trust, Salford, United Kingdom), A Fryer (Department of Clinical Biochemistry, University Hospitals of North Midlands NHS Trust, Stoke on Trent, United Kingdom)

        Background: Current recommendations state that people with type 2 diabetes mellitus (T2DM) should have thyroid function checked at diagnosis, but that screening should not routinely be performed thereafter. However, the literature suggests an increased prevalence of thyroid dysfunction in the T2DM population and there is anecdotal evidence that regular screening is being performed in this population.

        Aims: To audit the frequency of thyroid function testing and determine the incidence of thyroid dysfunction in a cohort of patients with T2DM.

        Methods: Using data from the laboratory information management system, 306 individuals with T2DM and no existing thyroid disorder were identified. Thyroid function test (TFT) data was collected on these patients from 01/01/2009 to 31/12/2017 (1901 test episodes). The frequency of TFT testing was examined and the incidence of abnormal results determined.

        Results: 96.4% of individuals had at least one TFT during the follow-up period. For consecutive pairs of tests performed in primary care, the mean interval between tests was 11.3 months (SD=8.0, range 0-81). When categorized by test interval in months, 12 months was the most frequently observed interval and a large proportion of intervals (39.4%) fell within the 10-14 month categories. During the 8 year follow-up period, 11.4% of patients had an abnormal TFT result. Of these, 66% were female.

        Conclusions: Despite the lack of guidance, the majority of people with T2DM were screened for thyroid dysfunction. Furthermore, a substantial proportion of this testing appeared to be scheduled in nature, possibly performed as part of the diabetes annual review. A significant proportion of patients had abnormal TFT results, suggesting that this testing strategy, though outside current guidance, may be warranted.

      • P108 Free thyroxine assays: how much can we trust the results? Disparities between TSH and free thyroxine

        A Hutchesson (Department of Laboratory Medicine, Royal Bolton Hospital, Bolton, United Kingdom), J Osborne (Department of Laboratory Medicine, Royal Bolton Hospital, Bolton, United Kingdom), G Wieringa (Department of Laboratory Medicine, Royal Bolton Hospital, Bolton, United Kingdom), C Williams (Department of Laboratory Medicine, Royal Bolton Hospital, Bolton, United Kingdom), D Darby (Department of Clinical Biochemistry, Salford Royal Hospital, Salford, United Kingdom)

        Following the introduction of Roche analysers, we observed discrepancies between TSH and free thyroxine (fT4) results in several patient samples. Samples appearing particularly discrepant were reanalysed by a neighbouring hospital using different methodology (Siemens). Of 150 samples, 100 (66.7%) showed a Roche fT4 result >20% above the mean for both methods. We derived method-specific log-linear relationships between fT4 and TSH for both the Roche assays and also for our previous analyser (Abbott Architect). Samples tested by the Siemens assay showed a similar relationship between TSH and fT4 to that seen in the Architect assay, irrespective of whether the Roche and Siemens fT4 results agreed or not. In contrast, those Roche fT4 results which disagreed with the Siemens result showed no relationship with TSH.

        Samples which showed reasonable agreement between Roche and Siemens fT4 values showed tight agreement between the two methods’ TSH results (r2=0.8689), with a mild positive bias for the Roche assay (median +6.67%, interquartile range +2.17% to +10.84%) on Bland-Altman analysis. However, the 100 samples which showed poor agreement for fT4 also showed poorer correlation between TSH results, with negative interference in the Roche (median -9.13%, interquartile range -32.22% to +3.80%) and/or positive interference in the Siemens assays (p<0.01, Mann-Whitney U test).

        In December 2017, we received 5826 requests for thyroid function, sent 17 to be checked and found significant discrepancy in the fT4 result in 12 (0.21% of all requests).

        This work confirms that interference in free hormone assays remains significant in a proportion of patients, with positive and negative interference in different assays; and illustrates the value of using physiological relationships to assess it.

      • P109 The impact of standardised advisory comments on appropriate follow-up of individuals with subclinical hypothyroidism

        K Langford-Smith (Clinical Biochemistry, University Hospitals of North Midlands NHS Trust, Stoke-on-Trent, United Kingdom)

        The UK Guidelines for the Use of Thyroid Function Tests (2006) advise that subclinical hypothyroidism (SCH) should be confirmed by repeat thyroid function testing (TFTs) within 3-6 months, and that the interval of monitoring in confirmed cases should depend on the result of thyroid peroxidase antibody (TPOA) testing. To support adherence to these guidelines, in April 2016 we introduced standardised comments reflecting this advice. The aim of this study was to determine whether these comments have improved appropriate follow-up for individuals with SCH. TSH/FT4 measurements made over a 7 day period in April 2017 were compared to a similar data set collected in June 2014. The impact on duty biochemist workload was as predicted, with a 1.6-fold increase in the number of TSH results held for review. The 6-month follow-up was assessed in 22 GP patients with a comment advising repeat testing in 3-6 months to confirm SCH, and 8 with confirmed SCH who were advised to measure TPOA. Compared to patients receiving results before introduction of the comments, no significant improvement in follow-up rates was observed. Repeat thyroid function testing was performed within 6 months for 37% of results with comments, and 44% without. TPOA testing was performed within 6 months in 25% of relevant individuals before and after introduction of the comments, although repeat thyroid function testing in this group increased from 38% to 63%. Therefore the advice provided has not had an impact on GP decision making. The reason for this is unknown and further engagement with GPs will be necessary in order to understand this and identify ways we can support appropriate follow-up in patients with SCH.

      • P110 An unexpected endocrine abnormality

        L Tooth (Department of Clinical Biochemistry, Imperial College Healthcare, London, United Kingdom), EL Williams (Department of Clinical Biochemistry, Imperial College Healthcare, London, United Kingdom)

        A 39 year old man presented to his GP with on-going complaints of sleep disturbance, tiredness and headaches. He had suffered a traumatic brain injury several months before where he had fallen down a flight of stairs at work on a construction site. He was not taking any regular medication, apart from analgesics for pain, and leading a fairly active lifestyle frequently cycling and going to the gym.

        On routine bloods he was found to have a TSH of 0.08 µ/L, with a FT4 of <5.2 pmol/L and FT3 of 3.7 pmol/L. His testosterone was grossly elevated at 85.3 nmol/L with a suppressed LH (<0.5 IU/L) and FSH (<0.5 IU/L). His prolactin was also raised at 437 µ/L with macroprolactin not detected. Further investigation of his androgenic status revealed androstenedione, DHEAS and oestradiol all within their reference ranges.

        On further examination he admitted to using Dermacrine, which is a topical cream used to boost male hormone levels, improving workout capabilities and general well-being. It is advertised as getting ‘the benefits of additional lean muscle mass and less body fat with the convenience of avoiding injections and harsh oral steroids or prohormones’.

        Our laboratory purchased some Dermacrine and found DHEA was listed on the ingredients list. DHEA can be converted to testosterone in the body. DHEA has also been linked to thyroid function, with decreased values seen in hypothyroidism, and increased values seen in hyperthyroidism. In rats DHEA has also been shown to reduce TSH, FT4 and FT3 concentrations.

        His endocrine abnormalities returned to normal after he stopped using the Dermacrine cream. This case highlights the importance of a detailed clinical history.

      • P111 Evaluation of two commercially available quantitative TSH receptor antibody immunoassays used in the diagnosis of Graves’ disease

        D Peiris (Blood Sciences, Basingstoke and North Hampshire Hospital, Basingstoke, United Kingdom)

        Background: Due to their involvement in Graves’ disease (GD) pathogenesis, thyroid-stimulating hormone (TSH) receptor antibodies (TRAb) are measured to aid diagnosis. Current assays measure all TRAb subtypes: thyroid stimulating immunoglobulins (TSI), TSH-stimulation blocking antibodies (TSBAb) and neutral antibodies. However, two newly developed TRAb assays have been identified as being more specific and sensitive for GD diagnosis, using recombinant human TSH receptors (hTSHR): the Siemens IMMULITE 2000/2000 XPi TSI assay, which specifically measure TSI, and the Thermo Fisher Scientific Phadia 250 EliA anti-TSH-R assay; a competitive ELISA that detects all TRAb subtypes. This study aimed to evaluate the clinical utility of both methods in comparison to a current method in use: RSR’s TSHR Autoantibody Third Generation ELISA kit.

        Methods: For the three-way method comparison, 51 patients with TRAb results from the current RSR method were chosen by stratified sampling and random selection to include equal numbers of ‘negative’, ‘borderline positive’ and ‘positive’ TRAb patients. These samples were run once on both the Siemens and Thermo Fisher Scientific methods. Linearity, trueness and precision were all assessed.

        Results: At the clinical decision cut-offs, both the Siemens and Thermo Fisher Scientific methods demonstrated acceptable linearity, trueness and precision. There was good agreement (80.4%) between the RSR and Siemens methods across the range, and moderate agreement (52.9%) between the RSR and Thermo Fisher Scientific methods. With regards to clinical utility, the Thermo Fisher Scientific method and Siemens method performed similarly, with the latter performing slightly better: 80.8% and 88.5% concordance respectively. This was however only in the 26 out of 51 samples that have been reviewed thus far.

        Conclusion: Whilst both methods are analytically sound, full review of the samples will help to further define clinical utility and decide on whether there are any benefits of only measuring TSI as opposed to all TRAb subtypes.

      • P112 Random cortisol levels versus short synacthen test outcomes in adrenal insufficiency: are our cut-offs appropriate?

        E Smith (Department of Clinical Biochemistry, Southmead Hospital, Bristol, United Kingdom), P Thomas (Department of Clinical Biochemistry, Southmead Hospital, Bristol, United Kingdom)

        Background: The first-line test for adrenal insufficiency is a 9 am cortisol measurement, with our laboratory cut-off for excluding this being >350 nmol/L (Roche GenII assay). A short synacthen test (SST) is advised at ≤350 nmol/L. Following withdrawal of support for performing SSTs in the community, there is a need to review advice. This study aimed to compare initial cortisol levels with subsequent SST outcomes, to reassess whether our 350 nmol/L cut-off is appropriate.

        Methods: Data was collected retrospectively from 128 patients over 11 months (01/10/16 to 31/08/17), and initial cortisol levels compared to SST baseline cortisol levels and outcomes.

        Results: 12/128 SSTs were abnormal (post-synacthen cortisol ≤420 nmol/L). Of these, 11 patients had an initial cortisol ≤350 nmol/L (range 20-296 nmol/L) with 1 acutely unwell ITU patient having an initial cortisol of 420 nmol/L, and all patients from primary care having an initial cortisol of <200 nmol/L. Of the 116 patients with a normal SST, 15 had an initial cortisol of 300-350 nmol/L; 84 patients (72%) had an SST baseline cortisol higher than their initial random cortisol; 32 (38%) increased to >350 nmol/L. All patients with a sub-optimal SST had a SST baseline cortisol <250 nmol/L; all primary care patients had a SST baseline cortisol <150 nmol/L. Considering only results where initial cortisol samples were collected at an appropriate time (6 am-12 pm), 71/100 patients (71%) with a normal SST had an increased SST baseline cortisol compared to their initial cortisol; 28/71 (39%) increased to >350 nmol/L.

        Conclusions: Lowering our cortisol cut-off for advising SSTs in GP requests to ≤250 nmol/L, or initially advising a repeat cortisol for patients, has the potential to reduce unnecessary SSTs by approximately 40%. However there is still a clear need to advise SSTs in acutely unwell patients or those with a high index of suspicion.

      • P113 Phenoxybenzamine concentrations in breast milk are much higher than expected

        J Jeffery (Derriford Combined Laboratory, Derriford Hospital, Plymouth, United Kingdom), B Flower (Department of Nephrology, Derriford Hospital, Plymouth, United Kingdom), R  George (Derriford Combined Laboratory, Derriford Hospital, Plymouth, United Kingdom) P Rowe (Department of Nephrology, Derriford Hospital, Plymouth, United Kingdom)

        Aim of Study: Phaeochromocytoma in the peripartum period is rare and as a consequence there is little data on the concentrations of alpha-adrenoceptor blocking drugs in breast milk and the risks that this may pose to a neonate. The British National Formulary states that phenoxybenzamine may be present in breast milk, but there is no published data in support of this statement. Side-effects associated with phenoxybenzamine include hypotension and respiratory depression in newborn infants. A patient presented to Derriford Hospital with a diagnosis of phaeochromocytoma in pregnancy where attempted localisation and resection of the tumour during a caesarean section operation was unsuccessful and she continued to take phenoxybenzamine whilst breast feeding. Phenoxybenzamine concentrations were measured in samples of plasma and breast milk at 3 different time points; pre-dose, 1 h and 6 h post-dose on three consecutive days using a liquid chromatography mass spectrometry method.

        Methods: The patient was prescribed 60 mg phenoxybenzamine bd. Concentrations of this drug were measured in EDTA plasma and breast milk samples by a commercial laboratory using the method outlined below; sample extraction involving protein precipitation with acetonitrile containing the internal standard dextromethorphan, an additional hexane extraction (for breast milk samples), evaporation and reconstitution in aqueous methanol followed by quantification of phenoxybenzamine using liquid chromatographic separation on an ACE 5 CN 75x2.5 mm HPLC column and tandem mass spectrometry detection of the phenoxybenzamine and dextromethorphan using an Applied Biosystems Sciex 3000 mass spectrometer.

        Results: Concentrations of phenoxybenzamine ranged from undetectable to 6.2 ng/mL in plasma samples and from 90 to 740 ng/mL in breast milk samples.

        Conclusions: Phenoxybenzamine is excreted in significant amounts in breast milk, with post-dose concentrations over 100 times that of plasma. This may subject a breast-fed neonate to a risk of side-effects from the drug.

      • P114 Aldosterone:renin ratio for screening of primary hyperaldosteronism using Liaison in Asians: need for population and method specific cut-offs

        L Ong (Department of Laboratory Medicine, National University Hospital, Singapore), S Saw (Department of Laboratory Medicine, National University Hospital, Singapore), S Sethi (Department of Laboratory Medicine, National University Hospital, Singapore), WL Cheng (Department of Laboratory Medicine, National University Hospital, Singapore), N Siew Fong (Department of Laboratory Medicine, National University Hospital, Singapore), C Yin (Department of Laboratory Medicine, National University Hospital, Singapore), S Chew (Department of Laboratory Medicine, National University Hospital, Singapore)

        Background: The advent of automated renin concentration measurements caused confusion amongst our endocrinologists with regards to appropriate aldosterone: renin (ARR) cut-offs. We evaluated the Liaison direct renin concentration (DRC) and aldosterone to determine the need for population specific ARR cutoffs. We also evaluate risk of cryoactivation for DRC.

        Methods: We compared aldosterone and DRC measured on Liaison against aldosterone measured on radioimmunoassay (ZenTech, Belgium) and plasma renin activity (PRA) measured on Liaison (Diasorin, Italy) for samples requested between September 2016 and January 2018. A positive screen for primary hyperaldosteronism was defined by ARR >550 and aldosterone >415 pmol/L. Cryoactivation studies for DRC were performed with 3 healthy volunteers with 3 EDTA tubes collected in ice, 3 collected at room temperature (RT), and kept at varying durations at RT before analysis.

        Results: Liaision aldosterone concentrations correlated well with ZenTech results (r2=0.9), while DRC correlated fairly with PRA (r2=0.8). Seventy patients (median age 55.5 years old, male: female ratio of 36:34, Chinese: Malay: Indian: other ethnic groups ratio of 53:9:2:6) screened for secondary hypertension were included. Based on receiver operator characteristics, the best sensitivity (100%) and specificity (96.8%) were at aldosterone:DRC ratio of 35.6 and aldosterone concentrations >369 pmol/L.

        Compared to DRC measured immediately in EDTA tubes drawn at RT, samples kept in ice showed a mean +10% bias. For plasma samples frozen immediately and thawed before analysis, thawing at 370C or RT did not make a significant difference.

        Discussion and conclusion: Our ARR differed from international guidelines (45.6 in Endocrine Society guidelines; 100 in Australian guidelines) suggesting there may be population or ethnic related factors. Limiting factors include small sample size and further studies with larger sample size and looking at ARR in different ethnic groups may be helpful.

      • P115 The UK Acromegaly Register: centralised vs. local measurement of insulin-like growth factor-I

        P Monaghan (Blood Sciences Department, The Christie Pathology Partnership, Manchester, United Kingdom), P Trainer (Endocrine Unit, The Christie NHS Foundation Trust, Manchester, United Kingdom)

        Introduction: Acromegaly is a rare disease with a prevalence estimated at 40 per million. It is apparent that the condition is associated with increased morbidity and mortality. The UK Acromegaly Register was established to provide sufficient patient numbers to address key epidemiological and therapeutic issues. The project has been adopted by the UK Clinical Research Network (UKCRN ID 6687).

        Methods: In order to eliminate assay variation between centres, this project has undertaken centralised measurements of insulin-like growth factor-I (IGF-I) to enable better comparison of patient outcomes. Centralised measurement of IGF-I was performed on the Immunodiagnostics Systems (IDS) iSYS automated immunoassay analyser with the assay calibrated and traceable to WHO International Standard 02/254.

        Results: Twelve local centres participated in this study; 251 paired measurements (centralised vs. local) for IGF-I were collected based on 4 different analytical platforms. Passing-Bablok linear regression showed a mixed systematic proportional (slope 1.21) and constant bias (intercept -15.14), Pearson’s correlation coefficient, r2 0.94, 95% CI 0.92–0.95 (p<0.0001). Bland Altman analysis showed agreement between centralised vs. local IGF-I; bias 9.7% [23.1 ng/mL]. Comparison of results interpretation based on centralised vs. local reference intervals revealed good qualitative agreement; 99% of results being classified within the reference intervals and 93% classified as elevated above the reference intervals. However, only 20% of the results reported as below the reference intervals (low) by the local centres were classified as such on centralised measurement. The local centres reporting these results as low were using different assay platforms compared to centralised testing.

        Conclusion: Clinical decisions on modification of treatment in acromegaly would be consistent based on centrally or locally determined results and furthermore validates that laboratory quality management systems, including EQA at all centres is fit for purpose.

      Gut Nutrition Trace Elements

      • P116 Case epidemiology from the first 3 years of laboratory-based surveillance for children with elevated blood-lead concentrations in England, 2014-17

        B Sampson (Department of Metabolic Medicine, Imperial College Healthcare NHS Trust, London, United Kingdom), D Roberts (Environmental Epidemiology Group, Centre for Radiation Chemical and Environmental Hazards, Public Health England, Chilton, United Kingdom), H Crabbe (Environmental Epidemiology Group, Centre for Radiation Chemical and Environmental Hazards, Public Health England, Chilton, United Kingdom), A Busby (North East & North Central London Health Protection Team, Public Health England, London, United Kingdom), S Reshat (North East & North Central London Health Protection Team, Public Health England, London, United Kingdom), T Owodunni (Environmental Epidemiology Group, Centre for Radiation Chemical and Environmental Hazards, Public Health England, Chilton, United Kingdom), H Gordon-Brown (Environmental Epidemiology Group, Centre for Radiation Chemical and Environmental Hazards, Public Health England, Chilton, United Kingdom), R Close (Environmental Epidemiology Group, Centre for Radiation Chemical and Environmental Hazards, Public Health England, Chilton, United Kingdom), G Dabrera (Consultant in Public Health, Public Health England, London, United Kingdom), R Ruggles (Consultant in Public Health, Public Health England, London, United Kingdom), G Leonardi (Environmental Epidemiology Group, Centre for Radiation Chemical and Environmental Hazards, Public Health England, Chilton, United Kingdom)

        Children incur lead toxicity even at low blood-lead concentrations (BLCs). Overt symptoms and signs appear late, accompanying higher BLCs, so timely diagnosis relies on clinical suspicion. Testing in England, however, is opportunistic. Most tests are performed by 6 supra-regional laboratories, with whom the Lead Poisoning in Children (LPIC) surveillance system (SS) was initiated in 2014 to ensure timely notification to health protection teams (HPTs) for case-management. We aimed to describe LPIC-case epidemiology and lead exposures, to inform opportunities for prevention of lead exposure in children in England.

        Surveillance population: children <16 years and resident in England during the reporting period June 2014-2017. Case definition: incident cases with BLC ≥0.48 μmol/L (10 μg/dL). System description: sentinel, passive laboratory-based SS. We extracted demographic and postcode data on lead exposure cases from HPZone (HPTs’ case-management system), and linked these with LPIC data using HPZone-number, and population data and socio-economic status (SES) using postcode. Households were defined by cases with unique postcode and surname. We described case demography, notification incidence, and BLCs. We also retrospectively obtained exposure data from HPTs, by online survey, for incident and prevalent cases reported March 2014-15.

        There were 85 incident cases in 79 households. Regional 2014-17 notification rates ranged from 0.39 to 8.83 per million 0-15 year-olds (nationally 2.72 per million); 59/81 (73%) cases were male, 49/85 (58%) were aged 1-5 years, 61/85 (72%) cases lived in the 2 lowest SES-quintile areas. BLC was <1.93 μmol/L in 78/82 (96%) cases. Exposure data for 31 cases revealed 18/31 (58%) had pica; exposure was linked to paint in 6/31 (19%), soil in 5/31 (16%).

        A heterogeneous notification rate alongside BLCs most typical of subclinical toxicity suggests variable clinician awareness and testing practice. Interventions should prioritise recognition/testing and prevention in higher-risk populations (under-fives, children with pica, case siblings) and settings (with ingestible lead hazards, deprived areas).

      • P117 Evaluation of sample stability for a quantitative faecal immunochemical test and comparison of two sample collection approaches

        M de Ferrars (Department of Clinical Biochemistry, York Teaching Hospitals NHS Foundation Trust, York, United Kingdom), S Mellen (Department of Clinical Biochemistry, York Teaching Hospitals NHS Foundation Trust, York, United Kingdom), C Chapman (Department of Clinical Biochemistry, York Teaching Hospitals NHS Foundation Trust, York, United Kingdom), S Hoggart (Department of Clinical Biochemistry, York Teaching Hospitals NHS Foundation Trust, York, United Kingdom), D Turnock (Department of Clinical Biochemistry, York Teaching Hospitals NHS Foundation Trust, York, United Kingdom)

        Background: Faecal immunochemical testing (FIT) is increasingly being used to triage symptomatic patients for suspected colorectal cancer. However, there are limited data on the effect of pre-analytical factors on faecal haemoglobin (f-Hb) when measured by FIT. The aims of this work were (a) to evaluate the effect of different sample storage conditions on f-Hb results in patient samples, and (b) to compare f-Hb stability in samples extracted by patients and sampled by the laboratory.

        Methods: A quantitative immunoturbidometric method (HM-JACKarc) was used to measure f-Hb. (a) Six patients provided faecal samples for f-Hb which were sampled at baseline and 1, 2, 3 and 7 days post collection after storage at room temperature or 4°C. (b) A total of 137 patients collected both a sample for f-Hb using a collection device and provided a faecal sample for laboratory f-Hb sampling. Results were compared categorically and discrepancies in categorisation were assessed against clinical outcome.

        Results: (a) f-Hb concentration declined rapidly within a day of storage at room temperature but results remained ≥10 µg Hb/g faeces in 5/6 patients after two days. Higher f-Hb concentrations were obtained in samples stored at 4°C compared to room temperature at day 3 and day 7 post-collection. (b) Results obtained from patient sampled faeces were significantly different to results from laboratory sampled faeces.

        Conclusion: There is considerable heterogeneity in f-Hb sample stability. Samples should therefore be collected rapidly into collection devices to prevent degradation and false negative results. Patient use of collection devices may lead to false positive results compared to laboratory sampling at a cut-off of 10 µg Hb/g faeces.

      • P119 Verification of a faecal bile acids kit: a new way to diagnose bile acid diarrhoea?

        L Hughes (Department of Clinical Biochemistry, Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), C Ford (Department of Clinical Biochemistry, Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), R Gama (Department of Clinical Biochemistry, Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), M Brookes (Department of Gastroenterology, Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom)

        Introduction: Irritable Bowel Syndrome (IBS) is a common condition, estimated to affect 11.2% of people globally. Whilst not life-threatening, it can significantly diminish the quality of life for sufferers. However, between 16.9% and 50% of patients diagnosed with diarrhoea-predominant IBS (IBS-D) instead have bile acid diarrhoea (BAD). BAD is caused by decreased reabsorption of bile acids in the intestine, leading to increased bile acid excretion in the faeces. This misdiagnosis is caused by lack of easily available diagnostic assays. Immunodiagnostik (IDK) have developed an enzymatic kit designed to measure total faecal bile acids (FBA) on a random stool sample. The aim of the study was to verify the kit, to determine whether it was suitable for use in a study assessing biomarkers of BAD against the current gold-standard radiological SeHCAT test.

        Method: Verification was performed on routine post-analytical stool samples. Imprecision, recovery, linearity, limit of blank (LoB), limit of detection (LoD) and comparison studies with the IDK reference laboratory were performed. Acceptable performance was assessed by comparison to performance measures in the kit insert.

        Results: LoB and LoD obtained were 0.03 and 0.055 µmol/g. The assay was linear between 0.24 and 20.5 µmol/g. FBA’s were stable at -20oC over a period of 6 weeks. No significant difference was seen between results obtained at New Cross Hospital and the IDK reference laboratory (t=0.96, p=0.36). Average recovery of cholic acid spiked into stool extract was 92.2%. Inter-batch extraction CV was 11% at low FBA level and 17.7% at high levels.

        Conclusion: The assay meets acceptable criteria for use in measuring total FBA in a random stool sample. Further work is required to determine a suitable reference range, and to assess whether a random stool sample provides an accurate reflection of overall faecal bile acid excretion.

      • P120 Hair loss? Not worth testing your zinc over, unless you've had a bariatric surgery!

        S McCann (Clinical Biochemistry, Stepping Hill Hospital, Stockport, United Kingdom), N Perry (Clinical Biochemistry, Stepping Hill Hospital, Stockport, United Kingdom)

        Received zinc requests are vetted ensuring they are clinically indicated, have not been assessed in the previous fortnight and that the albumin is not very low (<25 g/L) prior to being referred externally for analysis. It had been noticed that around 20% of all requests were on female patients with hair loss. We wanted to find guidelines for what tests to assess in these patients and assess our practice to this. Up-to-Date and NICE CKS was searched for ‘female hair loss’ which suggested considering checking FBC, TFT and ferritin and where there were features of androgen excess (irregular menstrual cycle, hirsutism and acne) testing androgens. The British Obesity & Metabolic Surgery Society (BOMSS) guidelines stated zinc should be monitored annually in people with gastric bypass/duodenal switch and tested as needed if unexplained anaemia, hair loss or changes in taste occur.

        We audited our 593 zinc requests from 08/2015 to 08/2017 of which 105 (17.7%) stated ‘hair loss’. Of these 105, 90 (85.7%) had FBC, ferritin and TFT tested meeting the guidelines, whereas 1 (0.95%) mentioned gastric bypass surgery to support zinc testing.

        Of the 105 samples tested, 7 had results below the reference range (10-21 µmol/L); 5 results were >9.2 µmol/L which were not deemed significantly abnormal. One patient (Zn 7.6 µmol/L) had a gastric bypass 8 years earlier and was appropriately managed with replacement. The last patient, a 2 year old girl (Zn 7.3 µmol/L) had hair loss following scarlet fever and did not attend her hospital referral or have a repeat test was diagnosed as Telogen effluvium. Zinc requests are only processed in cases of hair loss if there is a history of bariatric surgery and this has now been communicated to our clinical users and should save the lab around £600 annually.

      • P121 Evaluation of BÜHLMANN CALEX® Cap stool extraction devices for the extraction of faecal calprotectin

        S Kaur (Department of Clinical Chemistry, Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), C Ford (Department of Clinical Chemistry, Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), R Gama (Department of Clinical Chemistry, Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom)

        The gold standard method for faecal calprotectin extraction is the manual weighing method. This method however is time consuming and prone to human error. To improve the extraction phase and throughput of the faecal calprotectin assay, Royal Wolverhampton NHS Trust evaluated a commercial alternative to the manual weighing method known as BÜHLMANN CALEX® Cap stool extraction devices. The performance of the CALEX® devices was assessed and compared to the gold standard manual weighing method.

        A comparison was performed by extracting 67 homogenised stool samples (ranging from faecal calprotectin <20 to >1932 µg/g) using both the CALEX® devices and the manual weighing method. Stool samples with low, medium and high concentrations of faecal calprotectin were homogenised, aliquoted and extracted using both extraction methods 10 times over different days to calculate inter-batch imprecision. Stability was assessed by storing 3 samples extracted with CALEX® devices at 4oC for 14 days and faecal calprotectin was measured in the extracts on days 0, 5, 8 and 14 days post-extraction. Extracts were analysed using BÜHLMANN fCAL® turbo reagent on an Abbott ARCHITECT c16000 platform.

        CALEX® devices showed good agreement to the manual weighing method. Results for CALEX®-extracted external quality assurance samples also compared better to the method laboratory trimmed mean than those extracted using the manual weighing method. Inter-batch imprecision using the CALEX® devices was not significantly different to the weighing extraction method: 19.6% and 19.4%, respectively. Stability studies showed that faecal calprotectin was not significantly affected by storage at 4oC over 5 days.

        In conclusion, CALEX® devices demonstrate similar imprecision and accuracy to the gold standard manual method for extracting faecal calprotectin. Although more costly, CALEX® devices are easy to use and offer a quicker extraction process compared with manual weighing and will therefore enable a faecal calprotectin service to increase its throughput.

      • P122 Plasma citrulline and parenteral nutrition dependency in small bowel resection versus intestinal graft verse host disease/colitis

        RNK Ranasinghe (Department of Clinical Biochemistry, King's College Hospital NHS Foundation Trust, London, United Kingdom), S Tentis (Department of Gastroenterology, King's College Hospital NHS Foundation Trust, London, United Kingdom), B Holt (Department of Nutrition & Dietetics, King's College Hospital NHS Foundation Trust, London, United Kingdom), K Friend (Department of Nutrition & Dietetics, King's College Hospital NHS Foundation Trust, London, United Kingdom), M Houghton (Department of Nutrition & Dietetics, King's College Hospital NHS Foundation Trust, London, United Kingdom), G de Costa (Department of Clinical Biochemistry, King's College Hospital NHS Foundation Trust, London, United Kingdom), A Aggarwal (Department of Pharmacy, King's College Hospital NHS Foundation Trust, London, United Kingdom), M Mahmood (Department of Pharmacy, King's College Hospital NHS Foundation Trust, London, United Kingdom), P Dubois (Department of Gastroenterology, King's College Hospital NHS Foundation Trust, London, United Kingdom), RP Vincent (Department of Clinical Biochemistry, King's College Hospital NHS Foundation Trust, London, United Kingdom)

        Introduction: Plasma citrulline concentration can be used as a non-invasive biomarker of small bowel functional capacity, hence dependency of parenteral nutrition (PN) in short bowel syndrome (SBS). In SBS (residual small bowel length (rSBL) <100 cm), long-term PN dependency threshold is considered as ≤20 µmol/L. We compared the clinical utility of citrulline to assess PN dependency in SBS and intestinal graft verse host disease (iGvHD)/colitis.

        Method: This retrospective audit reviewed adult inpatients, who required PN for SBS (with documented rSBL) and histology proven iGvHD/colitis between January 2015 and December 2017. Pathology and hospital IT systems were used to obtain relevant data. Citrulline was measured using ion-exchange chromatography with post-column ninhydrin derivatisation and photometric detection of amino acid-ninhydrin compounds at 440 nm and 570 nm. LoQ was 2 µmol/L with a 5.3% CV at 27 µmol/L.

        Results: In total there were 13 (4M, eight small bowel resections and five iGvHD/colitis) patients, aged 52(29-71) [median(range)] years. All had eGFR >50(52->90) mL/min. PN contains precursors for glutamate, glycine and proline. rSBL was <100 cm (citrulline 6-21 µmol/L) in six and <130 cm (2,21 µmol/L) in two. All received long-term/home PN except one with 70 cm (11 µmol/L) who was weaned off PN after 64 days. Correlation between rSBL and cirtrulline was poor r=0.37 (-0.45-0.85) (p=0.36). The iGvHD/colitis patients had citrulline ≤21 µmol/L but, received PN only during hospital stay (average 31 PN days). Seven (54%) and four (31%) patients had elevated glycine and glutamate respectively. None of the other amino acids correlated with rSBL or long-term PN in this small study.

        Conclusion: In our cohort, citrulline ~21 µmol/L was a strong indicator of long- and medium-term PN dependency in SBS and iGvHD/colitis respectively. Thus, citrulline as a marker of bowel functional capacity has a useful clinical utility in consideration of PN administration.

      • P123 Development of a simple and accessible method for faecal bile acid measurement

        A Sanders (Department of Clinical Biochemistry, Dudley Group NHS Foundation Trust, Dudley, United Kingdom), AW Usmani (Department of Clinical Biochemistry, Dudley Group NHS Foundation Trust, Dudley, United Kingdom), H Ashby (Department of Clinical Biochemistry, Dudley Group NHS Foundation Trust, Dudley, United Kingdom), P Mohammed (Department of Clinical Biochemistry, Dudley Group NHS Foundation Trust, Dudley, United Kingdom)

        In conditions where the terminal ileum is damaged, bile acid (BA) reabsorption can be compromised. This bile acid malabsorption (BAM) means BA enter the colon where they stimulate electrolyte and water secretion leading to watery diarrhoea, bloating and incontinence. The SeCHAT test, involving the ingestion of 75Se-homocholic acid taurine followed by 7 day monitoring for its loss through faeces is the usual test for diagnosis. Whilst other tests exist, they are also relatively inaccessible and suggested as the reason many of the estimated 1% of UK population with BAM remain undiagnosed. Here we aim to demonstrate how a commercially available BA kit can be adapted for use on a routine chemistry analyser for a quick and cheap means of measuring faecal BA.

        BA measurement was performed on the Roche c702 module using Diazyme Laboratories reagents. Mass of faeces, concentration of methanol and HCl, along with varying incubation time and temperature were evaluated for effect on yield of BA. Replicate extraction and BA measurement in 2 faecal specimens over 5 days was used to estimate precision. Specimens were analysed by the Diazyme-Roche method were compared to the commercially available IDK® Bile Acids Test for faeces (Immunodiagnostik,AG).

        The optimised extraction conditions were 0.1 g faeces into 5 mL extraction solution (50%(v/v) methanol, 40%(v/v) saline (0.9%w/v), 10%(v/v) 12M HCl), vortex 1 min, then 5 min centrifugation at 3500 rpm. Varying reaction time and temperature had little effect. Pool1 mean=94 umol/L, CV=11.33%; pool2 mean=89 umol/L, CV=2.81%. The Diazyme-Roche results compared well with the Immunodiagnostik kit with a linear fit of y=0.9005x+0.0067, r²=0.8352.

        We have demonstrated that following a quick extraction step, routine analysers can be used for faecal BA measurement. With further clinical validation it has the potential to provide a quick and accessible screen for BAM.

      • P124 Clearance of gadolinium from the body following multiple magnetic resonance imaging scans

        K Raja (Clinical Biochemistry, Viapath Analytics, King's College Hospital, London, United Kingdom), C Gordon (Clinical Biochemistry, Viapath Analytics, King's College Hospital, London, United Kingdom)

        Background: Gadolinium-based contrast agents are often used during MRI scans to enhance signal and improve the visibility of internal body structures/vessels. These contrast agents are bonded to ligands to form stable complexes and thus avoid gadolinium toxicity and facilitate clearance. Although the gadolinium is believed to be rapidly excreted renally, several recent reports have indicated that gadolinium accumulates in the body even in patients with adequate renal function. Patients may thus experience more widespread effects, with neurological and musculoskeletal consequences reported. Retention is higher in patients/individuals receiving multiple scans. There is currently no consensus regarding clinical significance and treatment of gadolinium accumulation.

        Aim and Methods: Following a direct approach, gadolinium concentrations were measured in urine samples obtained from a patient who had exposure to a contrast medium (Gadovist) MRI every quarter for the last two years. Measurements were made in samples obtained pre- and post-chelation with the chelating agent Succimer (dimercaptosuccinic acid, DMSA). DMSA is commonly used in the treatment of lead, mercury and arsenic poisoning. Gadolinium was measured using inductively coupled plasma mass-spectrometry in standard mode and using rhodium as the internal standard. Concentrations were corrected for creatinine.

        Results: Gadolinium excretion was significantly elevated (~75-fold) in sample collected during chelation therapy with DMSA. Increased excretion was still evident two days after chelation therapy had been halted. Excretion seemed to normalise 9-10 d after chelation therapy.

        Conclusions: These measurements provide evidence of gadolinium build-up following MRI scans. However, it is unclear if the accumulated gadolinium is in free form (Gd3+) or associated with ligand(s). It is feasible that the accumulation could contribute to clinical complications.

      • P125 Use of calculated bound copper index as an internal quality control test

        B Sampson (Department of Metabolic Medicine, Imperial College Healthcare NHS Trust, London, United Kingdom), N Martin (Department of Clinical Biochemistry, NW London Pathology, London, United Kingdom), A Courtney (Department of Clinical Biochemistry, NW London Pathology, London, United Kingdom), N Unsworth (Department of Clinical Biochemistry, NW London Pathology, London, United Kingdom), L Seyani (Department of Clinical Biochemistry, NW London Pathology, London, United Kingdom)

        The primary copper binding protein in serum is caeruloplasmin (Cpl), which will bind almost all of the copper in serum. In Wilson’s disease there is a very low serum [Cpl] and a low [Cu]. There can be an increased amount of non-Cpl bound Cu, which is reflected in the increased urine Cu excretion seen in these patients. Although the increased [Cu] relative to [Cpl] can be reflected in the calculated free Cu index this cannot be used as a diagnostic indicator as there is imprecision in both assays which makes this unreliable. However, we show that using this calculation can highlight errors in either assay on the basis of individual samples and batches.

        Methods: Copper and Cpl measurements in the authors’ laboratory since 1997 have been included in the analysis. During this period Cu has been measured by flame atomic absorption (FAAS) analysis and by inductively coupled plasma mass spectrometry (ICPMS) using collision cell. Cpl was measured by immunonephelometry using a series of platforms. The laboratory participated in EQA with generally acceptable performance throughout this period. Free Cu index was calculated using the formula 100*([Cu]-47*[Cpl])/[Cu].

        Results: There have been minor changes in yearly average concentrations of Cu and Cpl over the period. There were 41400 samples with average [Cu] 19.33 µmol/L and [Cpl] 0.381 mg/L. Average free Cu was 6.4% (range -18235-88%). After exclusion of fliers the slope of [Cpl]v[Cu] was 46.7 and the 3SD range was -15 to 30%. For samples collected since 2015 the slope was 48.7 and the 3SD range was -11 to 30%. There are changes in the running mean of the calculated free Cu during the year which can be compared to IQC performance for both assays.

        We conclude that calculated free Cu can be used to detect analytical errors and changes in assay performance.

      • P126 Application of laser ablation-ICPMS to detect implant derived metal distribution in tissues from patients with orthopaedic implants

        B Sampson (Department of Metabolic Medicine, Imperial College Healthcare NHS Trust, London, United Kingdom), I Osman (Department of Clinical Biochemistry, NW London Pathology, London, United Kingdom), I Swiatkowska (Institute of Orthopaedics and Musculoskeletal Science, University College, Stanmore, United Kingdom), A Hart (Institute of Orthopaedics and Musculoskeletal Science, University College, Stanmore, United Kingdom)

        Wear and corrosion of metal components from orthopaedic implants is recognised as a significant cause of implant failure and patient morbidity and has led to the issue of advice and product warnings from regulatory authorities.

        We have reported metal distribution in tissues from patients with metal-on-metal (MOM) and metal-on-plastic (MOP) implants using X-ray spectroscopy at the Diamond Light Synchrotron. Access to this resource is limited and it cannot be used for screening large numbers of samples.

        Laser ablation inductively coupled mass spectrometry (LA-ICPMS) is a technique for determining the spatial distribution of metals in tissues. The sample is scanned by a laser beam which ablates the sample into a stream of inert gas and into the ICPMS. The resulting data can be transformed into a 2-dimensional map of the sample. The disadvantages of this technique compared to synchrotron spectrometry are that the sample is destroyed by the process and that the spatial resolution is lower. Synchrotron spectrometry generally has a spatial resolution of 2-4 microns whereas LA-ICPMS has a resolution of greater than 10-15 microns.

        Hip capsule tissue and liver, heart and spleen described in previous reports using synchrotron spectrometry were analysed by LA-ICPMS. A quadrupole-based inductively coupled plasma mass spectrometer coupled to a laser ablation system was used to study distribution of Co and Cr in samples of cardiac, hepatic and splenic tissue.

        In synovial tissue from patients with MOM implants there is a general low level signal from Co59 seen in all samples whereas there are much more localised areas of Cr52 with a high signal. Distributions of metals in heart, liver and spleen from patients with MOP implants show lower levels of Cr and Co. These patients were not expected to show significant Cr and Co deposits. In all cases the Cr and Co were not co-located.

      • P127 Five year survey of blood and urine mercury analysis data

        E Fox (Specialist Laboratory Medicine, Leeds Teaching Hospitals, Leeds, United Kingdom), C Lippiatt (Specialist Laboratory Medicine, Leeds Teaching Hospitals, Leeds, United Kingdom), J Booth (Specialist Laboratory Medicine, Leeds Teaching Hospitals, Leeds, United Kingdom), B Hill (Specialist Laboratory Medicine, Leeds Teaching Hospitals, Leeds, United Kingdom)

        Background and aims: Indications for mercury testing include occupational exposure and signs of suspected toxicity such as tremor, paraesthesia, visual deterioration, and renal damage. The aim of this survey was to address the following questions: Who requested mercury testing and why? How many results were elevated and what were the causes? How many requests were from patients concerned about mercury amalgam fillings and what did the results show?

        Methods: All blood and urine results from the previous 5 years were gathered from the laboratory database. Request location, requesting consultant and reason for request were recorded.

        Results: A total of 281 blood mercury results and 86 urine mercury results were reported. The top 3 requesters were ophthalmology (n=65), neurology (n=29) and occupational health (n=19). Common clinical details were optic atrophy, neuropathy/paraesthesia and tremor. Just 6 requests originated from renal departments. No clinical details were available for 87 requests. Elevated blood mercury (>30 nmol/L) was found in 27 samples and elevated urine mercury (>15 nmol/24 h or >5.5 nmolHg/mmol creatinine) was found in 5 samples. These represented 11 individual cases: exposure to metallic mercury (n=1); exposure to mercury in topical cream (n=1); probable contaminated fish consumption (n=3); cause unknown (n=6). Mercury amalgam fillings were mentioned in 19 requests from 15 individual patients: the highest blood concentration was 14 nmol/L; the highest urine concentration was 4.5 nmolHg/mmol creatinine.

        Conclusions: Where details were available, the majority of requests referenced appropriate signs of mercury toxicity or occupational exposure. The low number of requests from renal teams should be explored. The commonest cause of elevated blood mercury was presumed to be the consumption of contaminated fish. Our data provided no evidence that mercury amalgam fillings increase blood or urine mercury concentration. Measurement for this reason should be discouraged.

      • P128 Early experience with the interpretation of dp-ucMGP: an indicator of vascular calcification

        M Bradley (Guy's, King's and St Thomas' School of Medical Education, King's College London, London, United Kingdom), R Gorska (Human Nutristasis Unit, Viapath, St Thomas' Hospital, London, United Kingdom), AEB Moore (Osteoporosis Screening & Research Unit, Guy's Hospital, King's College London, London, United Kingdom), G Hampson (Osteoporosis Screening & Research Unit, Guy's Hospital, King's College London, London, United Kingdom), DJ Harrington (Human Nutristasis Unit, Viapath, St Thomas' Hospital, London, United Kingdom)

        Matrix Gla protein (MGP) is a vitamin K-dependent protein with an essential role in regulation of bone and soft tissue calcification, and is the strongest inhibitor and reverser of arterial calcification currently known. Increased levels of uncarboxylated MGP reflect increased susceptibility for vascular calcification, and reduced vitamin K levels in the vessel wall.

        We measured the desphosphorylated-uncarboxylated form of MGP (dp-ucMGP). An automated IDS-iSYS assay based on chemiluminescent technology was used to begin to evaluate the clinical utility of this marker in patients with established vascular risk factors.

        We analysed a total of 34 plasma samples from patients with: type 2 diabetes mellitus (T2DM), (n=4); chronic kidney disease (CKD), (n=4); rheumatic disease (RD), (n=3); hypertension (HTN), (n=6); raised total cholesterol (>4 mmol/L), (n=11); and cardiovascular disease (CVD) (n=6). Additionally, we analysed 80 plasma samples from postmenopausal osteoporotic women aged ≥55 y, with low serum vitamin K1 concentrations ranging from 0.05-0.35 µg/L [fasting reference range; 0.17–0.68 µg/L].

        The MGP cut-off value <750 pmol/L was adopted from the manufacturer's reference range. It was calculated from 132 apparently healthy donors aged 18 to 59. We found dp-ucMGP levels above the cut-off in 22% of those with cardiovascular risk factors. We found a significant correlation between age (r=0.31, p<0.05), BMI (r=0.4, p=0.001) and dp-ucMGP, but did not find any relationship between vitamin K and dp-ucMGP in the osteoporosis cohort.

        Our results show more >750 pmol/L values in patients at high risk of CVD, and a positive correlation with age and BMI: both CVD risk factors. The vitamin K1 concentrations were below the reference range median; thus evaluation with a larger range is needed. Studies have shown that vitamin K supplementation reduces dp-ucMGP and improves vascular health. Monitoring dp-ucMGP levels may allow modulation of arterial calcification and CVD progression, particularly in those with increased age and BMI.

      Haematology

      • P129 Spurious but malignant hyperkalaemia

        RM Valentine (Department of Clinical Chemistry, The Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), R Gama (Department of Clinical Chemistry, The Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom), A Barkhuizen (Department of Haematology, The Royal Wolverhampton NHS Trust, Wolverhampton, United Kingdom)

        An 87 year old asymptomatic woman in primary care was found to have persistent hyperkalaemia in the absence of renal disease. This was therefore considered spurious, the likely result of delayed centrifugation from the time of sample collection. After discussion with the GP, the patient was further investigated. Blood was drawn sequentially into a Sarstedt lithium-heparin-gel tube, Sarstedt serum/z4 gel tube and a Sarstedt EDTA KE/2.7 tube. The lithium-heparin-gel tube was centrifuged immediately and the separated plasma potassium analysed for a renal profile (sodium, potassium, urea and creatinine) and bicarbonate. A full blood count (FBC) was analysed soon after collection from the EDTA sample tube. The serum gel tube was centrifuged after two hours and the separated serum was analysed for a renal profile.

        The plasma potassium was normal (4.1 mmol/L) but serum potassium was raised (5.4 mmol/L), therefore confirming spurious hyperkalaemia and excluding not only genuine hyperkalaemia but also spurious hyperkalaemia due to poor venesection technique, sample contamination, inadequate sample transport conditions, inappropriate centrifugation and delayed separation. The FBC identified thrombocythaemia (platelet count 1185 x109/L; reference range 150 - 450 x109/L) as the cause of the spurious hyperkalaemia and excluded familial pseudohyperkalaemia, a rare benign autosomal dominant condition of increased erythrocyte membrane permeability to potassium.

        The patient was referred for urgent haematological assessment and a diagnosis of JAK-2 positive essential thrombocythaemia was confirmed. After commencing three weeks of hydroxycarbamide treatment the serum potassium normalised following the resolution of her thrombocytosis. This case highlights the importance of excluding haematological disorders before attributing persistent spurious hyperkalaemia to other pre-analytical causes; we present an algorithm for the investigation of spurious hyperkalaemia. It should be made aware that spurious hyperkalaemia may be the sole presenting feature of malignant haematological disease.

      Immunology

      • P131 An audit of adherence to NICE guideline NG20: "Coeliac disease: recognition, assessment and management" in the North West region

        SL Hanton (Neuroscience Laboratories, The Walton Centre NHS Foundation Trust, Liverpool, United Kingdom)

        Background: Coeliac disease is a chronic inflammatory bowel condition, affecting approximately 1 in 100 people in the UK. Affected individuals may present with a variety of often non-specific signs and symptoms, making effective diagnostic procedures essential. In September 2015, the National Institute for Health and Clinical Excellence (NICE) published a guideline (NG20) for the recognition, assessment and management of coeliac disease, as a replacement for the previous guidance (CG86). NG20 contains specific information relating to requirements for laboratory investigations in suspected coeliac disease.

        Aim: The aim of the audit was to determine whether laboratories in the North West region have implemented the advice relating to laboratory investigations given in NG20.

        Methods: An audit questionnaire was circulated electronically to all clinical biochemistry audit leads in the North West region. Responses were collated and the data reviewed. Supplementary questions were circulated as appropriate.

        Results: Completed questionnaires were received from eight laboratories. Four perform coeliac screening in-house, with the remainder referring testing to external laboratories. Four laboratories were compliant with all of the criteria that applied (these varied depending on whether testing was performed in-house), while the remainder were compliant with 60-80% of the criteria. The main discrepancies were noted around measurement of total IgA and cut-offs for IgA deficiency; investigation protocols for paediatric patients also differed in some laboratories.

        Discussion:

        This audit shows that laboratory screening procedures for coeliac disease in the North West region are generally compliant with NICE guidance. Inconsistencies in cut-offs for total IgA deficiency may be method-related; although NICE recommends a cut-off of 0.07 g/L, EQA data indicates considerable inter-method variability that may render the suggested threshold inappropriate.

      Lipids

      • P132 Use of sterol/cholesterol ratio improves diagnostic performance of the absorption sterols in investigation of hypercholesterolaemia

        E Middling (Department of Blood Sciences, Royal Victoria Infirmary, Newcastle-upon-Tyne, United Kingdom), M Appleton (Department of Blood Sciences, Royal Victoria Infirmary, Newcastle-upon-Tyne, United Kingdom), D Neely (Department of Blood Sciences, Royal Victoria Infirmary, Newcastle-upon-Tyne, United Kingdom), A Bowron (Department of Blood Sciences, Royal Victoria Infirmary, Newcastle-upon-Tyne, United Kingdom)

        Background: Analysis of the plasma absorption sterols sitosterol, campesterol and cholestenol is performed in patients with suspected familial hypercholesterolaemia in whom no genetic cause has been identified in order to exclude hyperlipidaemia due to a defect in sterols absorption e.g. sitosterolaemia. Non-cholesterol sterols may be elevated in hypercholesterolaemia, making interpretation of the results difficult. Use of the sterol/cholesterol ratio may improve the diagnostic accuracy of the test.

        Aims: To review reference ranges for absorption sterols, to prepare reference ranges for sterol/cholesterol ratios and to assess whether the ratios improve diagnostic accuracy in patients with hypercholesterolaemia.

        Methods: Sterol profiles were measured plasma samples from 70 patients aged >18 years with no evidence of disorders of absorption or metabolism of cholesterol or other sterols (controls) and 15 patients with hypercholesterolaemia (serum total cholesterol >5.6 mmol/L) who were not on any lipid-lowering therapy. Sterols analysis was performed by hexane extraction and gas chromatography-mass spectrometry. Results from controls were used to derive reference ranges for each sterol and the sterol/cholesterol ratio. These were used to assess the results from the patient group and to determine the diagnostic performance of the test in patients with hypercholesterolaemia.

        Results: Reference ranges were: sitosterol <8.7 µmol/L, campesterol <13.7 µmol/L, cholestenol <12.5 µmol/L, sitosterol/cholesterol 0.21-1.90 µmol/mmol, campesterol/cholesterol 0.26-2.80 µmol/mmol, cholestenol/cholesterol 1.00-2.70 µmol/mmol. All hypercholesterolaemia patients had an increase in one or more absorption sterol; 87% had increased sitosterol. When the ratios were used, only 33% of hypercholesterolaemia patients had an increased sterol/cholesterol ratio and 20% had an increased sitosterol/cholesterol ratio.

        Conclusion: Absorption sterols are increased in patients with hypercholesterolaemia, making the investigation of non-FH causes of hypercholesterolaemia difficult. Use of the sterol/cholesterol ratio improves diagnostic accuracy of the test, avoiding falsely diagnosing patients with sterol absorption disorders.

      • P133 Hypertriglyceridaemic acute pancreatitis in a patient taking capecitabine chemotherapy: a case report

        A Lomas (Clinical Chemistry, Northern General Hospital, Sheffield, United Kingdom)

        Aim: To report a case of suspected capecitabine-induced hypertriglyceridaemia which led to acute pancreatitis.

        Background: Capecitabine, a pro-drug of 5-fluorouracil, is a chemotherapeutic agent used in the treatment of a variety of cancers. Hypertriglyceridaemia due to capecitabine is uncommon, and therefore may go under-diagnosed.

        Case Summary: The patient is a 42 year-old male, treated with capecitabine and oxaliplatin adjuvant chemotherapy for a splenic flexure adenocarcinoma (T3 N1 M0). Other than a fatty liver on ultrasound scan, he had no relevant past medical or family history. His body mass index was 31.8. Review of his historical biochemistry revealed persistently elevated triglycerides (earliest result: 5.3 mmol/L, highest level 14.9 mmol/L). He had never received lipid-modifying therapy.

        His baseline triglycerides before commencing capecitabine were 4.3 mmol/L. Three weeks after starting capecitabine, levels had risen to 38 mmol/L, and by five weeks were beyond our laboratory’s upper limit of quantification (>50 mmol/L), when he attended the emergency department with abdominal pain. A CT scan confirmed pancreatitis with no obstructive gallstones. The patient received atorvastatin 80 mg once daily and capecitabine was stopped. Around a week after these interventions, the triglyceride level had fallen to 5.6 mmol/L.

        Following this the patient was seen for the first time in the lipid clinic. Capecitabine had been restarted three days earlier, alongside atorvastatin. Clinic investigations revealed a new diagnosis of diabetes mellitus with an HbA1c of 52 mmol/mol. A blood test obtained one week after restarting capecitabine revealed that the triglycerides had increased again, this time to 17.8 mmol/L. A blood sample was collected for ApoE2 genotyping.

        A Naranjo Adverse Drug Reaction Probability Scale score of 8 suggests a ‘probable’ reaction to capecitabine.

        Conclusion: It may be beneficial to monitor lipid profiles in patients receiving capecitabine chemotherapy, especially in those who have high baseline triglyceride levels.

      Management

      • P134 Critical value reporting in a tertiary hospital: effectiveness of telephonic reporting

        T Pillay (Department of Chemical Pathology, University of Pretoria and National Health Laboratory Service, Pretoria, South Africa), B Chale-Matsau (Department of Chemical Pathology, University of Pretoria and National Health Laboratory Service, Pretoria, South Africa)

        Aims: Reporting of critical values is a regulatory requirement and an important quality indicator. Critical values are those that significantly deviate from the reference range and require prompt reporting to ensure effective patient management. Our main objective was to evaluate how effectively critical results data were handled in a tertiary hospital laboratory by assessing the efficiency of each sample handling step, which included collection, reception, registration, testing and results reviewing, using 60 minutes as a target.

        Methods: Retrospective critical values data were obtained from the National Health Laboratory Service (NHLS) archives of samples analysed at chemical pathology laboratory Tshwane Academic Division (TAD) between January and December 2015. The data were grouped according to patient location and further into specific units to assess which patient profiles were mostly affected. Analyses were performed for various sample handling steps including collection, receiving, registration, testing and review of values.

        Results: A total of 7197 critical values were generated. A significant time lapse was observed between collection and receipt of the samples. Patient data on the request form was registered as soon as possible across all patient locations. On average there was a prolonged interval (2±1.87 h) between testing and critical value reporting. More than 40% of critical values from emergency units were not reported within 60 minutes. Further stratification of the data showed that most samples from high care were received as soon as they were collected. A large proportion (67%) of test requests from orthopaedic ward were reviewed early.

        Conclusion: The study has demonstrated that owing to the high volumes of tests sent to the laboratory, telephone reporting of critical values cannot keep up with the required efficiency of critical value reporting suggesting that other forms of automated electronic reporting should be employed and sample transportation and processing should be improved.

      Molecular Genetics

      • P135 Hormonal and genetic biomarkers for predicting risk of perinatal depression

        E Frith (School of Life Sciences, Coventry University, Coventry, United Kingdom), D Grammatopoulos

        (Clinical Biochemistry Department, University Hospital Coventry and Warwickshire, Coventry, United Kingdom)

        Perinatal depression (PND) affects 1 in 7 women and is a major health burden. PND often remains undiagnosed, resulting in profound consequences for both mother and baby. The failure of the maternal hypothalamic-pituitary-adrenal (HPA) axis to re-equilibrate following delivery has been identified as a putative pathogenic mechanism. Abnormal levels of inflammatory mediators, neurotrophic factors and neuroactive hormones have been identified as potential contributors to this abnormal HPA activity, especially in individuals with increased genetic vulnerability due to specific molecular genotypes.

        In this cross-sectional study, depressive risk was determined at 24-28 weeks of gestation using the Edinburgh Postnatal Depression Scale (EPDS) for 100 pregnant women. Cut-off scores of >10 and >16 were used to classify women with increased risk for minor or major antenatal depression respectively. Single nucleotide polymorphisms (SNPs) of the glucocorticoid receptor (BclI), the corticotropin releasing hormone receptor 1 (rs242939 and rs242924) and the serotonin transporter (rs6354 and rs2020936) were determined by PCR and high resolution melting curve analysis. Circulating levels of interleukin-6, brain-derived neurotrophic factor, progesterone and its metabolite allopregnanolone levels were quantified using immunoassays.

        Significant associations were identified for allopregnanolone/progesterone ratio, with the mean ratio lower (p=0.04) in women at risk of major depression (0.14 vs 0.19) and for IL-6, with an increased concentration (0.87 vs. 0.59 pg/mL) correlating with scores ≥10 (p=0.0004). The ratio of BDNF/IL-6 was also significantly lower in patients with an EPDS score ≥10 (p=0.03). In patients with rs2020936, a SNP associated with depression through reduced functionality of the serotonin transporter, the correlation between IL6 and EPDS score was lost.

        These pilot findings suggest that a combined use of biochemical biomarkers and SNP-based molecular signatures might offer a screening strategy suitable for identifying women at risk of antenatal depression. Large cohort studies are required to confirm this.

      • P136 Is there a correlation between patient's single nucleotide polymorphism score and familial hypercholesterolemia Welsh score?

        F Palmer (Department of Clinical Biochemistry, University Hospitals Bristol NHS Foundation Trust, Bristol, United Kingdom), L Gritzmacher (Department of Clinical Biochemistry, University Hospitals Bristol NHS Foundation Trust, Bristol, United Kingdom), K Haralambos (Department of Medical Genetics, Cardiff and Vale UHB, Cardiff, United Kingdom), A Hills (Department of Blood Sciences and Bristol Genetics, North Bristol NHS Trust, Bristol, United Kingdom), M Williams (Department of Blood Sciences and Bristol Genetics, North Bristol NHS Trust, Bristol, United Kingdom)

        The Welsh score is frequently used to determine which patients undergo genotyping to confirm/exclude familial hypercholesterolaemia (FH) diagnosis. Clinicians calculate a patient’s Welsh score from lipid results, personal/family history of premature cardiovascular disease and examination findings. Scores: ≥6 = eligible for genotyping; <5 = only in exceptional circumstances. If no FH pathogenic variant is found a single nucleotide polymorphism (SNP) score is ranked in deciles which gives a risk of polygenic hypercholesterolemia: 1-3 (low), 4-5 (average), 6-10 (high).

        By reviewing genetic, clinical and biochemistry reports from UHBristol patients who underwent FH genotyping between July 2016 - October 2017, we investigated if there is a relationship between SNP and Welsh score.

        In 88 pathogenic variant negative patients: 21 (24%) had SNP 1-3 (Welsh score mean 3: range -4 to 8); 57 (65%) had SNP ≥6 (Welsh score mean 4: range 0 to 11). In 37 pathogenic variant positive patients: 8 (22%) had a SNP 1-3 (Welsh score mean 5: range 2-10); 14 (38%) had a SNP ≥6 (Welsh score mean 8: range 4-15).

        Results show a higher proportion of patients with SNP scores of ≥6 in FH negative vs. FH positive patients (64% vs 38%). Pathogenic variant positive patients had a higher mean Welsh score in those with a high SNP score compared to a low score (8 vs 5). However, there is no significant correlation between SNP and Welsh scores in either FH negative or positive patients (r² = 0.0142 and r² = 0.052 respectively).

        In conclusion, further work is required. Results indicate SNPs might alter the phenotype of pathogenic variant positive FH, with an increased number of LDL-C raising SNPs leading to a higher Welsh score, which reflects a more severe phenotype. The Welsh score relies on patient’s detailed understanding of their family history; this is often limited and may therefore explain the lack of correlation between SNP and Welsh scores.

      Oncology

      • P137 LC-MS/MS analysis of 5-fluorouracil in plasma for clinical research

        S Balloch (Scientific Operations, Waters, Wilmslow, United Kingdom), G Hammond (Scientific Operations, Waters, Wilmslow, United Kingdom), L Calton(Scientific Operations, Waters, Wilmslow, United Kingdom)

        There is a high degree of inter-individual variability for 5-fluorouracil (5-FU) metabolism. An accurate, analytically sensitive and specific method may play a role in researching the pharmacokinetic and pharmacodynamic effects of 5-FU administration.

        Matrix-matched calibrators (20-2000 ng/mL) and quality control samples (40, 350, 750 and 1500 ng/mL) were prepared using in-house stocks of 5-FU and pooled plasma. Samples (50 µL) were spiked with 5-FU-13C15N2 internal standard prior to liquid-liquid extraction using ethyl acetate. Run time was 3 minutes using a HSS PFP UPLC column (2.1 x 100 mm, 1.8 µm) on the Waters ACQUITY UPLC® I-Class FTN and a water/acetonitrile gradient. 5-FU was quantified using a Waters XEVO® TQD mass spectrometer operating in electrospray ionization in negative mode and multiple reaction monitoring detection mode.

        Analytical sensitivity was calculated to be 7.5 ng/mL (n=10 extractions, over five occasions, ≤20% CV). Linearity was demonstrated over the concentration range 14-2600 ng/mL and system carryover was negligible in samples ≤10000 ng/mL. Precision studies (n=5, over five occasions) demonstrated repeatability and total precision ≤9.0%. The mean recovery for 5-FU pooled plasma samples (n=3, 40 and 1500 ng/mL) was between 87.6-110.6 -% in the presence of high concentrations of endogenous and exogenous compounds. Qualitative ion suppression experiments, using post-column infusion, showed that 5-FU elutes in a region of minimal ion suppression using these conditions. Six individual plasma extracts were used in testing and compared to a solvent blank. Negligible matrix effects were observed at low (40 ng/mL) and high (1500 ng/mL) concentrations as indicated by respective mean internal standard adjusted matrix factors in the range 0.94-1.05.

        The clinical research method developed for the quantification of plasma 5-fluorouracil demonstrated good linearity, analytical sensitivity and precision with negligible matrix effects.

        For Research Use Only, Not for use in diagnostic procedures.

      • P139 Do Prostate Cancer Risk Management Programme (PCRMP) PSA limits really predict prostate cancer? A local retrospective biochemistry-histology correlation study

        J Newmark (Department of Laboratory Medicine, Royal Bolton Hospital, Bolton, United Kingdom), G Wieringa (Department of Laboratory Medicine, Royal Bolton Hospital, Bolton, United Kingdom), A Hutchesson (Department of Laboratory Medicine, Royal Bolton Hospital, Bolton, United Kingdom), C Williams (Department of Laboratory Medicine, Royal Bolton Hospital, Bolton, United Kingdom), P Waugh (Department of Laboratory Medicine, Royal Bolton Hospital, Bolton, United Kingdom)

        Aim: Prostate cancer is the most common male cancer in the UK, disproportionately affecting those >65 years (85% new diagnoses). Although national screening is not currently advocated, a risk management programme exists allowing men to request a serum prostate-specific antigen (PSA) test after counselling. PCRMP PSA limits, beyond which prostate biopsy should be considered, are ≥2.0, ≥3.0, ≥4.0 and ≥5.0 ng/mL for those 40-49, 50-59, 60-69 and 70-79 years respectively. These were later revised to ≥3.0 ng/mL for those 50-69 years, and variants of these limits are currently used across Manchester. We aim to assess how accurately the unrevised PCRMP PSA limits, adopted locally, predict prostate cancer within our population cohort.

        Method: We identified all men, aged 40-79 years, who underwent prostate biopsy, resulting in histology, between 1 January and 30 April 2017. Histology results, and the last PSA level requested by the referring GP, were correlated. We excluded patients ≥80 years, known pre-existing prostate cancer, and those without a relevant PSA level. Sensitivity and specificity were calculated, and ROC curve analyses were used to determine the diagnostic accuracy of the age-stratified PSA limits.

        Results: Ninety-three patients (1 patient 40-49 years, 16 patients 50-59 years, 33 patients 60-69 years, and 43 patients 70-79 years) were identified. For those 40-79 years, PSA had a sensitivity of 97.5% and specificity of 3.8% for prostate cancer. PSA failed to discriminate between malignant and benign pathology ≥70 years, with ROC curve analyses giving an AUC of 0.54 for 70-79 years versus 0.71 for 40-69 years.

        Conclusions: The majority of PSA requests, and new prostate cancer diagnoses, were in those 70-79 years; however, PSA did not discriminate between malignant and benign pathology in this cohort. Our results support current guidance from Public Health England, which does not recognise the utility of PSA-based referral ≥70 years.

      Proteins Enzymes

      • P140 Thiopurine methyl transferase: genotype-phenotype correlations (2014-2017)

        RL Griffiths (Department of Clinical Biochemistry, Sandwell & West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), L Eccles (Department of Clinical Biochemistry, Sandwell & West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), D Oakley (Department of Clinical Biochemistry, Sandwell & West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom), J Berg (Department of Clinical Biochemistry, Sandwell & West Birmingham Hospitals NHS Trust, Birmingham, United Kingdom)

        Introduction: The thiopurine methyl transferase (TPMT) gene exhibits genetic polymorphism, with 1:10 Caucasian adults having low activity (<68 mU/L), and 1:300 having complete deficiency (<10 mU/L). Patients with low/deficient TPMT activity are at high risk of haematopoietic toxicity from thiopurine medications; therefore treatment should be adjusted according to TPMT phenotype.

        At Sandwell & West Birmingham Hospitals (SWBH) we provide front-line TPMT phenotyping, with follow-up TPMT genotyping for common deficient alleles TPMT*2 and TPMT*3A/*3C, which account for 60-95% of mutant alleles. Here we review the correlation between TPMT phenotype and genotype over a 3-year period (2014-2017).

        Methods: Data was gathered for TPMT requests with both phenotype and genotype results. Results were reviewed to determine the distribution of TPMT activity and the frequency of reported genotype in each TPMT activity category.

        Results: A total of 6435 paired results were available; 3420 (53%) of samples had low TPMT activity, with 519 (8%) showing deficient TPMT activity. Of the deficient samples, 97% were identified to have at least one deficient allele, with 92% being homozygous or compound heterozygous. Fourteen samples had deficient activity but a wildtype genotype. For samples with TPMT activity <35 mU/L, 940 (77%) had at least one deficient allele, but 23% had a wildtype genotype. In this cohort of samples, 88% were documented to have a low haematocrit level. Of the 3015 samples with normal TPMT activity levels, 88% had a wildtype genotype, whereas 362 (12%) were found to have at least one deficient allele.

        Conclusion: Overall we have demonstrated excellent concordance between TPMT phenotype and genotype. However, there are a significant proportion of patients with either low/deficient activity and wildtype genotypes, or normal activity with one or more deficient alleles. This audit highlights that front-line TPMT phenotyping screen with genotype confirmation is the most appropriate screening strategy for patient safety.

      Renal Disease

      • P141 A new urine retinol binding protein assay for identification of tubular proteinuria in adults

        R Salota (Chemical Pathology, Epsom and St Helier University Hospitals NHS Trust, Carshalton, United Kingdom), M Dockrell (South West Thames Institute of Renal Research, Carshalton, United Kingdom), E Nabi (South West Thames Institute of Renal Research, Carshalton, United Kingdom), M Lapsley (Chemical Pathology, Epsom and St Helier University Hospitals NHS Trust, Carshalton, United Kingdom)

        Introduction: Retinol-binding protein4 (RBP) assays using polyclonal antibodies (pRBP) have major problems of non-linearity of dilution and a very small useable dynamic range. Our objective was to develop a specific assay with a wider dynamic range to detect tubular proteinuria.

        Materials and methods: Manual immunoassays, mRBP (monoclonal capture and second antibody with colorimetric detection) and fRBP, (polyclonal capture and monoclonal second antibody with fluorescence detection) were developed and compared with pRBP. Two hundred and ninety patient samples (62 with a history of kidney stones, 104 type 2 diabetes (53 no microalbuminuria and 51 microalbuminuria), 60 renal disease (TPCR ≥100 mg/mmol), 12 tubular proteinuria and 52 healthy controls) were analysed by mRBP. Thirty eight samples were tested with fRBP.

        Results: mRBP assay showed better linearity on dilution and wider analytical range than pRBP. There was poorer detection of the tubular component in patients with albuminuria and glomerular proteinuria and inadequate separation of patient groups with glomerular and tubular proteinuria. The fRBP assay also showed improved assay characteristics, with better sensitivity but also with good separation of patient groups.

        Intra-assay CV for fRBP was <6% between 15 and 323 µg/L, with inter-assay CVs of <8% using controls between 3.8 and 107 µg/L. The analytical range for fRBP was 2.3-600 µg/L. fRBP was linear on dilution within the analytical range using samples from patients with glomerular and tubular proteinuria. Correlation (r) was 0.8722 (95% CI 0.7621 to 0.9333, p< 0.0001), Wilcoxon matched-pairs signed-ranks test showed a significant difference (p<0.0001) between fRBP and pRBP results, with fRBP showing a mean negative bias of 247 μg/mmol.

        Conclusion: The combination assay with fluorescence detection proved more discriminatory than a purely monoclonal system especially in patients with significant proteinuria and has advantages of better linearity on dilution and wider analytical range than the existing pRBP assay, whilst maintaining good sensitivity.

      • P142 A case of hepatitis C virus negative cryoglobulinaemia

        L McGavin (Clinical Biochemistry Laboratories, Wishaw General Hospital, Lanarkshire, United Kingdom), J Price (Renal Medicine, Monklands Hospital, Lanarkshire, United Kingdom), J McGuire (Clinical Biochemistry Laboratories, Hairmyres Hospital, Lanarkshire, United Kingdom), I Godber (Clinical Biochemistry Laboratories, Monklands Hospital, Lanarkshire, United Kingdom)

        A 23 year old female was reviewed at renal outpatients for membranoproliferative glomerulonephritis (MPGN) caused by type II cryoglobulinaemia. She presented oedematous and nephrotic. Abnormal laboratory results included serum creatinine 130 µmol/L, eGFR 46 mL/min/1.73m2, serum albumin 23 g/L, and urine microalbumin:creatinine ratio 803.1 mg/mmol creatinine.

        She was diagnosed five years previously, following kidney biopsy after presenting with haematuria and proteinuria. There was no evidence of end-organ dysfunction caused by cryoglobulinaemia other than renal manifestations, though she had a history of vascular rash and Raynaud’s phenomenon. Investigations were negative for hepatitis B virus (HBV), hepatitis C virus (HCV) and human immunodeficiency virus (HIV). She was successfully treated with mycophenolate, which was stopped nine months prior to this presentation, and she was lost to follow-up.

        Her IgM kappa cryoglobulin was detected at 0.34 g/L, having previously been around 0.06 g/L (reference range <0.06 g/L), indicating that she manifested clinically with low cryoglobulin levels though these increased with worsening symptoms. Despite re-commencing treatment with mycophenolate and subsequent plasmapheresis, her condition deteriorated further. She started regular haemodialysis and was taken off all medications. Following a later illness, she received further plasma exchange and four weekly treatments with rituximab.

        Complement proteins (C3 and C4) and cryoglobulin levels are being monitored to determine the patient’s disease state. In monitoring her cryoglobulin levels, the mean concentration detected for this patient is 0.026 g/L. This information will be used to generate evidence of disease remission for transplant.

        The procedures for cryoglobulin sample collection, transport and quantification must be followed closely to reduce variability and false negatives. However, it is uncommon to monitor cryoglobulin levels.

        This is a case of HCV negative cryoglobulinaemia that caused a severe case of MPGN progressing to end-stage renal failure, and cryoglobulin measurement was used to monitor disease progression.

      • P143 Primary care requesting patterns for estimated GFR and compliance with NICE chronic kidney disease guidelines

        RN Ranasinghe (Department of Clinical Biochemistry, King's College Hospital NHS Foundation Trust, London, United Kingdom), A Simhadri (Department of Clinical Biochemistry, King's College Hospital NHS Foundation Trust, London, United Kingdom), N Rao (Department of Clinical Biochemistry, King's College Hospital NHS Foundation Trust, London, United Kingdom)

        Introduction: Chronic kidney disease (CKD) is a significant cause of mortality, morbidity and escalating health costs related to cardiovascular disease and renal replacement therapies. CKD is stratified according to estimated GFR (eGFR) and urine albumin creatinine ratio (ACR). In 2015, NICE updated their guidance (CG182) on identification and investigation of individuals with CKD including frequency of monitoring of eGFR in each CKD category. We compared frequency of monitoring of eGFR in primary care with NICE recommendation.

        Methods: This retrospective audit focused on adult requests from primary care to clinical biochemistry for eGFR and/or ACR between January 2017 to February 2017. Patients were categorised according to CKD grades based on eGFR and ACR categories: G1 >90 ml/min/1.73 m²; G2 60-89; G3a 45-59; G3b 30-44; G4 15-29; G5<15 and A1 <3 mg/mmol; A2 3-30; A3 >30 respectively. eGFR monitoring frequency was determined over a one-year period for each patient and compared with NICE recommendations. Serum and urine creatinine were measured by Jaffé method and albumin by immunoturbidometry (both Siemens Advia 2400). eGFR was calculated using the 4 variable modified MDRD equation.

        Results: A total of 117 patients (55 males) aged 58 (±14) [mean (±SD)] years were reviewed. Twenty-four (20.5%) were G1A1; 10 (8.5%) G1A2; 43 (36.7%) G2A1; 8 (6.8%) G2A2; 4 (3.4%) G2A3; 7 (6%) G3aA1; 2 (1.7%) G3aA2; 3 (2.6%) G3bA1, 1 (0.9%) G3bA2 and 1 (0.9%) G4A3. Only ACR was requested in 14 (12%) cases. Overall compliance with NICE recommendation was found in 42 (36%) while 75 (64%) were non-compliant.

        Conclusion: CKD monitoring was not adequately performed in line with NICE recommendations. The laboratory is the key link to engage primary care and renal physicians to develop local policies for monitoring CKD. Testing of individuals at high risk and early identification of patients with progressive CKD is a priority.

      • P144 Uncertainty of measurement of urine albumin and protein to creatinine ratios: computational approach highlights 'hidden' variability

        A O'Brien (Clinical Biochemistry, Cork University Hospital, Cork, Ireland), S Costelloe (Clinical Biochemistry, Cork University Hospital, Cork, Ireland)

        Background: Medical laboratories must regularly review uncertainty of measurement (UM) for all measurands (ISO15189:Cl5.5.2.4). For calculated tests, UM depends on the imprecision of component values. Numerous combinations of component values may yield the same result. Thus, a UM range (UMR) might apply to many values of a calculated test. Here, UMRs for the urine-albumin-to-creatinine-ratio (ACR) and the urine-protein-to-creatinine-ratio (PCR) are considered.

        Aims: Model UMRs for all possible ACR and PCR values. Estimate how UMRs affect classification of albuminuria and proteinuria.

        Method: Imprecision profiles for urinary-creatinine ((UCR): 1-37 mmol/L), urinary-protein ((UPRO): 17-1599 mg/L), and urinary-albumin ((UALB): 2-504 mg/L) were generated using residual specimens. UMs (UM=1.96*%CV) were calculated for each UALB/UCR and UPRO/UCR, using appropriate imprecision data, by Monte Carlo modelling. UMRs for each value of ACR and PCR were constructed. UCR, UPRO, and UALB data for 2017 were gathered. Using the UMRs, misclassification of albuminuria and proteinuria (using KDIGO guidelines) was estimated. Analyses were run in R.

        Results: For all ACRs and PCRs modelled, UM was 4.1%–31.4% and 3.9%-46.1% respectively. The UMRs at cut-offs were: 4.4%-20.6% at ACR=3.0 mg/mmol (moderate albuminuria); 4.4%-5.1% at ACR=30 mg/mmol (severe albuminuria); 4.4%-24.8% at PCR=15 mg/mmol (moderate proteinuria); and 4.1%-16.1% at PCR=50 mg/mmol (severe proteinuria). Patients may have been misclassified for albuminuria in 101/11063 (0.91%) cases and for proteinuria in 414/22337 (1.85%) case in the period examined.

        Conclusion: Many laboratories rely on rudimentary estimates of UM for calculated tests. This study demonstrates that UMRs for ACR and PCR vary widely, with the widest UMRs observed at lower values of ACR and PCR, which will affect diagnosis of mild albuminuria and proteinuria. At higher values of ACR UMRs are narrower; but for PCR a 4-fold variability in UM is observed at the cut-off for severe proteinuria depending on UPRO/UCR. Nephrologists are almost certainly unaware of this variation and laboratories should address this deficit with clinical users.

      • P145 Hospital-acquired acute kidney injury prevalence in adults at a South African tertiary hospital

        K Fenna (Department of Clinical Biochemistry, Royal Surrey County Hospital, Guildford, United Kingdom)

        Background: Hospital-acquired acute kidney injury (HA-AKI) prevalence had not yet been analysed in a South African tertiary hospital setting. In this study we investigated the prevalence of HA-AKI, in line with the KDIGO definition, along with clinical characteristics and outcomes according to gender and age. The aim was to provide evidence for the need of earlier treatment intervention initiatives to improve patient outcomes, such as the recent UK NHS initiative of automated electronic alerts in the laboratory information system.

        Methods: Retrospective analysis of laboratory and clinical data was analysed for a given 6 month period (01/12/2015 – 01/06/2016) at Tygerberg Hospital, Cape Town. Serum creatinine (SCr) laboratory results and clinical records were analysed and collated into gender and age group specific results before subsequent analysis.

        Results: HA-AKI occurred in 6.2% of the hospitalised patients admitted to Tygerberg Hospital for the given 6 month period of analysis. The highest incident of HA-AKI occurred in females aged 18-39 and males aged 40-59. The most common AKI stage reached was AKI stage 1. HA-AKI increased length of hospital by an average of 4.6 days and 20% of patients were readmitted to hospital at a later date with abnormal renal function.

        Conclusions: AKI prevalence is significant and associated with adverse patient outcomes. Initiatives that allow front-line healthcare professionals to treat and manage an AKI, such as the introduction of automated electronic alerts in the laboratory information system, should be considered. Similar initiatives have been implemented in all UK NHS hospitals with positive impacts.

Updated 11 June 2018

© Copyright The Association for Clinical Biochemistry and Laboratory Medicine
The Association for Clinical Biochemistry and Laboratory Medicine, 130-132 Tooley Street LONDON SE1 2TU
Tel: 020 7403 8001 Fax: 020 7403 8006