Proceedings of the Australasian Association of Clinical Biochemistry and Laboratory Medicine’s 2021 Virtual Scientific Conference

. 2021 Dec;42(Suppl 1):S1–S36.

Proceedings of the Australasian Association of Clinical Biochemistry and Laboratory Medicine’s 2021 Virtual Scientific Conference

PMCID: PMC9199447 PMID: 35719697

CBR. 2021 Dec;42(Suppl 1):S7.

A1 DYNAMIC SPECIMEN VALIDITY ANALYSIS IN URINE DRUG SCREENING UTILISING LC-QTOF-MS

D Allen ^1,^✉, C Warnholtz ¹, C Bachmeier ¹, B McWhinney ¹

Introduction

The assessment of specimen validity is an important component of urine drug screening to identify when the detection of illicit drugs/medications may be compromised or subverted. The sophistication of synthetic urines also presents analytical challenges due to the replication of commonly monitored urine parameters such as creatinine, urea, pH, specific gravity and colour. LC-QTOF-MS data-independent acquisition provides vast amounts of chemical information from a sample. A comprehensive panel of endogenous urine analytes was proposed and subsequently extracted from LC-QTOF-MS urine sample data for assessment as specimen validity markers. This study demonstrates the utility of five of these analytes for the detection of synthetic urines in drug screening, and the adaptability of LC-QTOF-MS in providing robust specimen validity analysis.

Methods

Data acquired by LC-QTOF-MS (Waters G2-XS) from 50 urine drug screening samples (Creatinine range <0.9–28.6 mmol/L) was retrospectively analysed for the presence of a large panel of endogenous urine compounds. Included in this panel were the compounds tryptophan, 3-methylhistidine, phenylalanine, phenylacetylglutamine and p-cresyl sulfate. Data from four previously detected synthetic urines was also analysed for the presence of these validity markers. The utility of phenylacetylglutamine as a proposed human specific metabolite was also assessed by the analysis of three canine urines.

Results

The validity markers tryptophan, phenylalanine, and p-cresyl sulfate were detected in all 50 urine samples. Phenylacetylglutamine and 3-methylhistidine failed to be detected in only one dilute urine (creatinine <0.9 mmol/L). All validity markers were absent in three of the synthetic urines. One synthetic urine contained only phenylalanine which was likely due to the use of an apple juice base. Phenylacetylglutamine could be detected in all canine urines.

Conclusion

LC-QTOF-MS allows for the simultaneous analysis of a comprehensive number of endogenous sample integrity markers in urine, providing dynamic and robust sample validity analysis in urine drug screening.

CBR. 2021 Dec;42(Suppl 1):S7.

A2 COVID-19 MONITORING APP: A TRIAL IN MOUNT GAMBIER’S MEDI-HOTEL

C Phay ^1,^✉, E Kararmaz ^1,^✉, G Wilson ¹, R Tirimacco ¹, S Carrison ², E Pretorius ²

Introduction

The COVID-19 Monitoring App was developed by the Integrated Cardiovascular Clinical Network South Australia (iCCnet SA). Following the rise in COVID-19 cases in early 2020, iCCnet identified the potential use of remote monitoring to reduce the need for healthcare workers to physically engage with potentially COVID-19 positive cases. The aim of the trial was to assess the feasibility of implementing this program state-wide.

Methods

iCCnet’s established home monitoring platform was adapted for COVID-19 monitoring and trialled in the Mount Gambier medi-hotel from 21 October to 25 November 2020. The app is formatted as an ‘interview’. Guests operated monitoring equipment and answered a series of clinical questions. Results were automatically uploaded to a secure database. A total of 18 participants enrolled in the program, 11 were ineligible due to age or personal circumstances, and 4 declined. Participants completed their interviews daily, iCCnet staff assessed the results and reported to the medi-hotel nurses.

Results

As no guests were COVID-19 positive during the trial, there is no data to support an association between the program and early identification of positive cases, or of clinical deterioration. The compliance rate for interview completion was 92.7%, with 100% of participants satisfied with their experience. 66.7% of participants felt more at ease knowing that their health was being monitored closely for COVID-19 symptoms, and 83.3% of participants were willing to participate in monitoring again.

Conclusion

The trial identified areas of improvement for the app’s future utilisation in identifying symptoms and early deterioration of COVID-19 positive cases. The app is flexible and could be adapted to a variety of settings outside of the hospital, with the overall goal to reduce the pressure on the acute healthcare system. Lessons learnt from this trial are being used to implement the app into a metro COVID-19 positive medi-hotel.

CBR. 2021 Dec;42(Suppl 1):S7.

A3 VALIDATION OF CEREBROSPINAL FLUID FOR PORPHOBILINOGEN TESTING USING MAUZERALL AND GRANICK METHOD

YC Leong ^1,^✉, C Chiang ¹, J Smith ¹

Introduction

Urinary concentrations of porphobilinogen (PBG) are elevated in acute hepatic porphyrias, especially during an acute attack. PBG concentration in CSF was requested recently in our laboratory for a patient with homozygous HMBS (hydroxymethylbilane synthase) mutations and neurological symptoms thought to be related to elevations in central nervous system PBG. We set out to validate CSF as a matrix using the Mauzerall and Granick PBG method and a modified Mauzerall and Granick method by Miyagi et al. which was suggested to be more sensitive.

Methods

The samples were tested in duplicate using the Mauzerall and Granick method and the modified method. CSF was validated as a matrix using serial dilution of a validated matrix (urine) with a high PBG into three different CSF samples. Limit of Blank (LOB), Limit of detection (LOD) and precision were determined with reference to CLSI guidelines. CSF samples from 40 subjects were tested to determine CSF PBG levels in patients without acute porphyria.

Results

Dilution of PBG in a validated matrix into CSF was linear. The LOB and LOD were 0.36 μmol/L and 0.43 μmol/L respectively. The coefficient of variation (CV) at the LOD was 2.65%. The upper limit of normal (97.5% percentile) for patients without AIP was 0.97 μmol/L. The homozygous HMBS patient’s CSF PBG was 3.0 μmol/L, which is comparable to the levels in affected patients published in the literature.

Conclusion

The study showed our current Mauzerall and Granick PBG method can be used to measure CSF samples. The modified method did not increase the sensitivity.

CBR. 2021 Dec;42(Suppl 1):S7–S8.

A4 THE ACCURACY OF VITAMIN A AND E MEASUREMENT IN SERUM IS AFFECTED BY THE PRESENCE OF VITAMIN E ANALOGUES

M Fitzpatrick ^1,^2,^✉, T Andersen ^1,³, J Bush ^1,⁴, R Greaves ^1,⁵, G Woollard ^1,⁶, K Hoad ^1,⁷, J Collie ^1,⁸, S Browning ^1,⁹, T Harrower ^1,⁹, W Punyalack ^1,⁴

Introduction

Monitoring serum vitamin A (retinol) and vitamin E (α-tocopherol) concentrations is common practice for assessing nutritional status. Measurement can be challenging due to several factors including traceability, harmonisation and potential interferences from their common analogues. Whilst the RCPAQAP Serum Vitamins Program assists participating laboratories in harmonisation, the materials provided do not contain the analogues of retinol and α-tocopherol that can interfere in quantification leading to over or under reporting of a patient’s nutritional status. We aimed to assess participants’ capacity to accurately report retinol and α-tocopherol in the presence of potential vitamin E analogue interfering compounds.

Methods

A supplementary series of four lyophilised samples were prepared which included a control and three matched samples spiked with α-tocopherol acetate, γ-tocopherol or both substances. The samples were distributed to each laboratory participating in the Serum Vitamins Program; all of which use chromatographic analytical techniques for measurement. Retinol and α-tocopherol results for each spiked sample were compared to the results of the control sample submitted by each participant. Acceptability of retinol and α-tocopherol results was determined based on the RCPAQAP allowable performance specifications (APS).

Results

Thirteen participants returned results for the supplementary sample series. Interference from α-tocopherol acetate was observed with results below the APS in 30% (n=4) of laboratories for retinol quantification and in 23% (n=3) for α-tocopherol quantification. One laboratory returned results above the APS for α-tocopherol when γ-tocopherol was present.

Conclusions

This supplementary sample series has shown that analogues of these vitamins can lead to the over or under estimation of nutritional status by some participants. Interference from α-tocopherol acetate indicates its use as an internal standard and the laboratories’ failure to note variations in internal standard concentrations in a chromatographic run. Whereas the interference from γ-tocopherol was present suggesting insufficient chromatographic separation. Affected laboratories are encouraged to review their analytical procedures.

CBR. 2021 Dec;42(Suppl 1):S8.

A5 CONGENITAL ADRENAL HYPERPLASIA DETECTION AND MONITORING IN DRY BLOOD SPOTS BY LIQUID CHROMATOGRAPHY TANDEM MASS SPECTOMETRY (LC-MS/MS)

AL Schick ^1,^✉, BC McWhinney ¹, R Swenson ¹, JPJ Ungerer ¹

Introduction

17-hydroxyprogesterone (17OHP) is commonly used by newborn screening programs as a screening tool for congenital adrenal hyperplasia (CAH), but is prone to false positives caused by cross-reactivity with endogenous compounds and low antibody specificity. Elevated 17OHP is also associated with low birthweight, infantile illness and prematurity. Second tier testing by tandem mass spectrometry is recommended for all programs. The purpose of this study was to validate a second tier panel for CAH which included the quantification of 17OHP, cortisol, androstenedione, 11-deoxycortisol and 21-deoxycortisol in dried blood spot samples by LC-MS/MS.

Method

17OHP, cortisol, androstenedione, 11-deoxycortisol and 21-d eoxycortisol were extracted from dried blood spot samples using 90/10 v/v methanol/water containing the deuterated internal standards. The extract was purified via an Oasis MAX μElution plate and 15 μL of supernatant injected onto a CORTECS C18 2.7 μm column and then eluted into the MSMS and ionised by ESI in positive ion mode. The analytes of interest and their deuterated internal standards were measured in multiple reaction monitoring mode. Cycle time between samples was 1.75 minutes.

Results

Analytical range was 0–250 nmol/L for 21-deoxycortisol and 0–1500 nmol/L for the other remaining analytes. The coefficient of variation (%CV) intra- and inter-run precision was ≤10%. LOB and LOD was <0.1 nmol/L for all analytes. Carryover was <0.1% between injections.

Conclusion

This LC-MSMS second tier assay provides excellent specificity and sensitivity in the detection of CAH. More importantly, it reduces stress to families and clinicians as well as the follow up burden to newborn screening scientists by minimising the need for retesting. Pathology Queensland have been running this assay for the past year, during which its discriminatory power in differentiating false positives has been very successful.

CBR. 2021 Dec;42(Suppl 1):S8.

A6 PREDICTIVE UTILITY OF BIOMARKERS FOR MAJOR CARDIOVASCULAR EVENTS IN PATIENTS WITH PERIPHERAL ARTERY DISEASE

TP Singh ^1,^2,^✉, R Velu ², F Quigley ⁴, J Jenkins ⁵, J Golledge ^1,^2,^4,⁵

Background

The aim of this study was to examine the utility of inflammation and thrombosis biomarkers in predicting major cardiovascular events in patients with peripheral artery disease (PAD).

Methods

1652 participants with a variety of PADs had serum C-reactive protein (CRP), fibrinogen, neutrophil-lymphocyte ratio (NLR) and homocysteine measured. Participants were followed for a median of 3.7 (inter-quartile range 1.7, 6.7) years to record outcome events including major adverse cardiovascular events (MACE; myocardial infarction, stroke or cardiovascular death) and peripheral artery surgery. The association of biomarkers with events was assessed using Cox proportional hazard analysis. The ability of the biomarkers to predict events was assessed using area under the curve of receiver operator characteristics, net reclassification index (NRI) and classification and regression tree (CART) analyses.

Results

After adjusting for other risk factors, serum fibrinogen (hazard ratio, HR, 1.10 95% confidence intervals, CI, 1.01, 1.19, per g/L) and homocysteine (HR 1.02, 95% CI 1.00, 1.03, per 6 μM) but not NLR or CRP were significantly associated with an increased risk of MACE. Serum fibrinogen (HR 1.17, 95% CI 1.09, 1.25, per g/L) but not NLR, homocysteine or CRP were independently significantly associated with an increased risk of peripheral artery surgery. According to CART analyses, homocysteine was the more important biomarker in stratifying risk of MACE, while fibrinogen was the most important biomarker in stratifying risk of peripheral artery surgery.

Conclusions

Serum homocysteine and fibrinogen may be useful biomarkers in stratifying risk of PAD-associated major cardiovascular events.

CBR. 2021 Dec;42(Suppl 1):S9.

A7 EVALUATION OF FACTORS IN THE POLYETHYLENE GLYCOL (PEG) PROCEDURE FOR MACROPROLACTIN INVESTIGATION

K Tran ^1,^✉, C Ariyawansa ¹, MJ Tanner ¹

Introduction

Macroprolactin is a large auto-antibody complex that results in apparent ‘false’ hyperprolactinaemia. Whilst it is routinely evaluated by polyethylene glycol (PEG) precipitation, currently there is no uniform method; with little evidence demonstrating how altering procedural steps change recovery. Additionally, without a commercially available Quality Assurance Program (QAP) in Australasia it has not been demonstrated that laboratories are producing comparable macroprolactin results. A patient and QAP pool were established to evaluate the macroprolactin procedure.

Method

Patient and RCPA QAP pools were used with prolactin recoveries of 68% and 99% respectively, and stored at −18°C. The method of PEG precipitation was sourced from PathWest QEII in Perth, WA. The effect of dilution on recoveries was determined by altering the patient:PEG and patient:diluent ratio from 1:1 to 1:6. Effect of diluents was simultaneously tested by changing the patient and PEG diluents with phosphate buffered saline (PBS), saline, water and Abbott Multiassay Diluent. To determine analyte stability the prolactin recovery and total prolactin were retested over a three-month period.

Results

The patient and QAP pools obtained a 3-month total prolactin stability CV of 2.51% and 3.17%, and 3-month prolactin recovery stability of 5.24% and 3.29%, respectively. As the diluent ratio was increased, the patient and QAP pools obtained correlation coefficients ranging between −0.65 to −0.91 and −0.38 to −0.92, correspondingly, across all diluent types. None of the curves were statistically different from each other (when P<0.5).

Conclusion

The QAP and patient pool have demonstrated comparable total prolactin and recovery stability over a three-month period. Changing the diluent type to water, saline, multiassay diluents or PBS has shown insignificant differences in this study. As the ratio of patient:diluent increases, the recovery seems to reduce proportionally. The stability of RCPA QAP material demonstrates potential use in a macroprolactin QAP program as a negative control.

CBR. 2021 Dec;42(Suppl 1):S9.

A8 RETROSPECTIVE REVIEW OF INCREASING VITAMIN B6 WHOLE BLOOD LEVELS IN AUSTRALIA

S Browning ^1,^✉, B Teis ¹, M Henman ¹, R Bain ¹, K De Silva ¹, T Harrower ¹, D Holman ¹, S MacDonald ¹, M Najdovski ¹, K Sharvell ¹, C Sklavos ¹, S Stewart ¹, G Daley ¹, R Flatman ¹, L Price ¹, D Kanowski ¹, G Ward ¹

Introduction

A recent review (2021) has stated that the number of cases of vitamin B6 toxicity worldwide is increasing. From an Australian perspective, the TGA released an adverse drug reactions bulletin (2008) and a safety advisory (2020) regarding vitamin B6 supplementation causing peripheral neuropathy. Further, our laboratory demonstrated <2% vitamin B6 deficiency in our population in 2014. This study examined the trend of vitamin B6 whole blood levels over the past 7 years in the Australian community.

Methods

Pyridoxal-5-phosphate (B6) was measured in whole blood using the Chromsystems vitamin B6 HPLC method (until February 2021) and then using the combined vitamin B1 and B6 UPLC method, both with fluorometric detection. B6 levels in 88,655 samples between January 2014 and June 2021 were reviewed.

Results

B6 levels in 53,956 samples (60.9%) were within the reference interval (RI) (20–190 nmol/L). Nine samples (0.01%) had a level <20 nmol/L and 34,690 samples (39.1%) were >190 nmol/L. Of these samples 5431 (6.1%) had levels of >1,000 nmol/L. The patient mean in 2014 was 235 nmol/L and has increased to 444 nmol/L in 2021.

Conclusion

B6 deficiency is uncommon in the Australian community and the number of patients with levels reported to be consistent with neurological symptoms (275 nmol/L) has increased between 2014 and 2021. A 2014 UK study revealed 7% of hospitalised patients had a B6 level >550 nmol/L which is similar to our community findings of 8% in 2014. Literature and the TGA suggest an intake of 50 mg/day can be associated with B6 toxicity. There is inter-individual difference in the metabolism of B6 and therefore dose does not correlate well with blood levels. However, patients taking 50 mg/day have a B6 level of approximately 400 nmol/L. Patients with neurological symptoms should have their B6 levels checked regardless of supplementation.

CBR. 2021 Dec;42(Suppl 1):S9.

A9 DEVELOPMENT AND VALIDATION OF AN LCMSMS METHOD FOR THE ANALYSIS OF ANTIMICROBIALS IN PLASMA

A Ireland ^1,^✉, M Fitzpatrick ¹, VL Nguyen ^1,², D Sullivan ^1,², S Van Hal ^1,^2,³

Introduction

Beta-lactam antibiotics and the oxazolidinone antibiotic Linezolid have been traditionally administered in an empiric fashion. Recent pharmacokinetic studies have shown that standard dosing regimes may not achieve drug concentration targets to effectively clear bacterial infections – particularly in critically ill patients with altered protein binding and drug clearance. In this study, we aimed to develop an LCMSMS method for the measurement of six beta-lactams and Linezolid in plasma to support dose optimisation in patients with altered pharmacokinetics.

Methods

Chromatographic conditions and detector settings were optimised after careful selection of analytical column and mobile phases. To extract the antibiotics, plasma samples were diluted with a buffer, followed by protein precipitation and addition of an internal standard solution prior to centrifugation. The supernatant was then diluted with a buffer prior to analysis. The method was validated to include recovery, precision, accuracy, linearity, limits of quantification, extract stability and matrix effects.

Results

During method validation, recovery for all analytes was shown to be >90% with minimal matrix effects noted. The within-batch imprecision was <10% and the between-batch imprecision was <5% across all drugs. A lower limit of quantification of 0.5 mg/L for each drug showed CVs of <20%. Negligible degradation was seen in the extracted sample over a 24-hour period. This method displayed good agreement with external quality assurance (INSTAND) assigned values and comparisons to another laboratory.

Conclusion

We have developed a simple, robust method for the measurement of select antimicrobials in plasma. This method has been shown to reliably support dose adjustment when factors such as protein binding and minimum inhibitory concentrations are considered during result interpretation.

CBR. 2021 Dec;42(Suppl 1):S10.

A11 MOVING TOWARDS HARMONISATION OF MACROPROLACTIN ACROSS NSW HEALTH PATHOLOGY

P Ward ^1,^✉, R Guy ¹, J Sherfan ², L Boscato ³, T Yen ⁴

Introduction

NSW Health Pathology (NSWHP) has developed a recommended approach for the screening of samples with elevated total prolactin through the measurement of macroprolactin. In Sydney there are 5 immunoassay platforms for prolactin in routine use, so a verification study of the Beltran monomeric prolactin reference intervals (RI) was performed. Verification of corrected monomeric prolactin against total prolactin RI was also performed.

Methods

The NSWHP harmonised macroprolactin analytical procedure was adopted by the participating laboratories. Thirty-seven normal samples (19F, 18M) were sent to six Sydney laboratories using a total of five immunoassay platforms (Abbott-Architect, Siemens-Immulite, Roche-Cobas, Beckman-DXL, Siemens-Atellica). Macroprolactin was PEG-precipitated by the harmonised procedure and monomeric prolactin was measured. Average monomeric recovery was determined on all platforms from the 37 samples.

Results

Four samples containing macroprolactin were identified and confirmed using the five platforms tested. RIs published by Beltran (2008) were verified for all platforms except Roche – a consequence of two females with elevated total and monomeric prolactin and no detectable macroprolactin. Verification was achieved if 18M and 18F samples were within the quoted RI. Verification against the monomeric prolactin RI published by Overgaard (2016) was only successful when using the Beckman-DXL. Failure to verify occurred because the upper limit of the female RI was much lower than those published by Beltran. Monomeric prolactin corrected for the average platform specific recovery of monomeric prolactin was verified against Total Prolactin RIs. Average recovery ranged from 67% using the Abbott-Architect to 102% with the Beckman-DXL.

Conclusion

Surprisingly, 4 of 37 subjects had macroprolactin identified giving an incidence of 11%. Macroprolactin was detected in the five platforms tested. The Beltran RI are suitable for method-specific interpretation, and reporting corrected monomeric prolactin against Total Prolactin RI has merit. Harmonisation/standardisation of the PEG procedure between labs is essential in order to compare results between platforms.

CBR. 2021 Dec;42(Suppl 1):S10.

A12 THE IMPACT OF FOLATE AND VITAMIN B12 STATUS ON COGNITIVE FUNCTION AND BRAIN ATROPHY IN HEALTHY ELDERLY AND DEMENTED AUSTRIANS: A RETROSPECTIVE COHORT STUDY

M Herrmann ^1,^✉, J Rabensteiner ¹, E Hofer ¹, G Fauler ¹, T Benke ², P Dal-Bianco ³, G Ransmayr ⁴, R Schmidt ¹

Introduction

Dementia, and in particular Alzheimer’s disease (AD), is a debilitating progressive disease with high prevalence in our ever-aging societies. Vitamin B12 and folic acid deficiency have been proposed as potential modifiable risk factors. However, previous studies reported inconsistent results. The present retrospective study aimed to explore the association between direct and indirect biochemical markers of B12 and folate status, cognitive function and MRI-based brain atrophy in cognitive normal elderly and AD patients.

Methods

Stored blood samples from 378 normal elderly healthy controls and 217 AD patients were used to measure homocysteine, folic acid, total and active vitamin B12 and MMA. A sophisticated panel of neuropsychiatric tests capturing memory, executive function and visuopractical skills analysed cognitive function of all participants. Brain atrophy was assessed by MRI in 155 control subjects and 217 AD patients. In a subset of participants, cognitive testing (n=234) and MRI (n=182) was repeated after an average 1.25–6.25 years.

Results

The average concentrations of all biochemical markers of vitamin B12 and folate status were within the respective reference intervals. Cross-sectional analyses and longitudinal analyses did not reveal significant associations between biochemical markers of B-vitamin status and cognitive function after correction for multiple testing. None of the biochemical markers of B-vitamin as well as folate status related to global or regional brain volume, cortical thickness or cortical surface area, neither in normal elderly individuals nor in AD patients.

Conclusion

Elderly Austrians are well supplied with B12 and folic acid. Variations of direct and indirect markers of B12 and folate status are not associated with cognitive dysfunction and brain atrophy. Therefore, supplementation of B12 and folic acid in these subjects is unlikely to exert beneficial short- to mid-term preventive effects on cognitive dysfunction and brain parenchymal loss.

CBR. 2021 Dec;42(Suppl 1):S11.

P1 MERRF SYNDROME WITH NO RAGGED RED FIBRES DUE TO AN UNCOMMON HETEROPLASMIC MITOCHONDRIAL DNA MUTATION M.4284G>A

F Chi-kin Wong ^1,^✉, CY Law ¹, G Yuan ², W Chun-yin Leung ², TK Ling ¹, KC Wong ¹, R Shek-kwan Chang ², CW Lam ^1,³

Introduction

A 34-year-old man had a diagnosis of epilepsy since the age of 11, with recurrent photosensitive myoclonic and generalised tonic-clonic seizures. One of his brothers had epilepsy with myoclonic seizures. Initially treated as juvenile myoclonic epilepsy, the patient’s seizure control had been fair only. At 23 years, he was diagnosed to have diabetes mellitus. His seizure control deteriorated at 29 years with frequent hospital admissions. Serum lactate was elevated (9.3mmol/L).

Methods

Genetic analysis was performed by next-generation sequencing using targeted gene capture technique on Illumina NextSeq 500 Sequencing system.

Results

Muscle biopsy of quadriceps did not show ragged red fibres (RRF). Targeted mutation studies by Sanger sequencing for common mitochondrial DNA mutations were negative. He subsequently developed cerebellar ataxia, hearing impairment and cognitive decline. Progressive myoclonic epilepsies were suspected. While no mutations were identified by next-generation sequencing in a nuclear gene panel for epilepsy, heteroplasmic m.4284G>A, a known pathogenic variant, was detected in the mitochondrial DNA. The mutation load in blood (i.e., heterplasmic levels), urine and muscle were 65%, 83% and 100%, respectively. The mutation was also detected in the patient’s brother with a milder phenotype (heteroplasmy in urine: 60%, heteroplasmy in blood: 40%) and mother (undetectable in blood; detectable in urine with heteroplasmy of 38%).

Conclusions

Notably, RRF were absent in all reported patients with m.4284G>A to date including this patient. The rare pathogenic variant, m.4284G>A identified in this case should expand our understanding of the spectrum of myoclonic epilepsy caused by mitochondrial variant.

CBR. 2021 Dec;42(Suppl 1):S11.

P2 MEASUREMENT OF DEXAMETHASONE ENHANCES THE CONFIDENCE IN THE DEXAMETHASONE SUPPRESSION TEST

Z-Q Ou ^1,^✉, G Woollard ¹, W Chiu ¹, C Kyle ¹

Introduction

Low dose (1 mg) dexamethasone suppression test (DST) is a well-established procedure to screen for autonomous cortisol production in Cushing’s syndrome (CS). Cortisol suppression to exclude CS should be <50 nmol/L. Dexamethasone is not routinely measured and levels depend on patient compliance and clearance by CYP3A4 which is variably induced or inhibited. The dexamethasone level required to achieve sufficient cortisol suppression to avoid false CS positivity was determined.

Methods

A chromatographic method for dexamethasone in plasma was designed and validated. It involved extraction into methyl-t-butylether with added dexamethasone-D4 internal standard and subsequent evaporation and reconstitution. Further clean-up was by on-line SPE with Turboflow followed gradient elution and TSQ Vantage MS detection. Cortisol was measured by Cobas 800 series immunoassay. Oral dexamethasone was administered at 11 pm–midnight and blood was sampled around 9 am next morning to test cortisol and dexamethasone levels. Initially healthy subjects (n=8) without CS were tested to estimate range of expected dexamethasone levels (median 6.6 nmol/L; range 3.2–18.2 nmol/L). It was extended to 140 patients under investigation for CS from the community.

Results

LLOQ for dexamethasone is 0.25 nmol/L (±3.33% CV). Cortisol levels track dexamethasone in a non-linear fashion (Spearman’s rank correlation coefficient 0.30). The relationship is best described by a power function. This helped to identify the lowest dexamethasone level needed to suppress cortisol. An inflection point at 3.1 nmol/L was found, below which cortisol suppression is not reliable. This observation agrees with literature findings. Ten patients were found to have adequate dexamethasone levels with post-dexamethasone plasma cortisol >100 nmol/L. Of these, 2 patients were confirmed CS, 3 were possible CS with adrenal mass and 2 were excluded. The remainder were indeterminate (lost to follow-up).

Conclusion

Measurement of dexamethasone improves the diagnostic CS confidence by demonstrating sufficient dexamethasone (≥3.1 nmol/L) to adequately suppress cortisol to <100 nmol/L, even though it cannot distinguish other conditions that induce cortisol.

CBR. 2021 Dec;42(Suppl 1):S11.

P3 QUALITATIVE ANALYSIS OF NICOTINE, COTININE AND ANABASINE IN URINE USING QUADRUPOLE TIME OF FLIGHT MASS SPECTROMETRY

CR Warnholtz ^1,^✉, DR Allen ¹, BC McWhinney ¹, JPJ Ungerer ¹

Introduction

Cessation of tobacco smoking is a mandatory requirement for lung transplantation, the use of medical oxygen and for some surgical procedures. Effective monitoring of this cessation is essential to ensure the patient is compliant with pre-operative protocols. We have developed a liquid chromatography-quadrupole time-of-flight (LC-QTOF) assay that qualitatively determines the presence of nicotine, cotinine and anabasine, a pyrolysis product produced by tobacco smoking. This assay effectively replaces the gas chromatography mass spectrometry (GC/MS) and LC-QTOF urine toxicology screens currently used to determine general smoking status.

Methods

Urine (100 μL) was mixed with water containing deuterated internal standards (20 μg/L), and 5 mM ammonium formate buffer at pH 3.0 in a sample vial. The samples were mixed for 5 seconds and injected into the LC-QTOF system and the analytes separated on a 150 mm UPLC HSS C18 column.

Results

The LOD was determined as 5 μg/L for all analytes. The assay correlated with the GC/MS method, as well as the current urine toxicology screening method on LC-QTOF. The addition of deuterated internal standards for nicotine and cotinine added extra confidence to interpretation. Total run time was 9 min between injections.

Conclusion

We have developed an LC-QTOF assay to detect nicotine, cotinine and anabasine. It shows excellent specificity and sensitivity. Sample preparation has been simplified and run time shortened by 12 minutes/sample when compared to the GC/MS method. The new assay adds the capability to determine whether the presence of cotinine is due to tobacco smoking or other non-tobacco sources.

CBR. 2021 Dec;42(Suppl 1):S12.

P4 REVIEW OF PERFORMANCE IN THE RCPAQAP SERUM INDICES PROGRAM

J Bush ^1,^✉, S Shepherd ¹, W Punyalack ¹, P Graham ¹

Introduction

The RCPAQAP Serum Indices program was first introduced in 2019 to enable laboratories to compare their haemolysis, icterus, and lipaemia results with their peers. We previously reported on the initial performance of enrolled laboratories, highlighting a lack of harmonisation of selected units and setup even within the same instrument group. We sought to update our review of performance in this program.

Methods

Quantitative and qualitative results from 2019 and 2020 were analysed using in-house software. Performance based on the average CVs for haemolysis (H), icterus (I) and lipaemia (L) for quantitative methods and the level of consensus for qualitative methods was reviewed.

Results

Seventy laboratories enrolled in the Indices program in 2019, and 280 in 2020. The initial 2019 results showed average CVs of 6.1%, 22.7% and 14.1% for H, I, L, respectively. While there was no significant improvement in quantitative performance in the 2020 program (average CVs 5.9%, 22.6% and 11.5%, for H, I, L), it was noted that the newly enrolling laboratories still had a lack of understanding as to what constitute quantitative vs qualitative/semi-qualitative values. Once this was sorted via QAP staff support, good consensus (>70%) was achieved for qualitative results.

Conclusion

Since the introduction of a new EQA Indices program, laboratories continue to require support on understanding their methodologies. This highlights another area where QAP can contribute to improved performance in an important area where indices results are used to make decisions on accepting or rejecting patient results.

CBR. 2021 Dec;42(Suppl 1):S12.

P5 VERIFICATION OF REFERENCE INTERVALS FOR THE ROCHE COBAS PROLACTIN II ASSAY

MP Nascimento ^1,^✉, MB Theophilos ¹, C Lynch ¹, KA Sikaris ¹

Introduction

Prolactin concentrations vary with pregnancy and postpartum physiological changes, with organic disease, therapeutic drugs and also macroprolactinaemia. Appropriate reference intervals are required to help distinguish these conditions, and to define reflex testing limits for macroprolactinaemia by the laboratory. We aimed to verify the prolactin reference intervals in different age groups and gender.

Methods

We extracted 159,504 (125,029 female, 34,475 male) prolactin results performed with the Roche Cobas Prolactin II assay between 2015 and 2020. Pregnant patients were excluded from analysis. We examined the trend in the median distribution with age and gender to identify physiological partitions. We then applied Bhattacharya analysis to determine the reference intervals for each transition and compared these with the manufacturer’s recommendations.

Results

We identified gender-related differences from age 16 y onwards, with female prolactin concentrations higher than males across all age groups. Women over 59 y had lower values compared to those under 50 y. Median prolactin concentrations increased from age 80 y for both genders.

Discussion

Prolactin concentration between genders differs in adults, with higher values seen in females. We were able to confirm this pattern from the age of 16 y. In women, we chose a cut-off of 59 y to increase the specificity for identification of post-menopausal status. A lower age partition for (presumed) post-menopausal women may be required for increased sensitivity in the detection of hyperprolactinaemia. The rising prolactin concentrations observed in our elderly population may be related to decreased clearance with declining renal function, as well as increased secretion associated with pathological conditions or pharmacological treatment. Reference interval studies controlling for these factors could help determine whether a physiologic increase in prolactin secretion is present in this age group.

CBR. 2021 Dec;42(Suppl 1):S12.

P6 MAKING IT COUNT: IOHEXOL-RELATED SAMPLE SEPARATION INTERFERENCE IN RADIOLOGICALLY GUIDED ENDOCRINE DYNAMIC TESTING

C Stead ^1,^✉, A Desra ¹, J Smith ^1,², R Czajko ¹, C Chiang ¹

Introduction

Iohexol is an intravenous high-osmolar, non-polar, iodinated contrast media commonly used in computed tomography (CT)-guided interventional procedures. The impact of chemical interference by contrast media is well documented in the literature but is under-appreciated both in and outside of the laboratory. The difficulty of recollecting samples obtained through interventional radiology is of particular concern. In this case report, a 79-year-old man with primary aldosteronism was admitted for a contrast CT-guided adrenal vein sampling (ADVS). Samples from this procedure collected in gel tubes failed to separate appropriately post centrifugation due to Iohexol 350 contamination. Attempts to salvage serum from below the upper gel layer were incomplete and the results produced post-analysis were potentially inaccurate and misleading. A study was performed using healthy volunteers to identify the concentration of Iohexol 350 required to interfere with gel tube separation.

Methods

Serum samples were collected from healthy volunteers and were immediately aliquoted into four sets of collection tubes spiked with 2%, 4%, 7%, 10% and 15% Iohexol 350 v/v. Samples were mixed and allowed to clot as per kit insert, centrifuged at 3500 rpm for 10 minutes, and visually inspected for separation interference.

Results

At 4% v/v Iohexol concentration and above, serum and plasma could not be fully separated via centrifugation due to the change in density of the sample. This resulted in the separating gel layer rising above the serum sample instead of settling between serum and blood cell layers.

Conclusions

Iohexol alters serum and plasma sample properties in a dose-dependent fashion. Interventional procedures such as ADVS require appropriate collection to prevent repeated procedures. As the presence of contrast media is unavoidable in ADVS collections, our laboratory has changed procedure to use gel-free tubes to avoid the risk of samples not separating appropriately.

CBR. 2021 Dec;42(Suppl 1):S12–S13.

P7 ‘FLYERS’ IN A TROPONIN I ASSAY – TWO YEARS’ EXPERIENCE WITH A DUPLICATE TESTING PROTOCOL

GRD Jones ^1,^✉, S Mukerji ¹, P Bonnitcha ¹, A Screnci ¹, R Prasad ¹

Introduction

Troponin assays are known to suffer from occasional non-reproducible results (NRR, sometimes knows as ‘flyers’). We have previously reported these with the Beckman-Coulter High Sensitivity troponin I assay (G Jones et al., Clin Biochem Rev. 2019;40:S29). We report on two years’ further experience with a protocol to manage this issue.

Methods

Starting from mid-June 2019, all troponin requests were run simultaneously in duplicate on the same DxI analyser (ACCESS hsTNI, REF B52699). Where both results are >10 ng/L and the difference >20%, further testing was performed. If a stable result was found on further testing the result was released. If no stable results were found a repeat collection was requested. We accept BD SSTII and PSTII tubes. The presence of NRR was compared with other factors.

Results

Over 2 years a total of 206 samples have shown NRR. The median increase in the flyer above the released value was 26 ng/L (77%), IQR 11–106 ng/L (38–330%). NRR are 0.6% of all TnI requests. This is lower than the 2.4% reported in 2019. NRR occurred in 1.1% of heparin tubes and 0.3% of serum tubes. The highest percentages were from intensive care (2.2%), other wards (1.1%) and ED (0.4%). These differences are not fully explained by sample type. CRP levels were higher in NRR samples (medians: non-NRR 4.2, NRR serum: 11.2; NRR heparin 125 mg/L). Heparin NRR samples also had lower albumin and higher WCC. 12 patients had 2 or more NRRs, all with raised WCC and CRP. The 80^th centiles for ED turnaround time are not affected by this protocol.

Conclusions

Falsely elevated troponin results have the possibility to adversely affect patient care. Our protocol has identified multiple significant NRRs without affecting turnaround time for most samples. It is possible that NRRs are related to inflammation in the patient.

CBR. 2021 Dec;42(Suppl 1):S13.

P8 URINE FREE CORTISOL BY LC-MS/MS: INTERFERENCES CAN STILL OCCUR

R Bahnisch ^1,^✉, N Burke ¹, H Martin ¹

Introduction

Traditional consensus holds that measurement by mass spectrometry techniques are inherently more accurate than immunoassay techniques. A clinical query regarding a urine free cortisol (UFC) result generated by our LC/MS/MS method caused us to question this tenet. This work details our discovery that two steroid metabolites, 20α- and 20β-dihydrocortisone, could interfere with our original UFC method, changes made to eliminate these interferences and a small study to identify the prevalence of the metabolites.

Methods

Our existing method on a Shimadzu Nexera 8050 system separated cortisol, d4-cortisol and cortisone on a Phenomenex Turboflow C18-XL column using a gradient of mobile phase A (1% formic acid) and mobile phase B (acetonitrile). Samples were diluted with acetonitrile and internal standard.

The optimised method separates the same 3 analytes and, in addition, 20α- and 20β-dihydrocortisone on a Shimadzu Biphenyl column using a staggered gradient flow of mobile phase A (0.5 mM ammonium fluoride) and mobile phase B (methanol). Sample preparation was similar except methanol:water (50:50 v:v) was used as diluent.

128 patient urine samples were compared using both methods. Accuracy was evaluated against RCPA EQAP material. Precision, linearity and limit of quantification (LOQ) were also evaluated.

Results

10 of the 128 patient samples (8%) had significant concentrations of the two metabolites; the increases in cortisol excretion this caused ranged from 48–2712 nmol/mmol creatinine. (7 clinically significant). With these 10 samples excluded, excellent agreement of cortisol and cortisone results was achieved; Passing-Bablok fits y=0.99x–3.94 and y=0.98x+1.82 respectively. Precision, linearity and LOQ were comparable between the two methods. Immunoassay results indicated variable recognition of the metabolites.

Conclusion

Laboratories should beware the assumption that mass spectrometry methods are interference-free. Chromatographic separation of 20α- and 20β-dihydrocortisone from cortisol should be verified for all UFC LC-MS/MS methods to ensure no erroneously high results are reported.

CBR. 2021 Dec;42(Suppl 1):S13.

P9 INVESTIGATION OF A FALSE POSITIVE CODEINE INTERFERENCE IN AN AACB/RCPAQAP URINE TOXICOLOGY SURVEY SAMPLE

D Allen ^1,^✉, B James ¹, T Andersen ¹, A Naimi ¹, B McConnell ¹, A Leibie ¹, E Jenkins ¹, H Martin ¹, J Pope ¹, T Smith ¹, S Handley ¹, C Cruickshank ¹, G Moore (Chair) ¹

Introduction

A false positive codeine interference was suspected following review of AACB/RCPAQAP Urine Toxicology survey result (Cycle 29 – Sample 2). Although 27% of respondents reported the detection of codeine (median recovery 109 μg/L) using LC-MS based techniques, the high percentage of laboratories failing to detect codeine prompted an investigation into the presence of a potential interference. A participant also reported a similar interference in samples containing high concentrations of oxycodone using a LC-MS/MS technique. An investigation was subsequently undertaken using LC-QTOF-MS to identify the interfering substance.

Methods

Sample data obtained by data-independent acquisition (Waters G2-XS QTOF) for survey 29-02 was interrogated using QTOF software elucidation tools including elemental formula prediction and in silico fragmentation. Presumptive identification was achieved using accurate mass and molecular structure obtained from the Chemspider.com database.

Results

Untargeted data acquired from survey 29-02 indicated the presence of oxycodone and associated metabolites. A prominent signal was also observed at m/z 318.1695 co-eluting with the disputed codeine result. Using QTOF-MS elucidation tools, the co-eluting compound was presumptively identified as 6-oxycodol (C18H23NO4; m/z 318.1695), a minor metabolite of oxycodone. It was proposed that MS in-source fragmentation of 6-oxycodol, resulting in a subsequent water loss, would produce a fragment of similar mass to codeine (m/z 300.1587). Prediction of the formation of this isobaric fragment was also confirmed by in silico fragmentation analysis using the 6-oxycodol molecular structure. Also predicted was the formation of additional product ions commonly used in the LC-MS identification of codeine.

Conclusion

In-source fragmentation of 6-oxycodol was presumptively identified as a cause of false positive codeine interference in some LC-MS based methodologies. Resolution of this interference may involve modification of chromatographic separation or the use of unique product ion ratios. Acquisition of certified reference material is recommended for definitive confirmation of the interference.

CBR. 2021 Dec;42(Suppl 1):S13–S14.

P10 SERUM PROTEIN ELECTROPHORESIS: A RETROSPECTIVE AUDIT OF MIGRATION PATTERNS AND PARAPROTEIN CONCENTRATIONS

SA Handley ^1,^✉

Introduction

Serum protein electrophoresis (SPEP) is, in part, used for the detection of paraproteins. The Department of Pathology, Royal Hobart Hospital undertakes SPEP and immunofixation for both in- and outpatients. Results were retrospectively audited, in particular the relationship between the paraprotein concentration measured by SPEP and the immunoglobulin concentration measured by nephelometry, and the electrophoretic region that the paraprotein migrated in.

Methods

Patient demographics, total protein, albumin, IgA, IgG, IgM and paraprotein concentration, together with any interpretive comments, were extracted from the laboratory information system; 2010–2020. Protein electrophoresis was undertaken using the Sebia Hydrasys gel electrophoresis system. Total protein, albumin, IgA, IgG, and IgM were measured by nephelometry.

Results

There were 3490 samples analysed over the audit period. A paraprotein (2757) was detected in 65% of samples. IgG kappa/lambda, IgM kappa/lambda, IgA kappa/lambda, and IgD kappa/lambda were detected in 1,203/612, 294/158, 179/96, and 6/6 instances, respectively. IgE kappa/lambda were not detected. All IgG paraproteins migrated within the gamma-region, 90 % of IgM paraproteins migrated within the gamma/gamma-beta-region, 47% and 41% of IgA paraproteins migrated in the gamma- and beta-regions, respectively. Immunoglobulin D paraproteins largely (70%) migrated within the beta-region (30% in gamma-region). All paraproteins measured by nephelometry correlated well with the concentration measured by densitometry (IgG: R=0.94, IgM: R=0.93, IgA: R=0.92). The slopes for IgG and IgA paraproteins were 1.06 (95% CI 0.95–1.17) and 1.15 (95 % CI 1.13–1.17), respectively, showing a very good correspondence with quantification by nephelometry. The slope for IgM paraproteins was 1.64 (95% CI 1.53–1.75), illustrating a systematic higher result using nephelometry than SPEP.

Conclusion

Immunoglobulin G paraproteins were the most common detected. Although there was a good relationship between densitometry and nephelometry for IgG and IgA, IgM paraproteins were overestimated by densitometry.

CBR. 2021 Dec;42(Suppl 1):S14.

P11 VERIFICATION OF THREE POINT-OF-CARE ANALYSERS FOR HBA_1C

BV Li ^1,^2,^✉, A Kumar ¹, C Finlay ¹, E Barnes ¹, M van Drimmelen ³, CM Florkowski ^1,²

Introduction

Glycated haemoglobin (HbA_1c) point-of-care (POC) testing is an integral part of our laboratory service to provide timely results to the diabetes clinic for management of patients with diabetes.

Methods

30 samples were evaluated on the Abbott Afinion 2 and Roche cobas b 101, while 21 separate samples were tested on the Siemens Atellica DCA, and compared to the Bio-Rad D-100 HbA_1c assay. Passing-Bablok regression and Bland-Altman plots assessed the relationship between each POC assay using whole blood with our laboratory reference assay. Day-to-day imprecision was evaluated with six whole blood samples over five days for the Roche cobas b 101 and Abbott Afinion 2, and one whole blood sample over five days for the Siemens Atellica DCA due to consumable constraints. Comparisons were made with Royal College of Pathologists Australasia (RCPA) analytical performance specifications (APS) (±4 mmol/mol up to 45 mmol/mol; ±8% >45 mmol/mol).

Results

Using whole blood on POC devices, with the Bio-Rad D-100 as the reference, for the Roche cobas b 101, the Passing-Bablok regression was HbA_1c(b101)=6.548+0.894×HbA_1c(D-100), day-to-day coefficient of variation (CV) 2.0%, and 6/30 results outside RCPA APS. For the Abbott Afinion 2, the Passing-Bablok regression was HbA_1c(Afinion)=2.939+0.98×HbA_1c(D-100), day-to-day CV 2.0%, and 3/30 results outside RCPA APS. For the Siemens Atellica DCA, the Passing-Bablok regression was HbA_1c(DCA)=5.209+0.891×HbA_1c(D-100), day-to-day CV 2.5%, and 2/21 results outside RCPA APS. The Roche cobas b 101 had a negative bias with higher levels of HbA1c and a positive bias for lower levels of HbA_1c, while Abbott Afinion 2 had a positive bias with lower levels of HbA_1c, when compared with our reference Bio-Rad D-100 HbA1c analyser.

Conclusion

Roche cobas b 101, Abbott Afinion 2 and Siemens Atellica DCA performed satisfactorily for HbA_1c POC purposes with most results within RCPA APS and acceptable CV <3.0% for monitoring purposes.

CBR. 2021 Dec;42(Suppl 1):S14.

P12 THE USE OF SEMI-QUANTITATIVE H-INDEX IN THE RAPID MEASUREMENT OF FREE HAEMOGLOBIN IN ROUTINE ANALYSIS

CC Otoya ^1,^✉

Introduction

Free haemoglobin (fHb) is an established independent predictor of mortality among patients on extracorporeal membrane oxygenation (ECMO) support. Previous comparisons have shown that the photometric estimation of haemolysis (H-index) on Roche c systems has correlated well with established methods of measuring fHb. Our objective was to assess the use of the SI2/H-index assay on Roche cobas c702 system as a suitable method for measuring fHb.

Methods

Precision of the SI2/H-index was assessed by using commercially available quality control materials targeted at 0.1 g/L and 0.5 g/L Hb. Accuracy was assessed by using existing external quality assurance program RCPA QAP serum indices. A patient correlation of 41 serum samples was performed with an existing photometric method on the Abbott Architect C system. A matrix comparison study was performed for measuring fHb in serum, lithium heparin plasma and EDTA plasma.

Results

Precision data for commercial controls: Level 1 mean=0.11 g/L, SD=0.923, CV=8.33%; and Level 2 mean=0.527 g/L, SD=0.858, CV=1.63%. With a method group median CV of 11.3%, all RCPA QAP samples were within performance specifications. Trueness at the critical decision point: measured average 0.56 g/L (group target of 0.58 g/L). The Roche SI2/H-index compared to the Abbott Architect C system showed good agreement: co-efficient of determination X²=0.99, average negative bias of 4.94%, regression equation y=0.99x + 0.74. Serum comparison with EDTA plasma: y=0.99x – 0.012. Serum comparison with lithium heparin plasma: y=1.03x – 0.007

Conclusions

Roche SI2/H-index assay is a capable method for measuring fHb at clinical decision points of 0.1 g/L and 0.5 g/L in serum, EDTA plasma or lithium heparin plasma.

CBR. 2021 Dec;42(Suppl 1):S14–S15.

P13 METHOD BIAS IN FERRITIN PATIENT SAMPLES

R Tiet ^1,^✉, MP Nascimento ¹, C Lynch ¹, W Louey ², D Angeleski ¹, K Sikaris ¹

Introduction

External quality assurance (EQA) samples generally show method-related bias for serum ferritin values and variability is also exhibited with the three World Health Organization (WHO) calibrators that manufacturers use for their ferritin assays. We aimed to see if the ferritin bias also exists using patient samples.

Method

Ten patient samples with Roche c701 ferritin levels that varied between 16–606 μg/L were distributed to several laboratories running a variety of common ferritin methods including the Roche c701, Beckman DxI800 and AU5800, Siemens Atellica and Abbott Alinity and Architect. We acknowledge the support of Austin Hospital, Dorevitch, Douglas Hanly Moir and Sullivan Nicolaides Pathology laboratories. The results were collated and correlation and regression analyses performed.

Results

Despite there being a very high correlation between methods (R²>0.99), there was a consistent bias between all samples for each method that varied generally between ±20% at high ferritin levels and ± 60% at low ferritin levels, when methods were compared to the all-method mean. The proportional component of the bias (regression slope) in ascending order was Beckman DX1800 (0.81), Siemens Atellica (0.92), Abbott Alinity and Architect (1.07), Beckman AU5800 (1.09) and Roche c701 (1.21).

Discussion and Conclusion

The RCPA has recommended a clinical decision limit of 30 μg/L below which iron depletion can be diagnosed. Unfortunately, a sample with a mean ferritin level of 30 μg/L could be reported between 17 μg/L (Beckman DxI800) and 40 μg/L (Roche c701), leading to over-detection or under-detection of iron deficiency. Hopefully the move to the third WHO calibrator will improve the comparability of results between laboratories.

CBR. 2021 Dec;42(Suppl 1):S15.

P14 CAPILLARY ELECTROPHORESIS MEASUREMENT OF HBA_1C USING FLUORIDE OXALATE TUBES

MB Theophilos ^1,^✉, K Parfrey ¹, RR Davidson ¹, C Lynch ¹, KA Sikaris ¹

Introduction

In January 2021 our laboratory implemented the Sebia CAPILLARYS 3 capillary electrophoresis analyser for the measurement of HbA_1c using blood from K₂EDTA tubes. We aimed to determine the validity of using fluoride oxalate (FlOx) tube samples to report HbA_1c results, in circumstances where standard EDTA tube samples are unavailable or unsuitable, or to confirm a potentially spurious EDTA result.

Methods

FlOx tube samples were selected for analysis where EDTA and FlOx tubes were collected for the same laboratory episode. A total of 105 candidate samples were selected to cover a wide clinical range of HbA_1c results based on their reported EDTA tube result, and analysed the day after the EDTA result was reported. For precision testing, one FlOx sample at each of three different HbA_1c results was analysed with 20 individual measurements.

Results

Comparison of HbA_1c measurements between EDTA and FlOx tubes showed very close correlation (r²=0.99) of the two measurements. Precision testing of FlOx samples with a single reported EDTA HbA_1c of 5.6%, 6.4% and 9.5% (NGSP) showed coefficients of variation of 1.2%, 0.86%, 0.75% respectively which were comparable to the imprecision obtained using EDTA tubes and within RCPAQAP performance specifications. The mean of these samples showed a small negative bias (5.4%, 6.2% and 9.3% respectively).

Discussion

Our results have shown that measurement of HbA_1c from FlOx tube samples using capillary electrophoresis correlates closely with standard EDTA tube results, and can therefore be used with sufficient confidence for the monitoring of long term blood glucose control when EDTA tube samples are inappropriate or unavailable. However, due to the observed negative bias in FlOx tubes, we would advise caution in relying on this method for diagnosis of diabetes, and would suggest confirmation using EDTA samples when clinically indicated.

CBR. 2021 Dec;42(Suppl 1):S15.

P15 RCPAQAP KIMMS RISK PROTOCOL

S Gay ^1,^✉, T Badrick ¹, M Whiley ²

RCPAQAP Key Incident Management and Monitoring Systems (KIMMS) Advisory Committee has defined a new risk calculation protocol that is in line with current standards and practices across Australasia.

Risk can be high because patient harm from an incident is high, probability of an incident is high, or incident detectability is low, or a combination of all three. Harm can’t be altered, however the probability of an incident occurring and the ability to detect the incident are under the control of laboratories. In summary, harm is equal to consequence times probability of occurrence. Risk is equal to the harm times the ability to detect the incident.

Consequence has been assigned by the RCPAQAP, probability is an assessment of the likelihood of an event occurring and is independent of the volume of testing being processed. This will be assigned by the laboratory for the activity and should be based on historical data analysis. These two, when combined in a 5 × 5 matrix, form the Harm assessment for laboratories.

To further assist, the incident harm is then subjected to an assessment of the incident’s detectability. If you cannot detect an event, then it is inherently more of a risk than one which you can detect. Detectability is an individual laboratory’s assessment.

To reduce Risk, laboratories can either decrease the probability of an incident or increase its detectability. The consequence is unalterable. The level of Risk for each event will help laboratories decide where best to assign their resources.

CBR. 2021 Dec;42(Suppl 1):S15.

P16 BIOCHEMICAL PROFILE OF SYNTHETIC URINE DESIGNED TO SIMULATE DRUG COMPLIANCE

D Allen ^1,^✉, C Warnholtz ¹, B McWhinney ¹, E De Waal ¹

Introduction

To evade drug detection, the substitution of patient urine with a synthetic urine product is an ever-present issue faced in the drug testing laboratory. Synthetic urines are designed to replicate many of the properties and components of human urine to deceptively provide favourable drug testing results. Synthetic urines commonly provide a negative result for illicit drugs to evade detection. Demonstrated in this study was the detection and analysis of two synthetic urine products designed to provide positive results for two opiate treatment medications, methadone and Suboxone. The presence of endogenous urine components and additional additives was also determined.

Methods

Synthetic urine samples were analysed for the presence of endogenous urine components using a range of available biochemical tests including LC-MS, LC-QTOF-MS and spectrophotometric methodologies. Urine parameters analysed included creatinine, urea, uric acid, phosphate, calcium, electrolytes, amylase, GGT, ammonia, amino acids, pH and specific gravity. The detection of unknown exogenous additives was performed using LC-QTOF-MS untargeted data acquisition, spectral library comparison and in silico fragmentation.

Results

Urine No. 1 contained the drug methadone and its primary metabolite EDDP. Urine No. 2 contained the drugs buprenorphine and naloxone. Endogenous urine components detected in both urines included only creatinine, chloride and low levels of urea and ammonia. Additional compounds detected included the yellow/orange-coloured compounds 2,4-dinitrophenol, metanil yellow, riboflavin and dinoseb. Both samples contained antimicrobial compounds including carbendazim and chlorhexidine which was surmised were added to retard organism growth in the synthetic medium. Also detected was the surfactant lauryl sulfate.

Conclusion

The detection of two synthetic urines employed to simulate compliance to opiate addiction medications provides an indication of the sophistication of synthetic urine products available. Analysis also provided insight into additional compounds added to replicate normal urine parameters as well as preserve the quality and appearance of synthetic urine products.

CBR. 2021 Dec;42(Suppl 1):S15–S16.

P17 DEVELOPMENT AND VALIDATION OF A HIGHLY SENSITIVE TANDEM MASS SPECTROMETRY METHOD FOR THE ANALYSIS OF ATROPINE IN PLASMA

O Mubaslat ^1,^2,^✉, M Fitzpatrick ^2,³

Introduction

Sublingual atropine in a small dose significantly reduces the saliva secretion in patients treated with the antipsychotic clozapine. The dose recommended is 0.6 mg. Measuring the atropine plasma concentration is used to study the pharmacokinetics of sublingual atropine to guide dose selection. A sublingual dissolving atropine tablet is under investigation. We aimed to develop a highly sensitive HPLC method for measuring atropine concentration in plasma and use the method in studying the pharmacokinetics of low dose sublingual and oral atropine.

Method

Chromatographic conditions were optimised after careful selection of analytical column and mobile phases. To extract atropine, plasma aliquots (200 μL) were mixed with 800 μL of acetonitrile containing 0.5 ng IS atropine-d5 (50 μL of the 10 μg/L in 10% methanol). The mixture was vortex-mixed for 30 seconds to precipitate plasma proteins, then centrifuged for 5 minutes at 4000 rpm. 900 μL of supernatant was transferred to a test tube and evaporated under vacuum at 50°C, then reconstructed with 200 mL of 10% methanol in water.

The method was validated to include recovery, precision, accuracy, linearity, and lower limit of quantification. The validated method was used in the analysis of blood samples collected over 8 hours in a multi-dose pharmacokinetic study in 7 participants.

Results

Atropine concentration was linear over an analytical range 0.025–10 μg/L (least-squares regression correlation coefficient R² >0.998). No ion potentiation or suppression at the atropine retention time was recorded after injecting an extracted blank matrix sample while running a post column infusion of atropine into the mass spectrometer via T-junction. This method displayed full agreement with the US FDA bioanalysis guidelines. The pharmacokinetics of sublingual and oral atropine in 7 participants were determined.

Conclusion

We developed a highly sensitive and accurate method for the measurement of atropine in plasma and used the method in determining the pharmacokinetics of atropine.

CBR. 2021 Dec;42(Suppl 1):S16.

P18 RETROSPECTIVE REVIEW OF THE INCIDENCE OF VITAMIN C DEFICIENCY IN AN AUSTRALIAN POPULATION

S Browning ^1,^✉, B Teis ¹, M Henman ¹, R Bain ¹, K De Silva ¹, T Harrower ¹, D Holman ¹, S MacDonald ¹, M Najdovski ¹, K Sharvell ¹, C Sklavos ¹, S Stewart ¹, G Daley ¹, R Flatman ¹, L Price ¹, D Kanowski ¹, G Ward ¹

Introduction

Recent studies in surgical and psychiatric patients have determined that vitamin C deficiency may be more common in Australia than expected. Symptoms of scurvy can occur with plasma levels <11 μmol/L. Plasma vitamin C levels over the past 7 years were examined to determine the prevalence of deficiency in our patient population.

Methods

Vitamin C was measured in plasma using the Chromsystems vitamin C HPLC method with UV detection. Vitamin C levels from January 2014 to June 2021 were analysed. Lithium heparin plasma samples were collected from fasting patients and frozen and protected from light within 2 hours of collection.

Results

The study included 56,081 samples. Results ranged from <3 μmol/L to 66,028 μmol/L. 14,006 (25%) were below the lower limit of the RI (26–85 μmol/L), 35,625 samples (63.5%) were within the RI and 6,450 samples (11.5%) were above the upper limit of the RI. 2657 samples (4.7%) were >100 μmol/L and 5995 samples (10.7%) were below the deficient level of 11.4 μmol/L.

Conclusion

This study demonstrated that vitamin C deficiency is more common now (10.7%), than in our previous study of 9976 samples in 2014 in which 6% were <11.4 μmol/L. Intravenous administration has also increased, with 4.7% of samples >100 μmol/L compared to 2.8% in 2014. Bhattacharya analysis revealed a reference interval of 25.1–158.5 μmol/L which is much higher than 6.0–87.2 μmol/L reported in 2014 and reflects intravenous administration rather than nutritional supplementation.

CBR. 2021 Dec;42(Suppl 1):S16.

P19 MACHINE LEARNING-BASED ALGORITHM FOR LIQUID CHROMATOGRAPHY-TANDEM MASS SPECTROMETRY (LC-MS/MS)-BASED URINE DRUG SCREENING

Wai-yuen Yu ^1,^✉, E Yi-ling Wong ¹, Chun-yiu Law ¹, F Chi-kin Wong ¹, Ching-wan Lam ^1,²

Introduction

Liquid chromatography-tandem mass spectrometry (LC-MS/MS) is the major laboratory method used in urine drug screenings. The positive identification of a drug would rely on the integration of a multi-dimensional LC-MS/MS data, including retention time, accurate mass, fragmentation pattern, peak shape, etc. and a manual interpretation is often required to safeguard the quality. We aim to create a machine learning (ML)-aided algorithm to assist manual interpretation of the LC-MS/MS data.

Methods

The drug and drug metabolites findings were extracted from LC-MS/MS data in the past three years for all reported urine drug screening cases. LC-MS/MS features (e.g. retention time, accurate mass, fragmentation pattern, isotopic match, etc.) were used to build a ML algorithm using Synthetic Minority Over-sampling TEchnique (SMOTE) k-nearest neighbor (KNN) classifier. The performance of the ML model was tested by comparing the ML results with two independent interpreters (a Chemical Pathologist and a Chemical Pathology trainee) using independent test sets.

Results

Total 99,560 reported items (including parent drugs and drug metabolites) were used to train the ML algorithm. Using the test sets, the overall accuracy, sensitivity and specificity were 0.88, 0.85 and 0.96 respectively.

Conclusions

This study showed promising performance with reasonable accuracy, sensitivity and specificity which can enhance the internal consistency among pathologists for the interpretation of urine drug screening results.

CBR. 2021 Dec;42(Suppl 1):S16–S17.

P20 PARAPROTEIN INTERFERENCE IN AUTOMATED PHOSPHATE MEASUREMENT: A CASE REPORT

H Leppinus ^1,^✉, SDC Thomas ¹

Introduction

The effect of paraprotein interference on automated chemistry and immune-assays is well documented. Commonly affected laboratory tests include bilirubin, phosphate, HDL, GGT, urea, CRP and glucose. We report a case of repeated spurious high phosphate in a patient with IgG paraproteinemia.

Methods

During routine testing of serum from a 46-year-old male, the automated chemistry analysers (Abbott Alinity c) were unable to produce a phosphate measurement due to exceeding the optical limit of absorbance. The patient was known to have an IgG paraproteinemia (IgG 68 g/L, total protein 133 g/L) with normal calcium and renal function. Historically, phosphate result for this patient has not been reported due to interference. Previously, measurement on the Roche assay reported significantly high phosphate. The patient sample was serially diluted with saline in an attempt to obtain a phosphate result. The index patient’s serum was used to spike five samples with normal results to investigate the effect of IgG paraprotein on the phosphate result.

Results

Consistent results were obtained when the sample was diluted 1:4 and 1:8 (0.88 mmol/L), total protein 70 g/L. The mean phosphate in the five normal samples was 1.01mmol/L and after addition of the index patient’s serum, the phosphate mean increased to 1.68 mmol/L (reference interval 0.8–1.5 mmol/L).

Conclusion

High concentration of paraprotein interferes with photometric phosphate measurement in routine automated chemistry, most likely due to spectral interference. Instructions for Users in automated chemistry assays commonly list paraprotein as an interfering agent. In the Abbott Alinity c phosphate assay, the protein concentration at which interference occurs is listed as 120 g/L, while the Roche assay does not specify a concentration. Recognition of spurious results due to interferents will avoid unwarranted investigations.

CBR. 2021 Dec;42(Suppl 1):S17.

P21 TIME FOR A CHANGE TO TUBE CHOICE AND PROCESSING OF SAMPLES IN MEASUREMENT OF GLUCOSE FOR GDM AND OGTT

G Dimeski ^1,^✉

Introduction

The NaF/Kox is the recommended tube of choice in glucose estimation for GDM and OGTT even with the knowledge it does not stop the glycolysis process for up to an hour. The common practice is to batch the tubes and centrifuge them hours later causing underestimation of DM prevalence and poorer patient outcomes from delayed DM diagnosis. The Greiner FC Mix tube is effective in stopping glycolysis but the uptake of the tube has been limited. Alternative suitable tubes are the rapid clotting serum tubes which have not been evaluated or utilised.

Methods

A pilot stability study was conducted from a single participant using BD tubes: lithium heparin (PST); NaF/Kox and Rapid Serum Tube (RST). Three sets of each tube types were collected. Set 1, baseline tubes were centrifuged within 10 mins of collection (10 mins, 3000g) and analysed immediately, then stored at 21°C for 4, 7 and 24 hrs and re-analysed. The 2^nd and 3^rd sets were left uncentrifuged for 1 and 2 hours before centrifugation. All samples were analysed on a Siemens Atellica chemistry system in triplicate and the average was used in estimating the percentage decrease from the baseline result.

Results

The baseline results (mmol/L) were: NaF/Kox – 5.745; PST – 5.746; RST – 5.740. The respective percentage decreases post 4, 7 and 24 hrs were: NaF/Kox – 1.3, 2.0 and 1.9; PST – 1.9, 2.9 and 6.4; RST – 0.6, 0.6 and 0.4. The respective percentage decreases post 1 and 2 hours delayed centrifugation were: NaF/Kox – 6.6 and 8.9; PST – 5.8 and 8.6, RST – 5.9 and 8.1.

Conclusions

The RST tube offers an improved alternative in stability when tubes are centrifuged within 10 mins. Additionally, the gel barrier eliminates remixing concerns, has decreased haemolysis, is becoming commonly used and additional tests can be performed for patient and service provider convenience.

CBR. 2021 Dec;42(Suppl 1):S17.

P22 DIRECT BILIRUBIN: ROCHE V RANDOX – A COMPARATIVE STUDY

C Cocotsi ^1,^✉, CF Lim ², A Connell ¹, J Cleary ¹, M Rozkin ²

Introduction

Within EHP we have been utilising the Roche direct bilirubin (diazo) assay on the Roche Cobas c702 platform. This assay suffers from interference at a low degree of haemolysis. The Paediatric Unit at Box Hill Hospital required a solution, as 98% of the neonatal samples were haemolysed sufficiently to exclude direct bilirubin measurement, requiring recollection. The Randox vanadate-oxidation assay claims a higher haemolysis index (HI) cut-off.

Objective

To compare the Roche and Randox assays to determine which assay is best fit-for-purpose in providing an adjunct to aid in the clinical management of neonatal patients.

Methods

21 Neonatal and 13 adult samples with a haemolysis index of <25 mg/dL were collected and analysed at Box Hill and Liverpool Hospitals using the vanadate-oxidation assay. 31 samples with a haemolysis index of <25 mg/dL were evaluated using both assays. Patient correlations were performed on 35 samples with haemolysis index >25 mg/dL. 4 samples with high bilirubin and haemolysis index of 0 were spiked with different dilutions of haemosylate, to determine the effect of haemolysis and the percentage of recovery.

Results

There was a good correlation of results between the Roche and the Randox assay when the HI was <25 mg/dL. At a HI of >25 mg/dL, recovery of direct bilirubin by the Roche assay decreased significantly. Further, there was no discernible correlation between haemolysis and the result obtained. The Randox assay, however maintained recovery to an HI value of 465.

Conclusion

The Randox vanadate-oxidation method was implemented in our pathology laboratory with great success and favourable responses from our paediatricians, as 96% of direct bilirubin requests now give a numerical result.

CBR. 2021 Dec;42(Suppl 1):S17.

P23 DRUG USE AND VIRAL INFECTIONS

M-S Shi ^1,^✉

Introduction

In European and American countries, because of drug abusers and sharing drug injection needles, the infection rate of HIV and hepatitis C infections has increased rapidly maybe 480,000–760,000 people infected with hepatitis C virus. There are about 50% (130,000) drug abuse injections, infected with HIV and hepatitis C virus at the same time. Although Taiwan is a country that strictly controls drugs, and if drug users are provided with free injection needles, whether their infection rate has decreased because of this.

Methods

87 people have been included in the Drug Abuse Program, aged 28–56 years, 74 males and 13 females. All 87 people agreed to check blood B and C hepatitis virus antigens and antibodies, HIV antibodies and syphilis serum.

Result

The results of this study found that up to 81 people (93.1%) had shared injection needles, and the positive rate of hepatitis B antigen was 8.0%, the positive rate of hepatitis B antibody was 90.8%, the positive rate of hepatitis C antibody was 87.4%. For syphilis the positive rate was 13.4%, while the HIV antibody positive rate was 10.3%. If a drug abuser is infected with hepatitis C, based on Fisher’s correlation statistical analysis, the gender (p=0.359), hepatitis B antigen positive (p=0.214), hepatitis B antibody positive (p=1.0), HIV antibody positive (p=0.596), syphilis positive (p=0.443), education level (p<0.001), sharing needles (p=0.025).

Conclusion

This study showed that the infection rate of drug abusers infected with hepatitis C was 87.4%, showing a high infection rate. Taiwan generally has an infection rate of about 6%. The study also found that the sharing of injection needles is closely related to the education level of drug abusers.

CBR. 2021 Dec;42(Suppl 1):S17–S18.

P24 DETECTION OF RARE FS A1AT VARIANT: A CASE STUDY

WHH Lau ^1,^✉, P Pavone ¹, JCG Doery ^1,², ZX Lu ^1,²

Introduction

Alpha-1 antitrypsin (A1AT) is a protease inhibitor (Pi) synthesised and exported by the hepatocytes. Genetic variants (>120) exist and some alleles may lead to decreased A1AT production, accumulation of abnormal protein in the hepatocyte causing liver damage or impaired activity in inhibiting neutrophil elastase (NE) or proteinase 3 (PR3) causing lung disease. A1AT deficiency is an under-diagnosed disease. Early detection of abnormal A1AT variant may prevent development of A1AT deficiency-associated diseases.

Patient

A 41y old man was referred for phenotype because of persistent isolated GGT elevation and borderline low A1AT concentration.

Results

The A1AT concentration was 1.0 g/L (reference interval 1.0–2.4 g/L) on Beckman Immage. The A1AT phenotype by isoelectric focussing using Hydragel 18 on Sebia Hydrasys2 revealed a characteristic pattern of three double-bands consistent with the F variant and a slow migrating S variant. The M (normal variant) was not present.

Discussion and Conclusions

The combinations of M, S (produces low levels with moderately low activity) and Z (produces very low level with little to no activity) variance encompass over 98% of the world population. The F is one of the rare variants. It produces a normal quantity of A1AT protein but the protein does not function properly, resulting in reduced ability to inhibit NE but not PR3. Our complex heterozygote FS findings are consistent with the borderline normal A1AT level. Isolated elevation in GGT is not a manifestation of A1AT deficiency, suggesting the patient’s liver dysfunction is due to other aetiology. In conclusion, the rare FS phenotype may have normal A1AT levels but is associated with increased risk for emphysema, especially in smokers, but not usually with liver disease. The clinical significance and appropriate management of borderline normal or low A1AT levels is greatly illuminated by A1AT phenotyping.

CBR. 2021 Dec;42(Suppl 1):S18.

P25 INTERFERENCE IN THE MEASUREMENT OF THIOPURINE METABOLITES IN INFANTS

M Mapp ^1,^✉, R Prentice ^2,^3,⁴, J Smith ⁵, EK Wright ^2,⁶, E Flanagan ^2,⁶, AR Ross ², W Hardikar ^2,⁶, Q Lam ¹, SJ Bell ^3,⁴, C Trambas ¹

Introduction

Laboratory detection of interferences requires close liaison with clinicians in order to identify clinically discordant results. We were contacted by our gastroenterology unit questioning the veracity of detectable 6-methylmercaptopurine (6-MMP) results in two 6-week-old babies on two separate occasions. While the mothers were being treated with thiopurine, one baby was fed with infant formula and the other breast-fed at the time of testing. After confirmation by repeat analysis, we investigated a possible chromatographic interference in neonatal samples.

Methods

Thiopurine metabolites were measured by high performance liquid chromatography and UV detection using a modified Dervieux-Boulieu method. Whole blood samples from a cohort of 10 infant subjects not exposed to thiopurine were tested along with the exposed study subjects. All infants tested were currently fed breast milk, formula or a combination. Six infant formulae were analysed using the same method.

Results

Along with the index cases, 2 of the 10 subjects not exposed to thiopurine showed significant co-eluting peaks within the interference window for 6-MMP detection. One subject was exclusively fed with formula, and the other a combination of breast milk and formula. All commercial formulae showed significant peaks in the window correlating with the peaks in the subjects’ chromatograms.

Conclusions

A peak co-eluting with 6-MMP was identified in blood samples from babies not exposed to thiopurine, as well as in infant formula. We hypothesise this co-eluting peak may be a nucleotide found naturally in breast milk and fortified in infant formula. Similar properties of the nucleotides uridine monophosphate, adenosine monophosphate, inosine monophosphate, guanosine monophosphate and cytidine monophosphate may lead to co-elution and false detection of 6-MMP in babies fed with breast milk or infant formula. Further studies are in process to definitively identify this co-eluting nucleotide.

CBR. 2021 Dec;42(Suppl 1):S18.

P26 ASSESSMENT OF TOTAL BILIRUBIN ON THE GEM 5000 BLOOD GAS ANALYSER

B Fitzgerald ¹, S Newton ¹, MM Salib ^1,^2,^✉, PE Hickman ^1,²

Introduction

There are minimal Australian data with regards to whole blood total bilirubin (Tbili) testing on the Werfen GEM 5000 blood gas analyser. We assessed Werfen Gem 5000 TBili performance against the Abbott Architect platform.

Method

A total of 24 samples from the 2020 RCPA EQAP Neonatal Bilirubin (NBili) program; 17 whole blood heparinised samples spiked with either patient plasma with elevated TBili or EQAP samples from the NBili program; and 20 whole blood heparinised samples from liver disease patients, were tested over a 1-month period on the GEM 5000 Cooximetry method and Abbott Architect c16000 colorimetric assay. Results were compared using Passing-Bablok analysis and RCPA-QAP analytical performance specification (APS).

Results

Tbili results ranged from 80.4–361 μmol/L for EQAP NBiLi samples, 148–393 μmol/L for spiked samples and 31–599 umol/L for liver disease patients. When compared to Abbott Architect, GEM 5000 TBili gave an average bias of −2.40%, −6.09% and −19.06% for EQAP samples, spiked samples and liver disease samples, respectively. Passing-Bablok analysis versus Architect TBili showed a slope and correlation coefficient of 0.896 (0.998) for EQAP samples, 0.953 (0.996) for spiked samples and 0.8556 (0.996) for liver patients results.

RCPA EQAP NBili samples showed conjugated bilirubin range of 5–46%, while liver patient samples were from adult population with mostly conjugated hyperbilirubinaemia and this could account for the observed difference in performance. The latter group showed haematocrit range of 17–53% and Abbott Architect haemolysis index of 0.04–0.44.

Conclusion

Despite acceptable correlation for the RCPA EQAP samples and the spiked whole blood samples, GEM 5000 TBili from liver disease patients gave negative bias exceeding APS and is hence not suitable for testing in this population. Further comparison using patient samples with unconjugated bilirubinaemia is required to confirm appropriateness of use in neonatal population.

CBR. 2021 Dec;42(Suppl 1):S18–S19.

P27 LIQUID-CHROMATOGRAPHY MASS SPECTROMETRY METHOD COMPARISON FOR THE QUANTIFICATION OF 24,25-DIHYDROXYVITAMIN D3 IN PATIENTS AND EXTERNAL QUALITY ASSURANCE SAMPLES

S Zelzer ^1,^✉, C Le Goff, S Peeters ², C Calaprice ², A Meinitzer ¹, D Enko ¹, W Goessler ³, M Herrmann ¹, E Cavalier ²

Introduction

In-house developed liquid-chromatography mass spectrometry (LC-MS/MS) methods are used more and more frequently for the simultaneous quantification of vitamin D metabolites. Among these, 24,25-dihydroxyvitamin D3 (24,25(OH)₂D₃) is of clinical interest. This study assessed the agreement of this metabolite in two validated in-house LC-MS/MS methods.

Methods

24,25(OH)₂D₃ was measured in 20 samples from the vitamin D external quality assurance (DEQAS) program and in a mixed cohort of hospital patients’ samples (n=195) with the LC-MS/MS method at the Medical University of Graz (LC-MS/MS 1) and at the University of Liège (LC-MS/MS 2).

Results

In DEQAS samples, 24,25(OH)₂D₃ results with LC-MS/MS 1 had a proportional bias of 1.0% and a negative systemic difference of −0.05%. LC-MS/MS 2 also showed a proportional bias of 1.0% and the negative systemic bias was −0.22%. Comparing the EQA samples with both methods, no systemic bias was found (0.0%) and the slope was 1%. The mean difference of 195 serum sample measurements between the two LC-MS/MS methods was minimal (−0.2%). Both LC-MS/MS methods showed a constant bias of 0.31 nmol/L and a positive proportional bias of 0.90%, respectively.

Conclusion

This study is the first to assess the comparability of 24,25(OH)₂D₃ concentrations in a mixed cohort of hospitalised patients with two fully validated in-house LC-MS/MS methods. Despite different sample preparation, chromatographic separation and ionisation, both methods showed high precision measurements of 24,25(OH)₂D₃. Furthermore, we demonstrate the improvement of accuracy and precision measurements of 24,25(OH)₂D₃ in serum samples and in the DEQAS program.

CBR. 2021 Dec;42(Suppl 1):S19.

P28 FALSE POSITIVE INTERFERENCE ON BECKMAN COULTER TESTOSTERONE ASSAY IN PROSTATE CANCER PATIENTS

S Mukerji ^1,^✉, P Bonnitcha ^1,², R Prasad ¹, A Screnci ¹, GRD Jones ¹

Introduction

Higher than expected testosterone levels were observed at SydPath with the Beckman Coulter (BC) total testosterone assay in prostate cancer patients receiving androgen deprivation therapy. Interference with a metabolite of Abiraterone acetate (therapy used in metastatic castration-resistant prostate cancer) was considered as a possible cause. We report on findings from a local protocol to identify possible interferences.

Methods

For oncology patients with initial BC testosterone results ≥1.7 nmol/L samples were routinely sent for measurement on Roche or LCMSMS testosterone assays. For a subset of patients, clinical notes were reviewed to determine whether patients were recorded as being on Abiraterone at the time of sample collection. Other available medication data was also recorded where available.

Results

Of a total of 156 samples with duplicate measurements, 63 (40%) had Roche or LCMSMS results <0.3 nmol/L. The rate of low results was similar for the Roche and the LCMSMS assays. 16/29 (55%) patients with very low Roche or LCMSMS results were recorded to be receiving Abiraterone on the date of sample collection. 0/24 patients without low Roche or LCMSMS results were recorded as taking Abiraterone. No other medications appeared to be related to the method differences.

Conclusions

A difference between the BC and other assays is common in samples from patients with androgen deprivation therapy. As the comparison methods included LCMSMS, it is likely the differences are due to positive interference in the BC assay. We observed that approximately half the patients with assay differences and information available were taking Abiraterone. By contrast, Abiraterone was not identified in any patients without low Roche or LCMSMS values. As Abiraterone itself is known to not affect the BC assay, interference from a metabolite is likely, as was seen for an older Abbott testosterone assay. Other causes for differences should also be considered.

CBR. 2021 Dec;42(Suppl 1):S19.

P29 STUDY OF CIRCULATING MIR-519D AND MIR-26B AS POTENTIAL THERAPY-INDUCED AUTOPHAGY CHEMORESPONSE MARKERS TO DOXORUBICIN-BASED TRANSARTERIAL CHEMOEMBOLISATION IN HEPATOCELLULAR CARCINOMA PATIENTS

OAI AbdelKarem ^1,^✉, OS Elassar ²

Introduction

Doxorubicin-based superselective transarterial chemoembolisation (ssTACE) has emerged as a bridging therapy for early stages of hepatocellular carcinoma (HCC) awaiting liver transplantation. Clinical resistance to doxorubicin is a major problem that involves multiple molecular mechanisms, including therapy-induced autophagy. Recently, several miRNAs as miR-519d and miR-26b have been found to be involved in autophagy modulation. This study aims at assessing the diagnostic accuracy of circulating miR-519d and -26b as potential therapy-induced autophagy markers of response to doxorubicin-based ssTACE in Egyptian HCC patients.

Methods

Forty-seven HCV-infected HCC patients were referred for doxorubicin-based ssTACE, and after three months follow-up period resulted in forty-one responders and six non-responders based on modified response evaluation criteria in solid tumors (mRECIST). Twenty matched individuals were included as a reference group for miRNA expression calculation. Pre- and post-ssTACE tumor volume and viability were assessed using a computerised tomography scan. Circulating pre-ssTACE miR-26b and -519d expressions were determined using real-time quantitative polymerase chain reaction

Results

Circulating miR-26b but not miR-519d was significantly upregulated in HCC patients (median=4.56) compared to the reference group (median=1.81), p=0.017. However, there was no statistically significant difference between responders and nonresponders in terms of miR-519d or miR-26b expression level (p=0.5, p=0.6), respectively. Post-ssTACE tumor volume was significantly lower than pre-ssTACE in responders, p <0.001. None of the miRNAs (miR-519d,-26b) was found as a significant discriminator to chemoresponse by Receiver operating characteristics (ROC) curve analysis with an area under the curve (AUC) of ( 0.575, 0.557) respectively and p-value of (0.556, 0.463) respectively.

Conclusion

Although miR-26b was significantly upregulated in HCC cases compared to the reference group, there is no significant association between the expression level of both miR-519d or -26b and response to Doxorobicin-based ssTACE. A large scale study involving other chemosensitive/chemoresistant miRNAs is highly recommended

CBR. 2021 Dec;42(Suppl 1):S19–S20.

P30 EVALUATING DIAGNOSTIC TIME POINTS OF LACTOSE TOLERANCE TESTS TO REDUCE TEST DURATION TO 60 MINUTES AND IMPROVING DIAGNOSTIC RATES

ND Palmrose ^1,^✉, SDC Thomas ¹

Introduction

The prevalence of lactose intolerance, due to deficiency or absence of lactase, is about 75% worldwide, with a wide ethnic variation. The gold standard of diagnosis is a duodenal biopsy for lactose activity. A genetic test is now available for a mutation that enables lactase activity to persist in adulthood. The dynamic study of glycaemia following ingestion of a lactose load is the most commonly used diagnostic test in the community. The following audit was performed to assess the diagnostic rates at different time points of sample collection and to correlate with symptoms.

Method

Lactose tolerance test (LTT) results obtained over a 2-year period were collated and the laboratory cut-point of rise in serum glucose of >1.5 mmol/L was used to determine the diagnostic rates using the 30 min sample and the 60 min sample. Symptoms reported by the patient after ingesting the lactose load were correlated with the diagnostic rates.

Results

Diagnostic rates taken at 60 min showed that 71% of patients would be given a positive pathological diagnosis compared to only 38% if the 30 min time point was used for the decision point. Diagnostic rates using the 30 min time point or any other time point were comparable (38%). Symptoms noted during the tests were similar at every time point, approximately 45–50% of patients in each diagnostic group. Among reported symptoms, abdominal pain and nausea had the highest rates of reporting.

Conclusions

The diagnostic rate of lactose intolerance was higher when the 60 min sample was used. This appears to be a more acceptable rate of diagnosis due to the high pre-test probability of our subjects, in line with published results that suggest a peak glucose at the 60 min time point in lactose tolerant individuals. Some studies suggest using the 30 min or a 45 min sample however this may lead to false negative results.

CBR. 2021 Dec;42(Suppl 1):S20.

P31 EVALUATION OF BD BARRICOR™ BLOOD COLLECTION TUBES – IMPACT ON TURNAROUND TIMES AND SAMPLE QUALITY

S Callen ^1,^✉, D Layugan ¹, T Wagner ¹

Introduction

Relatively recently Becton-Dickinson (BD) introduced Vacutainer® Barricor™ plasma blood collection tube. This tube utilises a mechanical separation device in place of the usual separator gel which allows shortened 3 min centrifugation. Comparison was made to BD Vacutainer® PST™ lithium heparin tube with gel separator. Comparison of turnaround times (TAT) and sample quality was made.

Methods

100 of each tube type, all collected by Emergency Department medical staff within The Queen Elizabeth Hospital were compared. TAT from time of receipt in laboratory to time of delivery of results to ED medical staff were evaluated for routine biochemical profile and cardiac troponin T testing. Sample quality determined by plasma haemolysis (H) index was compared by evaluating the distributions of indices and rates of non-haemolysed versus haemolysed samples.

Results

PST™ lithium heparin tubes showed 6 periods out of 21 with ACHS target below 80% for routine biochemical profile; for troponin, 8 periods out of 21 did not meet 90% ACHS goal.

Barricor™ tubes showed 1 of the 24 periods monitored with <80% samples meeting ACHS target for routine biochemical profile testing; for troponin, 3 periods were below the 90% ACHS target.

Distribution of plasma (H) index showed more Barricor™ tubes with elevated H index compared with PST™ lithium heparin tubes. Rate of samples with no measurable haemolysis (H<5) was 28% for PST™ lithium heparin tubes, while for Barricor™ tubes it was 14%. 20% of Barricor™ tubes displayed H index >50, while only 11% of PST™ lithium heparin tubes had H index >50.

Conclusion

We show a definite improvement in ability to meet ACHS TAT goals with BD Vacutainer® Barricor™ plasma blood collection tube. TAT was more consistent across each 24-hour period and more often met ACHS goals. Rates of haemolysis were higher in BD Vacutainer® Barricor™ plasma compared to BD Vacutainer® PST™ lithium heparin plasma.

CBR. 2021 Dec;42(Suppl 1):S20.

P32 USING MICROFLOW LIQUID CHROMATOGRAPHY AND HIGH-RESOLUTION MASS SPECTROMETRY TO IDENTIFY A COMMONLY OCCURRING INTERFERENT IN A HIGH SENSITIVITY OESTRADIOL ASSAY

C Hodgkins ^1,^✉, S Thompson ¹, R Sahertian ¹, K Mantik ¹, AR Horvath ¹

Introduction

Mass-spectrometry (MS)-based assays have become the recommended method for measurement in individuals expected to have low oestradiol (E2) concentrations. Whilst MS-based E2 assays tend to avoid cross-reactivity with analytes of similar structure, interferences from isobaric compounds remain possible. We previously reported on the presence of a commonly appearing interfering substance in our MS E2 assay which we were able to remove by an extended gradient chromatographic separation. The aim of our study was to identify this interfering substance using high resolution, accurate mass MS technology.

Methods

The usual interference was present in several samples collected from a 76-year-old female patient. Aliquots of serum extracts were separated by reversed-phase microflow chromatography and analysed by a combination of data-dependent (DDA) and data-independent-acquisition (DIA) on a quadrupole-time-of-flight (QTOF) tandem mass spectrometer, with data processing by manual review and automated library searching of MS/MS spectra.

Results

Used as a pseudo precursor ion scan, DIA analysis revealed the interfering compound precursor m/z (271.1) to be an in-source fragment of an unknown precursor with m/z 408.1. Cross-referencing the high-resolution MS/MS spectrum of the 408.1 m/z precursor with a library of drugs and their metabolites returned a library match score of 98.4 (out of 100) for ezetimibe, a commonly prescribed lipid-lowering medication.

We retrospectively reviewed the clinical information from 20 other samples in which the interference was present. Medication history was available for 13 of these patients – all of whom were also prescribed ezetimibe.

Conclusion

Using a combination of high resolution, accurate mass MS measurements on a single sample that returned an elevated result, we have confirmed a commonly prescribed drug interferes with our sensitive E2 assay. This knowledge has helped us optimise our routine method by utilising extended gradient chromatography, while still reducing total analysis times.

CBR. 2021 Dec;42(Suppl 1):S20–S21.

P33 SAMPLE TYPE AND PRE-ANALYTICAL PROCESSING EFFECTS ON GLUCOSE CONCENTRATION

M Charan ^1,^✉, P Ward ¹, D Chesher ¹

Introduction

A major source of pre-analytical error in measuring glucose is loss of glucose from blood specimens through glycolysis. Glucose is lost from whole blood samples at a rate of 5–7% per hour at room temperature. Delays in centrifugation and separation of the blood samples, particularly apparent in a large hospital setting, facilitate this loss of glucose. We assessed the effects of five different collection tubes and pre-analytical processing procedures on glucose measurement.

Methods

Blood was collected from 43 laboratory volunteers into five collection tubes using standard procedures. Fluoride oxalate tubes (FO) were chilled on ice before and after blood collection. Whole blood was taken from the lithium heparin tube for testing on an i-stat immediately after collection. Lithium heparin (LH), FO, one Granulated Fluoride Citrate (FC) and serum tubes were centrifuged 30 minutes after collection whilst the remaining FC tube was centrifuged 2 hours later. Glucose was analysed on the Abbott Architect.

Results

There was good correlation of glucose measurements between and across the different tube types. However, glucose measurements in the FC tubes were 0.52 mmol/L higher than those measurements made from serum, lithium heparin plasma and on the i-stat. Furthermore, glucose concentration in the FC tubes was on average 0.29 mmol/L higher than the FO tubes collected on ice. There was no difference in the glucose results coming from the two FC tubes that were collected at the same time but processed two hours apart.

Conclusion

The immediate centrifugation of LH tubes might lead to a better recovery of glucose compared to FO tubes. It would seem that FC tubes return glucose results significantly higher than the FO tubes that were collected according to the HAPO study protocol. Acid-induced cellular damage has been hypothesised as a mechanism that may lead to fluid shifts and an increase in plasma glucose.

CBR. 2021 Dec;42(Suppl 1):S21.

P34 USING URINE CORTISOL:CREATININE RATIOS TO IMPROVE THE UTILITY OF 24-HOUR URINE TESTING – A RETROSPECTIVE ANALYSIS

M D’Souza ^1,^✉, F Nouri-Girones ¹, C Woods ¹, MM Salib ^1,²

Introduction

24-hour urinary free cortisol (UFC) excretion is a diagnostic screening test used in the setting of suspected Cushing’s syndrome. However, the validity of the test can be challenged by improper collection and concurrent medical conditions. We assessed the utility of adding a calculated UFC: creatinine ratios to all reports and explored the factors that may explain discordant results.

Methods

We conducted a retrospective data audit from January 2014 to December 2019, including 937 episodes of 24-hour UFC tests performed at ACT Pathology by HPLC method. A UFC:creatinine ratio was calculated and the medical records were further reviewed to assess discordant results.

Results

There was an agreement between 24-hour UFC and UFC:creatinine ratios in 834 (89%) of results. 103 (11%) episodes were discordant, of which 25 (24%) episodes were explained by either an over- or under-collection. Of the remaining episodes, 38 (37%) involved subjects with a history of Cushing’s syndrome currently receiving medical management, and 11 (11%) cases included subjects with intermittent steroid use, chronic pain or obesity. The chronic pain and obesity groups both demonstrated a high 24-hour UFC with normal UFC:creatinine ratio. Clinical data for the remaining 29 (28%) subjects was not available.

Conclusion

Inclusion of a calculated UFC:creatinine ratio to all reports would correct for errors of urine collection and increase the sensitivity and specificity of the test. Factors such as obesity, chronic pain and intermittent steroid use were associated with discordance and need to be interpreted in the clinical context. In these cases, further testing is warranted to clarify the significance of laboratory findings.

CBR. 2021 Dec;42(Suppl 1):S21.

P35 EXTERNAL QUALITY ASSURANCE PROGRAM FOR THE DETECTION OF CORONAVIRUS SARS COV-2

G Moyo ^1,^✉, L Wienholt ¹, P Graham ¹

Introduction

The RCPAQAP is internationally recognised as an External Quality Assurance (EQA) provider for the provision of proficiency testing programs. The transmission of the SARS-CoV-2 virus has had a significant global impact. As SARS-CoV-2 serological assays became available to market, RCPAQAP developed an EQA program for interlaboratory comparison of SARS-CoV-2 antibody testing.

Method

Two surveys were distributed to a total of 73 Australian and international laboratories for SARS-CoV-2 antibody testing (IgA, IgG, IgM, total antibody and point-of-care IgG and IgM), as per routine laboratory protocols. The surveys consisted of two negative and two positive commutable serum samples. Qualitative and quantitative data were analysed.

Results

Laboratories used a mixture of commercial and in-house assays for testing. There was a high concordance in reporting SARS-CoV-2 IgG, total antibody and point-of-care IgG results. Discrepant reporting was seen for SARS-CoV-2 IgA and IgM antibodies.

Conclusion

The RCPAQAP Serology Coronavirus SARS-CoV-2 antibodies program offers quality assurance that is an effective tool for verifying the accuracy of testing procedures and interpretation of those results. This program can assist laboratories in assessing their performance for patient reporting and is a means for laboratories of evaluating new assays by method comparison using characterised samples.

CBR. 2021 Dec;42(Suppl 1):S21.

P36 ASSESSMENT OF ANALYTICAL BIAS IN FERRITIN ASSAYS AND THEIR IMPACT ON FUNCTIONAL REFERENCE LIMITS

KW Choy ¹, G Sezgin ², N Wijeratne ^3,⁴, J Calleja ⁵, R Liwayan ^1,^✉, G Rathnayake ^6,⁷, R McFarlane ⁶, A McNeil ³, JCG Doery ⁸, Z Lu ⁸, C Markus ⁹, TP Loh ¹⁰

Introduction

Serum ferritin is currently the recommended laboratory test to investigate iron deficiency. There have been efforts to standardise serum ferritin assays with implementation of traceability to the World Health Organization reference standard. We evaluate the analytical bias among five widely used commercial ferritin asMonasays in Australia. The relationship between serum ferritin and erythrocyte parameters was recently explored to derive functional reference limits.

Methods

Serum ferritin is currently the recommended laboratory test to investigate iron deficiency. There have been efforts to standardise serum ferritin assays with implementation of traceability to the World Health Organization reference standard. We evaluate the analytical bias among five widely used commercial ferritin assays in Australia. The relationship between serum ferritin and erythrocyte parameters was recently explored to derive functional reference limits.

Results

At clinically relevant ferritin decision points, compared to the Beckman method, the Roche assay showed higher results ranging from 6 μg/L (31%) at the lowest decision point to 575 μg/L (57%) at the highest decision point. In contrast, the Ortho method underestimated ferritin results at lower decision points of 20 and 30 μg/L, with estimated ferritin results of 16 μg/L (−19%) and 27 μg/L (−12%) respectively. The Abbott and Siemens assays showed a positive bias which was introduced at differing decision points. The comparison of the Siemens and Ortho methods present similar inflection points between the two assays in the establishment of functional reference limits for serum ferritin.

Conclusion

There remain significant biases among some of the commonly used commercial ferritin assays in Australia. More studies are needed to assess if functional reference limits are a way to overcome method commutability issues.

CBR. 2021 Dec;42(Suppl 1):S22.

P37 A SNAPSHOT OF FOETAL FIBRONECTIN TESTING ACROSS AUSTRALASIA

W Punyalack ¹, B Limoux ^1,^✉, R De Leon ¹, P Graham ¹

Introduction

Preterm birth can result in health consequences for neonates, including mortality. Early identification of preterm delivery allows opportunity to implement delay interventions, improving outcomes for neonates. Foetal fibronectin (fFn) is a glycoprotein produced between the amniotic sac and the endometrium. This is detectable in vaginal fluid close to full-term birth. Low concentrations of fFn (<50 ng/mL) during early gestation is predictive of a lowered risk of preterm delivery. PoCT enables detection and quantification of fFn. While PoCT devices are widely in use in maternity wards throughout Australasia, until recently, no local external quality assurance (EQA) program was readily available to monitor the quality of these results. In 2021, the RCPAQAP launched a fFn EQA program, enabling laboratories to regularly monitor their performance.

Methods

Survey material was sourced from an affiliate EQA scheme (WEQAS). Samples are amniotic fluid-based covering clinically relevant levels (30–370 ng/mL). Nine sites (from Australia and New Zealand) were enrolled across three surveys between February and June 2021. Participating laboratories used the Hologic Rapid fFn 10Q or Hologic PeriLynx. Medians were used to set the target values and results were assessed against the Analytical Performance Specification (APS), when applied to the target value. Professional opinion, based on clinical risk, was used to derive the APS.

Results

Up to seven laboratories returned results for all three surveys. The determined APS was ±15 up to 200 ng/mL; ±15% >200 ng/mL. The majority of participants returned results within the APS. Between-lab imprecision (CV%) varied between 3.5–20.6%.

Conclusion

To our knowledge, this is the first Australian EQA program designed for fFn. Although the current participation rate is modest, results demonstrate acceptable inter-laboratory agreement, providing confidence in fFn testing/reporting. As more data becomes available the APS will undergo further review. Hopefully, as fFn testing gains prominence in the diagnostic space, greater participation in the program will follow.

CBR. 2021 Dec;42(Suppl 1):S22.

P38 URINE URIC ACID AND CALCIUM SCREENING TO IDENTIFY SYNTHETIC URINE SAMPLES

R Thurailingam ¹, R Taylor ¹, L Lennox ¹, D Suwandaratne ¹, H Cain ¹, L Nguyen ^1,^✉

Introduction

Ensuring specimen integrity is a key component of medicolegal urine drug testing. Substitution of urine with synthetic urine may be difficult to identify with the laboratory tests currently utilised. Our aim was to investigate whether measuring urine uric acid (UA) and calcium would assist in identifying synthetic urine specimens.

Methods

Commercially available synthetic urine, previously identified synthetic urine and patient urine samples were assayed for UA, calcium, urea and creatinine.

Results

In synthetic urines (N=23), the average UA and calcium concentrations were <0.01 mmol/L and 0.68 mmol/L (SD 0.27), respectively. In patient samples (N=100), the average UA and calcium concentrations were 2.86 mmol/L (SD 1.74) and 3.20 mmol/L (SD 2.61), respectively. Samples with low urea and creatinine were associated with lower UA and calcium concentrations. Using an UA cut-off of <0.1 mmol/L, there was 100% sensitivity and specificity for identifying synthetic urines. Using a calcium cut-off of <1.25 mmol/L, there was 100% sensitivity and 72% specificity for identifying synthetic urines.

Conclusions

Urine UA assay can be readily set up on analysers measuring creatinine. Low UA is a sensitive and specific indicator of synthetic urine. Adding UA as a routine part of integrity testing to samples undergoing urine drug testing will be useful in identifying synthetic urine specimens.

CBR. 2021 Dec;42(Suppl 1):S22.

P39 DIRECT BILIRUBIN: ROCHE V RANDOX – PHOTOTHERAPY INTERFERENCE

C Cocotsi ^1,^✉, CF Lim ², A Connell ¹, J Cleary ¹, M Rozkin ²

Background

The most common method used in Australia to determine the concentration of conjugated bilirubin is a diazo-based dye reaction. This is notoriously sensitive to interference from haemolysis. In our laboratory, we explored an alternate method of conjugated bilirubin measurement, by vanadate-oxidation method (Randox, Crumlin, UK). This method was shown to be significantly more resistant to interference by haemolysis; however, a there was the question of whether it measured isomers of bilirubin as well as the original bilirubin molecule. There have been reports that photoisomers interfere in the vanadate-oxidation method, falsely increasing direct bilirubin measurement. Reporting falsely elevated direct bilirubin could result in invasive investigations in an otherwise only mildly unwell infant.

Aim

To determine whether blue-light phototherapy created positive interference in the measurement of direct bilirubin by the vanadate-oxidation method.

Method

Serum samples with elevated direct bilirubin by both Roche and Randox methods were pooled, and aliquoted into 36 portions. These were then subjected to blue light via the Giraffe Blue Spot PT LiteLite (GE Healthcare, Chicago, US), as used in our neonatal unit. Each aliquot was subjected to a different length of phototherapy, from 0 minutes through to 24 hours. On removal from light, the aliquot was frozen, and thawed for batch analysis.

Discussion

There was no evidence in our samples of significant interference due to blue-light phototherapy. Each sample decreased in response to phototherapy.

Conclusion

The Randox vanadate-oxidation method was implemented in our pathology laboratory with great success and favourable responses from our paediatricians, as 96% of direct bilirubin requests now have a result.

CBR. 2021 Dec;42(Suppl 1):S22–S23.

P40 SIMULTANEOUS ANALYSIS OF 3 BUSULFAN METABOLITES IN PLASMA USING GAS CHROMATOGRAPHY MASS SPECTROMETRY

C R Warnholtz ^1,^✉, R Lawson ¹, BC McWhinney ¹, JPJ Ungerer ¹

Introduction

Busulfan is an alkylating agent used as part of conditioning regimens prior to haematopoietic stem cell transplants. Metabolites of busulfan may contribute to the efficacy and toxicity of the efficacy of the regimen and could play a role in the time-dependent clearance observed for busulfan. We have developed a gas chromatography mass spectrometry (GC/MS) assay that simultaneously measures three busulfan metabolites, tetrahydrothiophene (THT), tetrahydrothiophene 1-oxide (THTO) and sulfolane.

Methods

Plasma (200 μL) was mixed with methanol containing deuterated internal standards (2000 μg/L), saturated sodium chloride and chloroform:isopropanol. The samples were mixed via rotation for 5 min and centrifuged for 10 min. The solvent layer was injected into the GC/MS system and the analytes separated on a HP5-MS column.

Results

Analytes were linear to the range of 5–2000 μg/L. For each analyte, the coefficient of variation (%CV) was <4% (n=20) and <6% (n=13) for intra- and inter-run respectively. Recovery was 86–106% for all analytes and autosampler carryover was insignificant (<0.09%). No interferences or ion suppression issues were observed. Total run time was 15 min between injections.

Conclusion

We have developed a single GC/MS assay that measures 3 metabolites of busulfan in biological samples. It shows excellent precision, recovery and specificity. Sample preparation has been simplified and the analytical run time shortened when compared to other published methods. If busulfan metabolites are found to contribute to toxicity, efficacy or busulfan clearance, this new assay could be used in clinical practice to improve outcomes for patients undergoing HSCT utilising busulfan.

CBR. 2021 Dec;42(Suppl 1):S23.

P41 URINE IS A SUITABLE MATRIX FOR THE MEASUREMENT OF COPEPTIN

S Franke ^1,^✉, J Sherfan ¹, M Lee ¹

Introduction

Copeptin (also known as CT-proAVP) is a 39 C-terminal amino acid peptide of arginine vasopressin (AVP). Copeptin is secreted in equimolar amounts to AVP thus making it an ideal surrogate marker. Here we validate the measurement of copeptin in urine matrices.

Methods

Copeptin was measured on the BRAHMS Thermofisher Kryptor Compact Plus using the CT-ProAVP assay, based on the principle of Time-Resolved Amplified Cryptate Emission (TRACE). Recovery was assessed by serially diluting pooled serum and urine with the kit calibrator. Two samples from the recovery study were used for precision. To determine the suitable collection type and stability, a 24-hour urine was aliquoted into three 50 mL pots containing the following: 1 mL of 5M hydrochloric acid, 0.0032 g of thymol; and no preservative. Spot urines and 24-hour urine pots were then further aliquoted and stored at: room temperature, 4°C, and −20°C, over 7 days.

Results

Copeptin in serum and pooled urine produced mean recoveries of 111% (94–128%) and 109% (97–121%) respectively. The intra-assay imprecision at 6.77 pmol/L and 16.5 pmol/L produced CVs of 4.8% and 2.8% and the inter-assay imprecisions were 9.8% and 6.7% respectively. Creatinine corrected 24-hour urine collections with no preservative, HCl and thymol gave comparable recoveries in all storage conditions over 1 day, however, for long term storage, recoveries were better in urines without preservative and stored at −20°C for a period of 1 week. Acid preservative affected reaction kinetics and was deemed not suitable. Spot urines were found to be stable at room temperature and 4°C up to 7 days. Freezing of spot urines gave a higher mean recovery of 135% (113–158%).

Conclusion

The BRAHMS copeptin assay is suitable for measurement in urine.

CBR. 2021 Dec;42(Suppl 1):S23.

P42 RAPID CREATION OF AN EXTENSIVE PEPTIDE LIBRARY FOR THE SEVERE ACUTE RESPIRATORY SYNDROME CORONAVIRUS 2 (SARS-COV-2)

C Hodgkins ^1,², LK Buckton ^1,^✉, GJ Walker ^3,⁴, B Crossett ⁵, SJ Cordwell ^2,⁵, AR Horvath ¹, WD Rawlinson ^1,^3,^4,⁶

Introduction

Reverse transcriptase-polymerase chain reaction (RT-PCR) has been the cornerstone of diagnosing viral infections in clinical virology. RT-PCR is limited to confirming the presence of viral nucleic acids. Mass spectrometry (MS) can detect and quantify viral proteins. Used together, these techniques could provide complementary information on whether a virus is replication-competent and transmissible. In this project we produced a targeted liquid chromatography tandem mass spectrometry (LC-MS/MS) method suitable for respiratory virus detection in upper respiratory tract samples.

Methods

Signature peptides were empirically determined from analyses conducted using high-resolution MS with low-flow LC systems with a bottom-up proteomics approach to achieve the required sensitivity, specificity, dynamic range, and throughput.

To create a panel of target peptides for the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), we combined analyses from multiple high-resolution MS systems, including untargeted data-dependent-acquisition (DDA) on an Orbitrap MS, semi-targeted DDA using an inclusion list from the Orbitrap results on a quadrupole time-of-flight (QTOF), and untargeted data-independent-acquisition (DIA) on the same QTOF system.

Results

Detecting viral proteins in clinical samples using LC-MS/MS is challenging because these specimens contain a high background of other non-viral proteins. Cultured virus was more suitable for peptide discovery since it contains a higher concentration of viral proteins than clinical specimens.

Combining DDA and DIA on multiple instruments allowed us to build an extensive library of SARS-CoV-2 peptides. Overlapping results between different discovery methods provided higher confidence in identifications, and collection of MS/MS data in DIA provided valuable information for peptide selection without the need for additional targeted analyses.

Conclusions

While detecting viruses in clinical samples is a significant challenge, this workflow allowed for highly specific identification of SARS-CoV-2 with no cross-reactivity for other viruses in the same family. This methodology is rapid and transferrable to many applications including other viruses and human protein biomarkers.

CBR. 2021 Dec;42(Suppl 1):S23–S24.

P43 A CASE OF MACRO-AST IN A PATIENT WITH PROBABLE SERONEGATIVE RHEUMATOID ARTHRITIS

BV Li ^1,^2,^3,^✉, JD Falvey ², L Pike ^1,², C Sies ^1,², PT Chapman ², CM Florkowski ^1,^2,³

Introduction

A 19-year-old female Caucasian patient with probable seronegative rheumatoid arthritis and multiple symptoms, came to our attention following five weeks of consistently raised aspartate aminotransferase (AST) on three separate occasions (260–288 U/L (RI: 0–50)) in the setting of otherwise normal liver function tests undertaken before commencing sulphasalazine.

Methods

A non-invasive liver screen and abdominal ultrasound were completed. In addition, AST recovery post-polyethylene glycol (PEG) precipitation was performed and creatine kinase (CK) was measured on the Abbott ARCHITECT c16000 platform.

Results

Post-PEG recovery of AST was 4% (compared with 48% in a matched control sample), strongly suggestive of macro-AST. Normal CK (89 U/L (RI: 30–180)) suggested no substantial muscular source of elevated AST. Abdominal ultrasound showed normal liver. Iron studies showed iron deficiency rather than overload, with transferrin saturation 7% (RI: 16–45), and ferritin 49 μg/L (RI: 20–200) in context of C-reactive protein (CRP) 11 mg/L (RI: <5). Caeruloplasmin was normal (0.47 g/L (RI: 0.15–0.60)). Alpha-1 antitrypsin had normal MM phenotype and level (2.9 g/L (RI: 1.0–2.0)). Alphafetoprotein was non-elevated (<5 μg/L (RI: 0–16)). Hepatitis B and C serologies negative. Anti-nuclear antibody (ANA) detected at 1:320 in a homogeneous, chromosome positive pattern.

Conclusion

All the findings supported macro-AST in a patient with probable seronegative rheumatoid arthritis, including low post-PEG recovery of AST and substantial exclusion of muscular or liver related causes of AST elevation. Macro-AST is generally considered to be benign. Although pathogenesis is unclear, there have been other case reports of macro-AST in association with rheumatological conditions, hypothesised to be attributable to dysregulation of immune tolerance. Recognition of macro-AST as the cause of elevated AST is important to avoid misattribution of elevated AST and to avoid unnecessary invasive investigations.

CBR. 2021 Dec;42(Suppl 1):S24.

P44 ASSESSING COMMUTABILITY FOR ASSAYS WITH REAGENT LOT-TO-LOT VARIABILITY

M Albeos ^1,^✉, D Chesher ¹, Y Tan ¹

Introduction

Under the ISO 15189 laboratories are required to verify the performance of each new lot or shipment of reagent. Unfortunately, quality control materials are not always commutable and behave differently when assayed compared to patient samples. Undertaking patient comparisons with each new lot of reagents places a significant burden on the laboratory. The aim of this study was to determine for which assays there is an equivalent relationship between the results on patients and quality control materials, which could then allow acceptance testing to be limited to quality control materials alone.

Methods

A patient comparison was done by running 7–20 samples with widely distributed results and QC materials, on both old and new reagent lots. A relative deviation (RD) for QC samples was calculated as the absolute value of the average relative difference of patient results minus the average relative difference of QC results divided by the standard deviation of the differences for the patient results.

An assay was considered commutable if:

The RD at each QC level was <1.96; and
The above criteria was met for three consecutive lots.
If a commutability assessment failed the first criteria, then another 5 successive lot evaluations would need to pass the assessment.

Results

102 assays were evaluated and 61 were deemed as commutable. For assays that were not commutable, patient comparison was still required.

Conclusion

Our results confirm that manufactured quality control materials behave differently in many assays to patient samples and are not always suitable for assessing new lots. For non-commutable assays, patient correlation is recommended in addition to successful quality control check before the performance of an assay can be accepted with confidence. By identifying the commutable assays, our laboratory is able to limit the number of patient comparison studies required, which translates to savings in reagent usage, and staff time.

CBR. 2021 Dec;42(Suppl 1):S24.

P46 EVALUATION OF A POINT-OF-CARE BILIRUBIN ASSAY FOR MONITORING OF JAUNDICE IN THE NEONATAL INTENSIVE CARE UNIT AT PERTH CHILDREN’S HOSPITAL

S Trivic ^1,^✉, A Wilson ², E Byrnes ¹

Introduction

The Neonatal Intensive Care Unit (NICU) at Perth Children’s Hospital (PCH) requested the use of their existing Radiometer ABL90 Flex Plus blood gas analyser as an alternative to the Transcutaneous Bilirubinometer (TCB) as a point-of-care testing (POCT) device to screen neonates with jaundice. Neonates having routine blood gases (45μL) performed could have a bilirubin (tBil) result added to the profile without an increase in blood collection volume.

Methods

tBil results from 62 neonatal blood gas samples analysed using the Radiometer ABL90 Flex Plus analyser in the NICU at PCH were compared to samples collected at the same time and sent to the laboratory for plasma bilirubin measurement using the Ortho-Clinical Diagnostics Vitros 4600 neonatal bilirubin assay.

Results

There was good correlation of results between the POCT whole blood tBil assay on the ABL90 and the plasma bilirubin assay on the Vitros 4600. (Passing and Bablok slope = 1.013, R² = 0.9524, intercept = −8.7).

Conclusion

The Radiometer ABL90 Flex Plus blood gas analyser has the ability to provide an accurate and reliable bilirubin result at point of care with a quick turnaround without an increase in the volume of blood required (45μL). Of significant benefit is the negative predictive value eliminating the need for a formal blood collection to be sent to the laboratory except in cases where POCT tBil results are close to treatment decision markers. In these cases, a confirmatory result by way of blood sent to the laboratory for analysis using the Vitros 4600 may still be required.

CBR. 2021 Dec;42(Suppl 1):S24–S25.

P47 SERUM ASCITES ALBUMIN GRADIENT – COMPARISON OF TWO BCP AND BCG ALBUMIN METHODS

S Singh ^1,^✉, Y Tan ², S Thompson ¹, AR Horvath ¹

Introduction

The serum ascites albumin gradient (SAAG) is used to differentiate between portal hypertension-related and non-related ascites. The commonly used cut-off of 11 g/L was derived using the bromocresol green (BCG) albumin method. It has been suggested that this cut-off is no longer appropriate for the bromocresol purple (BCP) method which generally measures lower than the BCG method. Previous work indicated that the Abbott BCP assay provides a lower SAAG than the BCG method. The aim of this study was to determine whether a similar pattern is seen with the Roche albumin assay.

Methods

Albumin was measured on paired serum and ascitic fluid samples, on the Roche Cobas c701 (BCP)/Roche Cobas Integra (BCG) and Abbott Architect c16000 (BCG/BCP) analysers. SAAG was calculated for each method on each platform, and classified according to the cut-off of 11 g/L.

Results

16 paired adult specimens were included. Mean differences between BCP and BCG albumin were −5.8 g/L (Roche) and −2.4 g/L (Abbott) in serum; and −1.8 g/L (Roche) and −0.1 g/L (Abbott) in ascitic fluid. Consequently, both the Roche and Abbott BCP assays measured a lower SAAG than their BCG assays: mean difference −4.0 g/L (Roche) and −2.3 g/L (Abbott). The mean SAAG by the BCG assay was 19.8 g/L (Roche) and 17.8 g/L (Abbott), and it was nearly the same for the two BCP assays: 15.8 g/L (Roche) and 16.2 g/L (Abbott).

Using the cut-off of 11 g/L, 3/16 patients (Roche) and 2/16 patients (Abbott) would have had their SAAG classified differently dependent upon the albumin method used.

Conclusions

Our results confirm previous findings and indicate that the SAAG is lower with BCP albumin assays than with the BCG methods, irrespective of manufacturers. Consequently, the SAAG decision limit of 11 g/L warrants re-evaluation with the widespread use of BCP albumin assays. Clinical studies are required for establishing the new cut-offs.

CBR. 2021 Dec;42(Suppl 1):S25.

P48 ALDOSTERONE AND ADRENAL VEIN SAMPLING (AVS): EFFECT OF IOHEXOL AND IMPLEMENTATION OF AUTOMATED EXTRACTION FOR LIQUID CHROMATOGRAPHY MASS SPECTROMETRY QUANTIFICATION

BR Cooke ¹, MM Page ¹, K Vigors ^1,^✉, M Taranto ¹

Introduction

Adrenal vein sampling (AVS) is used to determine whether adrenal hypersecretion of aldosterone in patients with primary aldosteronism is unilateral or bilateral. Samples for the measurement of aldosterone and cortisol are taken by an interventional radiologist, using iodinated contrast media (iohexol; Omnipaque^TM in our centre) to facilitate catheter placement. Some apparently anomalous AVS results led us to investigate whether iohexol interfered with aldosterone and cortisol measurement by the routinely-used chemiluminescent immunoassay methods.

Methods

Serial concentrations of iohexol were added to a pool of EDTA whole blood and centrifuged. Cortisol and aldosterone in plasma were measured by immunoassay (Abbott Architect and Diasorin Liaison, respectively). Prior to LC-MS/MS analysis, plasma was extracted using MTBE and a Phenomenex Novum 96 well plate on an automated Eppendorf EpMotion 5075 platform. Extracts were chromatographically separated on a BEH C18 column and quantiified by mass spectrometry on a Waters Xevo® TQ-S.

Results

As the concentration of iohexol increased, the aldosterone concentration was expected to decrease due to dilutional effects only. Quantification of aldosterone via LC-MS/MS followed the predicted concentration pattern. The results from the Liaison method however were inversely proportional, implying positive interference of iohexol in the immunoassay, but not the LC-MS/MS method. Additionally, both methods for cortisol demonstrated negligible iohexol interference.

Conclusion

LC-MS/MS is superior for the analysis of AVS samples as the iohexol contrast used in sample collection causes interference in aldosterone quantification using the Liaison method. As iohexol contamination in AVS samples is unpredictable and variable between left and right adrenal samples, falsely increased aldosterone concentrations could lead to poor patient outcome. False non-lateralisation would deny a patient with unilateral disease the chance of a surgical cure. False lateralisation would lead patients with bilateral disease to pointless, potentially harmful surgery.

CBR. 2021 Dec;42(Suppl 1):S25.

P49 LC-QTOF-MS ANALYSIS OF CLANDESTINE SELECTIVE ANDROGEN RECEPTOR MODULATOR (SARM) CAPSULE

D Allen ^1,^✉, B Walker ¹, B Espley ¹, C Warnholtz ¹, B McWhinney ¹, E De Waal ¹

Introduction

Selective androgen receptor modulators (SARMs) are a class of non-steroidal androgen receptor ligands known to exert effects such as increased muscle growth. Primarily introduced as novel therapeutic agents, their popularity has subsequently increased within sporting activities such as body building due to their performance-enhancing effects. Despite the absence of appropriate approval for human use, their popularity increases due to their regular marketing as safe alternatives to anabolic steroids, and largely unregulated availability via the Internet. Recent studies have indicated that the true chemical composition of clandestine SARMs is often undisclosed with many online products containing additional unapproved substances. A male patient presenting to an emergency department with suspected glomerulonephritis declared using Testolone (RAD-140), a SARM used to increase muscle mass. A sample of the capsule was subsequently analysed by LC-QTOF-MS to determine the chemical composition.

Methods

Capsule contents were extracted and analysed via LC-QTOF-MS using untargeted data acquisition. Positive identification of chemical components was achieved by comparison with accurate mass, fragmentation and retention time data obtained from the LC-QTOF-MS scientific library. Presumptive identification of additional components was achieved using accurate mass, isotopic pattern and in-silico fragmentation.

Results

Analysis indicated that the capsule did not contain Testolone (RAD-140) as indicated by the product ingredients. The LC-QTOF-MS scientific library identified the presence of Ostarine (SARM) and caffeine. Review of untargeted data also provided a presumptive positive for the growth hormone secretagogue Ibutamoren, and the body building supplement creatine.

Conclusion

The unauthorised use of selective androgen receptor modulators (SARMs) is gaining popularity due to their performance-enhancing effects. With the absence of appropriate approval for human use and unknown chemical compositions, SARMs may present a risk to human health. This study demonstrates the inaccurately advertised ingredients of clandestine SARMs and highlights their relevance as emerging analytical targets in clinical toxicology.

CBR. 2021 Dec;42(Suppl 1):S25–S26.

P50 INTRODUCTION OF A POINT-OF-CARE KETONES EQA

S Shepherd ^1,^✉, R de Leon ¹, P Graham ¹

Introduction

Point-of-care ketone meters are in routine use for detecting and monitoring diabetic ketoacidosis (DKA). The RCPAQAP introduced a point-of-care (PoC) EQA for ketone meters in 2019 following requests from customers in hospital settings. The material is customised for specific meters and initially offered for StatStrip^® and Optium Neo^® meters, but to date, the main interest has been from StatStrip^® users. We reviewed the performance of StatStrip^® meters enrolled in our 2020/21 program to gauge the current average performance of PoC operators in hospital settings.

Methods

Results from six bi-monthly (two sample) surveys from June 2020 to April 2021 were analysed using RCPAQAP inhouse software. Participant/meter performance is reported using a “traffic light” assessment where green is within 1 Analytical Performance Specification (APS), orange between 1 and 2 APS and ed > 2 APS, where the APS is defined as +/−0.5 mmol/L up to 5 mmol/L then +/− 10%. The samples are aqueous liquid samples supplemented with ketone and provided in six linearly related concentrations. The medians of the samples range from 1.2 mmol/L to 6.4 mmol/L which covers the relevant medical decision limit (>3 mmol/L).

Results

Eight sites with a total of 1148 meters participated in the program. While participation varied over the course of the cycle, sites averaged good (green) performance for 75% of their results, acceptable (orange) 15% and poor (red) 10% indicating overall clinically acceptable performance for the detection and monitoring of DKA.

Conclusion

We have successfully established a PoC EQA for ketone meters that fulfills the need in hospital settings. The data from the RCPAQAP Ketones program demonstrates that well-trained competent PoC operators achieve satisfactory performance in keeping with good clinical practice.

CBR. 2021 Dec;42(Suppl 1):S26.

P51 URINARY MULTI-ELEMENT ICP-MS METHOD UTILISING SYNTHETIC MATRIX AND L-CYSTEINE STABILISATION OF MERCURY

D Allen ^1,^✉, E Connolly ¹, R Swenson ¹, B McWhinney ¹, C Pretorius ¹

Introduction

Biomonitoring in urine by inductively coupled plasma mass spectrometry (ICP-MS) allows the simultaneous measurement of numerous nutritional and toxic elements. Challenges encountered in the measurement of trace elements in urine include large biological variability in both trace element levels and sample matrix composition. The measurement of mercury in urine also presents analytical challenges due to memory effects in sample introduction systems. An ICP-MS method was developed for the measurement of zinc, copper, iodine, arsenic, cadmium, chromium, lead and mercury. A sodium chloride synthetic matrix was utilised to provide equivalent matrix matching of analytical standards. L-cysteine was also used as a stabilising agent due to the binding affinity of mercury to thiol functional groups.

Methods

Patient urine (100 μL), 2% L-cysteine (100 μL) and water 18 MΩ (100 μL) were diluted with 1800 μL of aqueous diluent containing ammonia, EDTA, 2-propanol, Triton X-100 and internal standards. For calibration, aqueous analytical standards (100 μL), 2% L-cysteine (100 μL) and 50 mM NaCl synthetic matrix (100 μL) were also diluted with 1800 μL of sample diluent. Analysis was performed using an Agilent 7900 ICP-MS. Calibration was performed using standard additions with blank background subtraction.

Results

Sufficient matrix matching of analytical standards was achieved using synthetic matrix providing low trace element background levels. The addition of L-cysteine to the sample preparation resulted in increased recovery of mercury due to reduced absorption of mercury to preparation tubes and instrument components. Inter- and intra-run precision was <10% for all elements. Carry-over for all elements was <0.05 % except for mercury at 0.46%. All elements showed excellent correlation with an existing ICP-MS methodology.

Conclusion

This method provides a sensitive, accurate and robust method for the measurement of 8 elements in human urine, successfully utilising synthetic matrix and effective stabilisation of mercury with L-cysteine.

CBR. 2021 Dec;42(Suppl 1):S26.

P52 THE EFFECT OF HAEMOLYSIS ON LACTATE DEHYDROGENASE ISOENZYME MEASUREMENT

M Lee ^1,^✉, J Sherfan ¹, N Perera ¹

Introduction

Lactate dehydrogenase (LDH) enzyme is an oxidoreductase enzyme of the anaerobic metabolic pathway, with 5 main isoenzymes. It is found in all tissues including red blood cells, muscles, liver and the heart. LDH isoenzyme measurement is performed using the QuickGel Kit (Helena Laboratories) and SPIFE 3000 electrophoresis system. Package insert from the manufacturer states haemolysed samples should not be used but provided no specific haemolysis cut-off level for rejection. We aim to study the effect of haemolysed samples on LDH isoenzymes to guide our pre-analytical sample handling procedure.

Methods

Packed red cells from 5 random donors were collected from blood bank and haemolysed by aspiration through a 25G needle, followed by vortexing. Different concentrations of haemolysate were added to 7 aliquots of pooled sera to give approximate H-indices of 5, 10, 15, 30, 50, 60 and 70 μmol/L. Total LDH and H-index (performed on the Roche Cobas c702), as well as LDH isoenzyme electrophoresis in duplicate were performed on these aliquots and unspiked sera.

Results

Total LDH of the initial pooled sera ranged between 98–163 U/L. At H-index of 70 compared to unspiked sera, total LDH increased by an average of 118U/L (range 109–127), while the percentage proportion of each LDH isoenzyme changed by an average of 4.3% (range 2.1 to 7.6) for LDH1, 1.5% (range −0.1 to 2.5) for LDH2, 1.5% (range −0.5 to 1.9) for LDH3, −2.5% (range −1.4 to −3.9) for LDH4, and −4.2% (range −2.4 to −6.0) for LDH5.

Conclusions

Haemolysis up to a H-index of 70 μmol/L increases the proportions of LDH isoenzymes 1,2 and 3, and decreases the proportions of LDH isoenzymes 4 and 5.

CBR. 2021 Dec;42(Suppl 1):S26.

P53 PLEURAL FLUID FATTY ACID SYNTHASE AS A NOVEL PROGNOSTIC MARKER OF MALIGNANT PLEURAL EFFUSION

XM Hu ^1,^✉, CY Law ², CW Lam ¹

Introduction

Overexpression of fatty acid synthase (FASN) is associated with tumour growth in many cancers. Our previous work has found that free fatty acids (FFAs) profiles differentiated cancer from noncancerous pleural effusion. This study aimed to test whether FASN in the malignant pleural effusion (MPE), the enzyme for fatty acid synthesis in cancer cells, is related to overall patient survival.

Methods

In this study, 66 patients with MPE were included for pleural fluid FFAs and FASN analysis. The quantification of pleural FFAs was performed using a free fatty acid colorimetric assay kit, and the pleural FASN concentrations were measured by the human FASN ELISA kit.

Results

Kaplan-Meier survival analysis demonstrated a significantly longer survival time in the low FASN group than the high FASN group. The hazard ratio (HR) was determined to be 3.28 (95% CI: 1.65–6.55) with a p-value = 0.0007 using a cut-off at 3.678 ng/mL.

Conclusion

The fatty acid synthase (FASN) is related to the overall patient survival of MPE patients. The high FASN in MPE is associated with shorter overall patient survival. The result is consistent with the overexpression of FASN in lung cancer cells. Taken together, pleural fluid FASN or FFAs profiles of MPE can serve as predictive biomarkers for MPE patients.

CBR. 2021 Dec;42(Suppl 1):S27.

P54 HELICOBACTER PYLORI SAMPLE SWAP AND STABILITY

K Tran ^1,^✉, MJ Tanner ¹, P Rayamajhi ¹, L Hainke ¹, S Lewis ¹, K Reid ¹, K Oh ¹, N Varis ¹

Introduction

There is currently no external quality assurance (QA) program available for the Helicobacter pylori breath test (HPBT) in Australia or internationally. The WA AACB QC Subcommittee established a patient sample exchange program to evaluate HPBT performance across WA laboratories. The stability of samples was assessed during the exchange and while in storage for 10 months.

Method

Between 2018 and 2021 sixteen breath test samples in scintillation fluid were selected including results that were classified as negative (<50 disintegrations per minute, DPM), indeterminate (50–199 DPM), and positive (>200 DPM) for Helicobacter pylori. The samples were from paediatric and adult populations. Samples were de-identified, light protected, and sent across four WA laboratories, with the initial laboratory re-testing at the end. All samples were collected utilising the TriMed PyTest collection kits and the ¹⁴C activity was determined using scintillation counters. Six samples were stored at room temperature and retested over 10 months.

Results

The CVs for positive samples in the sample exchange ranged between 2.3–6.2% (n=6), indeterminate between 3.8–11.2% (n=4), and negative between 21.8–114.9% (n=6). The intra-laboratory CV for positive samples ranged between 0.86–7.86%, for indeterminate 1.6–10.1%, and for negative 0–94.2%. The CVs of positive, indeterminate and negative samples stored for ten months were 1.8%, 2.6% and 39%, respectively, and for one month 1.9%, 3.4% and 18.7%, respectively.

Conclusion

This study showed HPBT samples in scintillation fluid are suitable for inter-laboratory comparison, with stability for up to 10 months when kept in routine storage conditions. These results suggest patient samples could possibly be used as a QA material across Australia. WA laboratories seem to show concordant results, and plan to continue comparison bi-annually.

CBR. 2021 Dec;42(Suppl 1):S27.

P55 HYPERNATRAEMIA OR EVAPORATION? A METHOD FOR INVESTIGATION OF SUSPICIOUSLY HIGH SODIUM RESULTS

MB Theophilos ^1,^✉, MP Nascimento ¹, E Sargeant ¹, J Waldron ¹, D Angeleski ¹, C Lynch ¹, KA Sikaris ¹

Introduction

Pre-analytical serum sample evaporation due to extended uncapping times can result in spuriously high serum sodium measurements due to sample concentration, with the consequential risk of erroneous clinical management if this error is not detected. In our laboratory EDTA tubes remain capped during sample processing and thus do not exhibit evaporative concentration. We have internally validated a method for measuring sodium from EDTA tubes to correlate with suspected evaporated serum tube samples

Methods

A total of 187 serum samples for inclusion in the training set were selected on the basis of analysis within 6 hours of arrival in the laboratory, and selected to give a range of serum sodium results between 119–158 mmol/L. Corresponding EDTA tubes were retrieved for analysis the day after serum measurement. Samples were analysed using the Roche Cobas ISE module

Results

There was good correlation of serum sodium and EDTA sodium values (R²=0.94) with a linear regression formula of y=0.87x + 16 in the training set. This formula was then used in a validation set which showed a predicted EDTA sodium level which correlated well with a measured EDTA sodium result. Samples were suspected to be evaporated when the predicted EDTA sodium result diverged from the measured EDTA sodium result with an absolute deviation of more than 4 mmol/L.

Discussion

The development of a formula to apply to sodium results where evaporation is suspected has enabled us to produce a worksheet which our validation scientists can use to compare serum sodium with EDTA sodium results. Validation staff entering these results into a worksheet are able to follow an algorithm which enables them to determine the best course of action for releasing these results in a straightforward and timely manner.

CBR. 2021 Dec;42(Suppl 1):S27.

P56 MASS SPECTROMETRY-BASED MACHINE LEARNING ALGORITHM FOR THE PREDICTION OF MALIGNANT AND TUBERCULOUS EFFUSIONS

Wai-yuen Yu ^1,^✉, Chun-yiu Law ¹, Xiaomin Hu ², Ching-wan Lam ^1,²

Introduction

We confirm the feasibility of a machine learning (ML)-based approach for the classification of pleural effusion (PE) using nuclear magnetic resonance (NMR) data. Here, we further substantiate our ML analysis using mass spectrometry (MS) data from malignant and tuberculous effusions.

Methods

We use the Synthetic Minority Over-sampling TEchnique (SMOTE) k-nearest neighbour (KNN) classifier to PE from a set of 50 liquid chromatography-mass spectrometry (LC-MS) data. 60% (30 cases) of the available dataset were used for the training set and 40% (20 cases) of the data were used for the validation set to test for the performance. The overall predictive performance in terms of classification accuracy and area under the receiver operating characteristic curve (AUC) was determined.

Results

The class ratio of malignant to tuberculous PE was 33:17. Total 3,028 non-preselected LC-MS features were identified and used to generate a ML-based algorithm. The optimal k-value in this KNN model is 7 which achieves a respective accuracy and AUC of 0.80 and 0.80 using the validation set. Given the imbalanced dataset, SMOTE was used for data augmentation in the minority class, the tuberculous PE. The optimal k-value was 9 in this SMOTE-KNN model which achieves a respective accuracy and AUC of 0.93 and 0.99 using the validation set.

Conclusions

Our ML-based method could classify malignant and tuberculous effusions with good confidence. Nevertheless, further evaluation with the use of test set will be necessary to test the overall performance. Chemical Pathology is a special clinical discipline which handles a large amount of laboratory raw data (big data). We envisage the power of ML in the era of big data and its potential clinical use in future laboratory medicine.

CBR. 2021 Dec;42(Suppl 1):S27–S28.

P57 MIDAZOLAM-FENTANYL SEDATION HAS RAPID AND SIGNIFICANT IMPACT ON ADRENAL CORTISOL AND ALDOSTERONE SECRETION DURING ADRENAL VEIN SAMPLING

JCG Doery ^1,^2,^✉, KK Lau ³, W Chong ³, PJ Fuller ^4,⁵, R Libianto ^4,⁵, NYN Chee ⁴, J Shen ^4,⁵, KL Wan ¹, ZX Lu ^1,², J Yang ^4,⁵

Introduction

Simultaneous bilateral adrenal vein sampling (AVS) is the gold standard for distinguishing unilateral and bilateral disease in primary hyperaldosteronism. Cortisol is measured concurrently with aldosterone to confirm successful cannulation and correct the aldosterone measurement for any dilutional effect related to proximity of the catheter tip to the orifice of the adrenal vein. Comparison of the aldosterone/cortisol ratios are then used to determine lateralisation of aldosterone secretion. The recent advent of a rapid semi-quantitative point-of-care (POC) cortisol kit has been recommended to inform likelihood of cannulation success whereby cortisol <276, 276–828 and >828 nmol/L indicating unsuccessful, equivocal and successful cannulations respectively. We report a sentinel case where midazolam-fentanyl sedation was given following the commencement of AVS.

Patient

A 47yr old female with confirmed primary hyperaldosteronism was referred for AVS. One minute after initial sampling 2 mg midazolam and 25 μg fentanyl was administered.

Results and Discussion

Within 10 minutes of drug administration adrenal vein cortisol fell from 4,700 to 570 nmol/L on the right and from 5,200 to 620 nmol/L on the left. Aldosterone/cortisol ratio increased 9.5-fold (0.2 to 1.9) and 3.3-fold (2.2 to 7.6) on the right and left respectively. A retrospective review of data from 34 other patients who underwent AVS with IV sedation (prior to Synacthen stimulation) showed significantly lower cortisol concentrations in the peripheral vein (median of 141 nmol/L vs 218 nmol/L), left adrenal vein (522 nmol/L vs 1284 nmol/L) and right adrenal vein (577 nmol/L vs 1578 nmol/L) (p <0.001 for all) compared to patients with no sedation (n=70).

Conclusions

Sedation during unstimulated AVS can reduce cortisol concentrations to levels deemed unsuccessful for cannulation. This may be misleading when POC cortisol kit is used to indicate whether cannulation is successful. The necessity for IV sedation during AVS should be carefully assessed.

CBR. 2021 Dec;42(Suppl 1):S28.

P58 PROTOCOL FOR DETECTING NON-REPEATABLE FALSE POSITIVE HCG ‘FLYERS’ ON BECKMAN COULTER ASSAY

S Mukerji ^1,^✉, P Bonnitcha ^1,², R Prasad ¹, A Screnci ¹, GRD Jones ¹

Introduction

Non-repeatable false positive serum βhCG results (‘flyers’) can result in treatment delays and inappropriate investigations for patients. For Beckman Coulter’s new Access Total βhCG (5^th IS) assay, improved rates of non-repeatable false positive results (flyer frequency 0.1%) were reported, compared to their previous assay (flyer frequency 0.6%). Following the identification of several hCG flyers, SydPath introduced a targeted re-run protocol for hCG.

Methods

Our re-run protocol involves re-centrifugation and re-analysis of samples with initial hCG results between 5 and 200 IU/L. We audited the frequency of flyers and examined whether elevated leucocyte count, CRP or tube type were associated. We reviewed results over a 6-month period, January 1^st to June 30^th 2021. We also reviewed 4 flyers identified in 2020, prior to introducing our re-run protocol. Samples were measured on Beckman Coulter Access Total βhCG (5^th IS) assay on DxI-800 analyser.

Results

Our re-run protocol identified 9 flyers out of a total 5932 hCG requests (0.15%), with a rate of 3.2% of the 279 samples re-tested. Initial results for flyers ranged between 5 and 69 IU/L. The result post re-centrifugation protocol was <2 IU/L for all flyers. When leucocytes and CRP were co-ordered on the same collection, leucocytes were elevated in 58% (7/12) and CRP was elevated (≥5 mg/L) in 89% (8/9). Number of flyers was similar across tube types (6 heparin plasma and 7 serum).

Conclusions

The overall flyer frequency is similar to the Beckman Coulter claims. While a consistent cause for hCG flyers was not identified, elevated leucocytes, which have been reported to be associated with flyers in a previous generation of this assay, and elevated CRP were seen in over half the flyer samples where these tests were performed. Based on our observed flyer rate and number of re-runs required, we will continue with this protocol.

CBR. 2021 Dec;42(Suppl 1):S28.

P59 DUODENAL TISSUE HOMOGENISATION FOR DISACCHARIDASE ANALYSIS: ULTRASONIC PROBE VERSUS BEAD BEATER

G Pool ^1,^✉, M Baxter ¹, S Wilschefski ¹, R Flatman ¹, G Daley ¹, D Kanowski ¹, L Price ¹, G Ward ¹

Introduction

Diagnosis of disaccharidase deficiency requires the measurement of disaccharidase activity of intestinal biopsies. Tissue homogenisation is required for sample preparation and can be achieved through various mechanical techniques. The aim of this study was to compare two homogenisation techniques, ultrasonic probe versus a bead beater.

Methods

Two extra duodenal biopsy tissues (5–10 mg) from patient samples (n=33) were used. The first piece of tissue of each patient sample was homogenised in a false-bottom tube containing 700 μL homogenising buffer with an ultasonic probe (SONICS VibraCell Ultrasonicator, 45% amplitude, 10 sec). After sonication the homogenate was centrifuged. The second piece of tissue was placed in a Precellys Lysing kit tube (containing ceramic beads) – CK28 for hard tissue homogenisation, homogenising buffer (500 μL) was added to approximately 5 mg tissue and then placed in a Precellys Evolution bead beater (Bertin technologies, 8400 rpm for 30 sec at room temperature). After homogenisation, homogenates were transferred to false-bottom tubes. All samples were analysed on Thermofisher Indiko PLUS analyser for protein concentrations (g/L) lactase, maltase and sucrase activities with in-house methods.

Results

Sucrase and maltase (U/g protein) activity showed no notable differences −0.33% and −5.02% respectively between ultrasonic and bead beating methods. Lactase results obtained from bead beater homogenisation (room temperature) were 22% lower compared to tissue homogenised (on ice) with ultrasonic probe.

Conclusion

In this experiment a cooling system was not provided with the Precellys Evolution - therefore the reduction in lactase activity could be due to heat generation during bead beating. Experimental data suggests that lactase is more sensitive to heat than sucrose and maltase and it is therefore crucial to protect heat sensitive samples with a compatible cooling unit for the bead beater. Results suggest that both ultra-sonication and bead beaters are successful in breaking down duodenal tissue for disaccharidase analysis.

CBR. 2021 Dec;42(Suppl 1):S28–S29.

P60 EXTERNAL QUALITY ASSURANCE PROGRAM FOR SMALL INTESTINE DISACCHARIDASES

J Bush ^1,^✉, P Graham ¹

Introduction

The RCPAQAP sought to develop an External Quality Assurance (EQA) for disaccharidase assays (primarily lactase, sucrase and maltase). These assays are performed to evaluate the enzyme activity of the small intestine to assist in the diagnosis of patients with enzyme deficiencies which can contribute to chronic bowel conditions. In practice, the sample tested is a piece of bowel lumen biopsied during colonoscopy.

Method

Enzyme supplements containing lactase, sucrase and maltase, sourced from a commercial supplier (Sigma-Aldrich) were used to spike a human albumin base solution. Low and high concentration pools were then blended to obtain six linearly related samples. The samples were aliquoted and freeze-dried to maintain stability.

Results

Pre-testing results showed acceptable linearity over six levels with enzyme activity at suitable clinical levels for maltase (4.8–313 U/g protein) and sucrase (6.2–354.9 U/g protein). Lactase was re-worked as the initial levels were too low (0–44.2 U/g protein).

Conclusion

The RCPAQAP Disaccharidase program was launched in February 2021 and offers the first of its type. Further review of the submitted results will help inform laboratories performing these assays on the degree of harmonisation of results.

CBR. 2021 Dec;42(Suppl 1):S29.

P61 METHOD EVALUATION OF THE VITROS B•R•A•H•M•S PROCALCITONIN ASSAY ON THE ORTHO VITROS 5600

C Cross ^1,^✉, JZY Chung ^1,², NJ Perera ^1,²

Introduction

Procalcitonin (PCT) is a biomarker of bacterial infection and sepsis, a potentially lifethreatening condition of which early recognition and treatment is important. Local pathways for the evaluation of febrile infants include a PCT cut-off as a trigger for a full septic screen. Hence it is imperative for laboratory assays to be traceable to published decision limits around the key clinical decision range of 0.5–1.0 μg/L.

Aim

To evaluate the analytical performance of the B•R•A•H•M•S PCT method on the Ortho Vitros Fusion 5600 platform.

Method

Method bias was evaluated by correlation with the Diazyme PCT and Abbott Architect PCT assays utilising patient samples (n=23). Samples from EQA programs (RCPAQAP, RIQAS) were also analysed.

Intra-day and inter-day precision was evaluated by analysis of three levels of VITROS B•R•A•H•M•S PCT manufacturer controls over five days, 20 times in total. In addition, three levels of third-party QC material (BioRad) were analysed over 8 days.

Sample stability was evaluated in patient samples () stored at 4°C for up to 7 days.

Results

Method comparison demonstrated good correlation with the Abbott assay (R²=0.98) and Diazyme assay (R²=0.94), however results from the Diazyme assay were significantly higher at the clinically relevant concentrations of 0.5–1.0 μg/L. All results of EQA sample analysis were within analytical performance specifications. Inter-day CV was determined to be 4.7% at 0.5 μg/L, 3.1% at 1.8 μg/L and 3.8% at 14.4 μg/L. Stability of lithium heparin plasma samples was confirmed to 7 days at 4°C.

Conclusion

The VITROS B•R•A•H•M•S PCT assay was found to be fit for purpose. The assay better enables result comparability with published decision limits in rule-out algorithms for the evaluation of patients with potential sepsis.

CBR. 2021 Dec;42(Suppl 1):S29.

P62 VERIFYING ROCHE FREE T3 REFERENCE LIMITS IN CHILDREN USING LABORATORY DATA

MP Nascimento ^1,^✉, MB Theophilos ¹, C Lynch ¹, KA Sikaris ¹

Introduction

Thyroid hormone levels change across childhood, and partitioning across childhood, as well as the partition to adult levels, should ideally be based on physiological evidence rather than arbitrary definitions (e.g. age 18 y). Following clinician complaints regarding the unnecessary flagging of elevated free T3 (fT3) in puberty, we aimed to verify fT3 reference intervals in childhood using an indirect technique.

Methods

We extracted the results of 62,949 thyroid function tests performed on Roche Cobas between 2011 and 2020. We looked at the trend in the median distribution with age to identify physiological partitions. We then applied Bhattacharya analysis to determine the reference intervals for each transition and compared these with the manufacturer’s recommendations.

Results

Our verification supported the published partitions up to age 11 years for both genders, however they could not be confirmed beyond that. We found that the adolescent transition was completed at age 16 in girls and age 18 in boys.

Discussion

We extended the fT3 upper reference limit to 7.7 pmol/L for these new gender-related age limits in adolescence. The new limits reduced reporting of false positives (i.e. isolated elevations of fT3). Further investigation is required to elucidate the reason for the noted differences in transition between genders, which could be related to physiological changes, or alternatively due to assay characteristics. Studies using alternative platforms would be required to determine if these findings are reproducible on other fT3 methods.

CBR. 2021 Dec;42(Suppl 1):S29.

P63 OBSERVED VS DESIRED EQA PERFORMANCE – Z-SCORES VS APS

A Taif ^1,^✉, P Graham ¹, T Badrick ¹

Introduction

External Quality Assurance (EQA) providers commonly report z-scores to guide on assessment of results. Analytical Performance Specifications (APS), based on biological variation are another means of assessment. RCPAQAP introduced z-scores in addition to APS for quantitative survey reports in 2019. z-scores are calculated using the mean and standard deviation of the submitted (observed) results and vary from survey to survey. APSs are consistent from survey to survey, regardless of the scatter of results; they can be set to drive the desired performance of a measurand. We sought to evaluate if a combination of z-scores and APSs provides the same or enhanced information on method performance.

Methods

z-scores and APSs was reviewed for the 2020 and 2021 RCPAQAP General Serum Chemistry, Special Lipids, Endocrine and Liquid Chemistry programs. Trends (either where both “tracked” each other or diverged) were noted and further investigated.

Results

Relative differences were minimal (<1) between z-scores and APSs for measurands that are traditionally performed well (both accurate and precise compared to a fixed reference target) (e.g. sodium, potassium, glucose) across all laboratories. Measurands with a wide scatter of results (e.g. Ca-125) where a reference (fixed) target was applied (e.g. vitamin D3) showed acceptable z-scores, and APSs requiring review. Where an APS reflected clinical use (e.g. urine calcium APS +/−2.0 mmol/L), and labs were precise, a small deviation flags for z-score review, but not for APS.

Conclusion

The combination of z-scores and APS for EQA performance assessment provides a balanced approach. z-scores allow for differences in commutability of the samples, APSs drive and maintain the desired performance of a method, particularly when the samples are commutable and fixed targets-based reference methods are applied.

CBR. 2021 Dec;42(Suppl 1):S29–S30.

P64 ANALYSIS OF CITRUS BIOFLAVONOID CONTENT AND DIPEPTIDYL PEPTIDASE-4 INHIBITORY POTENTIAL OF COMMERCIALLY AVAILABLE SUPPLEMENTS FROM USA, AUSTRALIA AND CANADA

A Gupta ^1,^✉, CK Narkowicz ¹, HA Al-Aubaidy ^1,², HF Jelinek ^3,⁴, DS Nichols ⁵, JR Burgess ^1,⁶, GA Jacobson ¹

The study involved the analysis of 10 commercially-available citrus bioflavonoid supplements from three different countries: Australia, USA and Canada. The supplements were tested for their citrus bioflavonoid content which varied from 0.8–33.3 %w/w. CBF-3 showing highest %w/w and CBF-6 lowest %w/w. The daily bioflavonoid dose varied from 19 mg to 560 mg. The highest flavonoid dosage was by CBF-4 and lowest by CBF-6. Hesperidin was the major citrus bioflavonoid in nine out of ten supplements. One supplement was found to contain <10% of the quantity of rutin claimed to have been added. The DPP-4 inhibitory potential, compared through an estimation of rutin equivalence, ranged from 1.9–400 mg per day. The maximum DPP-4 inhibitory activity with respect to rutin equivalence was shown by CBF-7 and minimum activity by CBF-6.

This data highlights the variability between the supplements in their potential to inhibit DPP-4 for subsequent health benefits.

CBR. 2021 Dec;42(Suppl 1):S30.

P65 EFFECT OF URINE DIPSTICK IMMERSION ON TRACE METAL ANALYSIS BY ICP-MS

S Keys ^1,^✉, S Wilschefski ¹, M Baxter ¹, G Pool ¹, R Flatman ¹, D Kanowski ¹, L Price ¹, G Ward ¹

Introduction

Urinalysis dipstick testing is used for screening prior to more definitive testing. Manufacturer instructions recommend dipping the strip in urine and blotting the excess with a tissue before interpreting the colour change. Urine specimens with multiple test requests are occasionally shared between instruments, which introduces potential contamination from prior dipstick testing. This study investigated the potential for contamination of urine samples by dipsticks.

Methods

Two patient urine samples (S1 and S2) were aliquoted into two 10 mL Technoplas tubes, with the original sample treated as a control. A Siemens Multistix 10 SG dipstick was rapidly immersed into the first aliquot of urine (T1) and another dipstick was placed in the second aliquot for 10 minutes (T2). The samples were then analysed for 15 trace elements through ICP-MS and 26 routine urine analytes on Abbott c16000 analyser.

Results and Conclusion

False positive results were observed at T1 and T2 for arsenic, iron and iodine. T1 arsenic concentration increased by approximately 16 μg/L for S1 and 42 μg/L in S2. T1 iodine concentration increased >100 μg/L in both S1 and S2. Iron increased >120 μg/L in S1 and >390 μg/L in S2. At T2 the levels for S1 and S2 were as follows: arsenic >480 μg/L; iodine >4900 μg/L; and iron >6000 μg/L.

Arsenic levels increased above the biological exposure standard of 35 μg/L recommended by Safe Work Australia. This is a notifiable result and arsenic speciation testing would be required. Iodine levels increased >100 μg/L. The WHO classification for iodine deficiency is <100 μg/L, therefore a patient with severe deficiency may falsely appear iodine replete causing significant clinical consequences. It is recommended that a dedicated sample should be used for dipstick urinalysis to prevent inaccurate and clinically misleading results.

CBR. 2021 Dec;42(Suppl 1):S30.

P66 HARMONISATION OF THE ANALYTICAL STEPS LEADING TO THE PRECIPITATION OF MACROPROLACTIN

R Guy ^1,^✉, T Yen ², L Boscato ³, P Ward ¹

Introduction

Despite the numerous publications on PEG precipitation for identification of macroprolactin, very few have validated the PEG precipitation procedure. Favresse et al. published results testing each step which was valuable information. A 2021 AACB survey revealed that dilution (1:2) and use of 25% PEG 6000 were standardised across laboratories, but there were significant variations in PEG handling and centrifugation. Harmonisation of macroprolactin testing requires standardisation and cooperativity from participating labs. The latter is enabled when evidence of variations in procedure are shown to be insignificant. The NSWHP harmonised macroprolactin analytical procedure was produced from local validation study and proposed for Australia and New Zealand wide harmonisation.

Methods

A random selection of samples was subjected to a number of experiments where the outcome was measured in terms of monomeric prolactin or the recovery of monomeric prolactin. Experiments included comparing high speed centrifugation (17000g) with normal (2000g); centrifugation at room temperature (RT) vs 4°C; PEG prepared in water vs that prepared in PBS and optimum incubation time for the precipitation reaction.

Results

There was no significant difference in monomeric prolactin when centrifugation at 17000g for 5 minutes @RT (85000gmins) was compared to centrifugation at 2000g for 30 minutes @RT (60000gmins). There was no significant difference in monomeric prolactin when samples were centrifuged at RT or 4°C. PEG prepared in PBS showed less variability but less binding than PEG prepared in water which was significant and so PBS was considered the superior solvent. Precipitation occurred immediately PEG was added to serum - an arbitrary time of 10 minutes was chosen.

Conclusion

Macroprolactin detection is both technical and analytical and small variations in the technical steps hamper progress to a single harmonised macroprolactin PEG precipitation procedure. Harmonisation is more likely when less restrictive technical steps are able to be accommodated especially when evidence is provided.

CBR. 2021 Dec;42(Suppl 1):S30.

P67 A SINGLE POINT CALIBRATION METHOD FOR LC-MS/MS URINARY FREE CORTISOL ASSAY

R Zhang ^1,^✉, M Rybak ¹, D Lau ¹, C Boyder ¹, J Joseph ¹

Introduction

A single point calibration (SPC) method has been established with an in-house developed LC-MS/MS urinary free cortisol (UFC) assay. This method is based on an instrument-specific near linear response range (NLRR) which can accommodate the desired assay concentration range at defined instrument operative sensitivity levels.

Method

Cortisol, cortisone and prednisolone are analysed by this assay. The LC-MS/MS response profile of these analytes has been studied by mapping out the entire instrument response range with an analyte concentration gradient. A NLRR is identified for these analytes based on the measurement deviations between the calculated values and the assigned values.

This assay uses solid phase extraction (SPE) method for sample preparation. 200 μL of urine sample is extracted by SPE, and analytes are detected by LC-MS/MS in the multiple reaction monitor (MRM) mode. A comparison study between the SPC method and multi-point calibration (MPC) method is carried out with more than 100 patients.

Results

Assay calibrated concentration range for UFC is 2–1363 nmol/L. Higher concentration samples can be measured with sample dilution. Assay recovery, linearity, carryover, LLOD/LLOQ, ULOQ, accuracy, intra-day and inter-day precision, sample stability were evaluated and they have met the clinical requirements. The comparison study between the SPC and MPC method has shown excellent agreement between the two methods.

Conclusions

The result of comparison study between the SPC and MPC method confirms that within an instrument-specific NLRR, SPC can have same level of accuracy as the MPC method.

CBR. 2021 Dec;42(Suppl 1):S31.

P68 MONOCLONAL OR POLYCLONAL GAMMOPATHY? LESSONS FROM A CASE STUDY

KW Choy ¹, E Georgiadis ¹, J Abcede ^1,^✉, D Azad ¹, R Liwayan ¹, M Pett ¹

Case

A 52-year-old female was referred to a haematologist for investigation of blood film abnormality (significant rouleaux) and high serum globulin (69 g/L). She had anaemia and neutropaenia with high erythrocyte sedimentation rate (ESR). Total immunoglobulin G (IgG) and IgA were high at 42 g/L (RI 7.0–16.5 g/L) and 7.4 g/L (RI 0.6–4.6 g/L), respectively. Serum free light chain kappa-to-lambda ratio was essentially normal. On serum protein electrophoresis (SPE) (capillary electrophoresis) and immunofixation (IFX), a restricted band was seen in the gamma globulin region, which mimicked the appearance of a monoclonal band. Immunotyping suggests polyclonal gammopathy. The differential for increases in IgG and IgA is broad, including chronic infections, autoimmune disease, and rarely, lymphoproliferative disorders. The patient was subsequently found to have newly diagnosed chronic infection. The polyclonal IgG was more than twice the upper limit of normal and this explains the rouleaux on blood film.

Discussion

Hypergammaglobulinaemia may be due to polyclonal and/or monoclonal immunoglobulin production. Polyclonal hypergammaglobulinaemia is usually associated with systemic inflammatory conditions and it is typically detected as a broad distinct band on SPE. However, pseudo-monoclonal gammopathy has been reported and incorrect diagnoses of monoclonal gammopathy may lead to unnecessary extensive workup including bone marrow biopsy.¹ Large polyclonal increases in IgG4 may produce a relatively broad restriction in the gamma globulin region that may mimic monoclonal gammopathy (even on immunotyping).²

A reliable and routinely available method is required to differentiate between monoclonal and polyclonal hypergammaglobulinaemia. Serum IgG subclasses measurement is useful if IgG4 disease is suspected. While mass spectrometry-based methods (utilising the molecular mass of monoclonal proteins) can assist in distinguishing monoclonal from pseudo-monoclonal gammopathy, they are not routinely available. Isoelectric focusing is an inexpensive and useful test to demonstrate lack of clonality in pseudoclonal or polyclonal cases.

References

1.Jawad MD, Go RS, Witzig TE, et al. Pseudo-monoclonal gammopathy: a report of four cases. Haematologica. 2017;102:e466–9. doi: 10.3324/haematol.2017.171694. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Finn WG, Gulbranson R, Fisher S, et al. Detection of Polyclonal Increases in Immunoglobulin G4 Subclass by Distinct Patterns on Capillary Serum Protein Electrophoresis: Diagnostic Pitfalls and Clinical Observations in a Study of 303 Cases. Am J Clin Pathol. 2016;146:303–11. doi: 10.1093/ajcp/aqw113. [DOI] [PubMed] [Google Scholar]

CBR. 2021 Dec;42(Suppl 1):S31.

P69 ARTEFACTUAL INCREASE IN THYROGLOBULIN IN A THYROID CANCER PATIENT. SAME ANALYSER, NEW INTERFERENCE

G Ghosh ^1,^2,^✉, C Ariyawansa ², S Carrivick ², V Panicker ¹

Background

In differentiated thyroid cancer, serum thyroglobulin and thyroglobulin antibodies are measured to monitor for persistent or recurrent disease.¹ Not recognising an artefactual cause for increased thyroglobulin in this setting can lead to unnecessary investigations, treatment and patient anxiety. Variability between different immunometric assays is described, thus serial thyroglobulin measurements are recommended to be done on the same assay.²

Case

We report the case of a 72-year-old woman with papillary thyroid cancer who had surgery and radioactive iodine approximately 10 years prior. She had an excellent treatment response and had been in biochemical remission with undetectable thyroglobulin and thyroglobulin antibody concentrations for several years. At routine annual biochemical surveillance, her thyroglobulin level was increased at 50 μg/L (Beckman access). Concurrent anti-thyroglobulin antibody remained <1 kU/L. Thyroglobulin repeated a few weeks later was 16 μg/L. Thyroglobulin antibodies were not detected on either occasion. Previous thyroglobulin concentrations were measured on the same analyser and laboratory.

The results were clinically discordant due to the patient being in remission for several years. Thyroid ultrasound and recombinant thyrotropin stimulated radioactive iodine uptake scan showed no evidence of recurrence. Serial dilution on the thyroglobulin level was found to be non-linear, supporting an artefactual cause of the result. The second sample was sent to another laboratory using a different immunometric analyser (Roche) and the thyroglobulin level was found to be <0.1 μg/L. The later result was felt to be reliable given the clinical picture and imaging results.

Conclusion

This case is curious as the patient seemed to newly develop assay interference despite using the same analyser for serial thyroglobulin tests in the past. The patient had experienced a severe hypersensitivity dermatitis requiring prolonged treatment in the interval between tests. We speculate that this event was the cause of the patient newly developing heterophile antibody interference.

References

1.Grebe SK. Diagnosis and management of thyroid carcinoma: focus on serum thyroglobulin. Exp Rev Endocrinol Metab. 2009;4:25–43. [Google Scholar]
2.Schlumberger M, Hitzel A, Toubert ME, et al. Comparison of seven serum thyroglobulin assays in the follow-up of papillary and follicular thyroid cancer patients. J Clin Endocrinol Metab. 2007;92:2487. doi: 10.1210/jc.2006-0723. [DOI] [PubMed] [Google Scholar]

CBR. 2021 Dec;42(Suppl 1):S31.

P70 EVALUATION OF THE SEBIA CAPILLARYS 3 FOR DETECTION AND QUANTIFICATION OF SERUM PROTEINS

MJ Boak ^1,^✉, KL Wan ¹, WHH Lau ¹, P Pavone ¹, ZX Lu ^1,²

Introduction

Capillary electrophoresis (CE) for serum protein electrophoresis (SPEP) has advantages of high throughput and less manual handling. We compared the performance of CE on Sebia Capillarys3 with Hydragel HR agarose gels (AGE) on Sebia Hydrasys2.

Methods

Between-run precision was assessed using Sebia control materials and patient samples with and without paraproteins. Correlation was performed using 111 patient samples, linearity by mixing patient samples (one without paraprotein; one with IgGK of 54 g/L), and carryover by analysing paraprotein samples alternating with water.

Results

For precision, CVs were: albumin (19–57 g/L) ≤1.3%; α1-globulin (2–2.6 g/L) ≤5.7%; α2-globulin (4.1–5.4g/L) ≤3.8%; β1-globulin (3–3.9 g/L) ≤3.8%; β2-globulin (2.1–2.9 g/L) ≤4.5%; γ-globulin (9.1–19.8 g/L) ≤2%.

CVs for paraproteins: IgA, 10.5% at 2 g/L and 0.6% at 58 g/L; IgG, 11% at 1.1 g/L and <0.9% at 13–48 g/L; IgM, 8.9% at 2.5 g/L and 1.3% at 42 g/L.

Good linearity was obtained for IgGK (measured IgGK=1.01 × expected IgGK+1.8, R²=0.996). No carryover was found for paraproteins up to 45 g/L.

Of the 111 samples, 87 intact paraproteins were identified by AGE but 86 by CE. The band (<1 g/L) missed during the initial interpretation was later identified on review. Good overall correlation for paraprotein quantification was obtained for 1–44 g/L (CE=0.983 × AGE+2.1, R²=0.96). CE gave higher values for some samples with IgA paraproteins at 2–5 g/L, IgG at 1–7 g/L and IgM at 2–10 g/L; many of these had bands in the β-region or with polyclonal background. Three samples with free light chains only were not identifiable by both methods but all had immunoparesis which warranted immunofixation.

Conclusion

The Capillarys3 provides satisfactory performance but experience is required to interpret chromatograms. As expected, quantification in β-region is challenging and results may differ between CE and AGE, highlighting the need for parallel testing for patient monitoring when changing method.

CBR. 2021 Dec;42(Suppl 1):S32.

P71 INTER-DAY VARIANCES OF VITAMIN K METABOLITES IN SUBJECTS WITHOUT SUPPLEMENTS

A Meinitzer ^1,^✉, D Enko ¹, S Zelzer ¹, E Fritz-Petrin ¹, N Alonso ¹, G Faustman ^2,³, J Roob ³, B Winkhofer-Roob ², M Herrmann ¹

Introduction

Vitamin K is involved as a cofactor in several physiological processes. To assess the biological day-to-day variance, we developed a liquid chromatography mass spectrometry (LCMS) method for determination of vitamin K1 (phyloquinone) and vitamin K2 (menaquinone-4, Mk-4; menaquinone-7, Mk-7) and investigated 12 healthy volunteers (6 females, 6 males) on five consecutive working days (Monday–Friday). The subjects were not taking nutritional supplements and were advised to avoid significant dietetic changes during this week.

Methods

For measurement of vitamin K analytes, we used a QTRAP 6500 LCMS device (SCIEX, Foster City, CA, US) with atmospheric pressure chemical ionisation. After liquid/liquid extraction for sample clean-up and separation on a special reversed phase column, detection was performed in the multiple reaction mode.

Results

By using 250 μL serum, the lower limits of quantification (LOQ) were 0.05 nmol/L for all analytes. Mean recoveries were between 60% (Mk-7) and 85% (VK-1), intra- and inter-assay CVs were all <15% in the specified working range (0.05–2.0 nmol/L). During the five-day measurements, a significant individual daily variance was observed. Vitamin K1 values ranged from 0.14 to 2.93 nmol/l (median 0.47), Mk-4 from 0.05 to 1.71 nmol/L (median 0.07), and Mk-7 from 0.38 to 2.08 nmol/L (median 0.67). Vitamin K1, the analyte with the shortest half-life, showed the highest inter-day variance.

Conclusion

A simple and rapid LCMS determination of endogenous blood levels of vitamin K and metabolites was presented. The observed high inter-day variance of vitamin K forms might reflect the individual daily dietary vitamin K intake.

CBR. 2021 Dec;42(Suppl 1):S32.

P72 FERRITIN: AN INDEPENDENT PROGNOSTIC MARKER OF MORTALITY IN ADULT HAEMOPHAGOCYTIC SYNDROME COMPLICATED BY HOOK EFFECT

A Hashmi ^1,^✉, D Thomas ¹, F Alvaro ¹, P Stewart ¹

Introduction

We report a case study of erroneously measured decreased ferritin in a patient admitted to ICU and treated for haemophagocytic lymphohistiocytosis (HLH). HLH is a hyperinflammatory syndrome associated with a high mortality rate characterised by extremely high ferritin. In addition to being of diagnostic value, recent studies also validated ferritin as a prognostic marker – the rate of decline of ferritin during treatment in patients with HLH is related to overall survival. Patients with severe SARS-CoV-2 infection may also present with features of HLH due to the severe cytokine storm that may be associated with COVID-19.

Case Report

A 70 yr old man diagnosed six months previously with HLH was admitted to ICU; serial ferritin measurements were requested during treatment. Serum ferritin was analysed on the Roche Cobas e601 (sandwich immunoassay, measuring range 0.50–2000 μg/L, automated onboard dilution if concentration >2000 μg/L).

Serum ferritin increased from 13,375 μg/L on admission to 862,722 μg/L on day 10. On day 13 the initial neat result of 1840 μg/L was not autovalidated only because of a delta check failure; further onboard and manual dilutions gave a final result of 987,000 μg/L. The ferritin concentrations then decreased until the patient expired on day 21.

Ferritin concentration >862,722 μg/L showed a hook effect consistent with a published study (hook effect for Roche assay >838,000 μg/L); the manufacturer claims no hook effect up to concentrations of 100,000 μg/L.

Conclusion

Despite recent advances in automated immunoassays, the hook effect still exists in modern sandwich immunoassays which may not be detected, especially with analytes whose pathologic concentrations span several orders of magnitude, such as ferritin. To avoid the ferritin high dose hook effect, studies have recommended the Roche ferritin upper assay limit be reduced from 2000 μg/L to 1400 μg/L, above which samples are auto-diluted.

[b1-cbr-42-s1] 1.Jawad MD, Go RS, Witzig TE, et al. Pseudo-monoclonal gammopathy: a report of four cases. Haematologica. 2017;102:e466–9. doi: 10.3324/haematol.2017.171694. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b2-cbr-42-s1] 2.Finn WG, Gulbranson R, Fisher S, et al. Detection of Polyclonal Increases in Immunoglobulin G4 Subclass by Distinct Patterns on Capillary Serum Protein Electrophoresis: Diagnostic Pitfalls and Clinical Observations in a Study of 303 Cases. Am J Clin Pathol. 2016;146:303–11. doi: 10.1093/ajcp/aqw113. [DOI] [PubMed] [Google Scholar]

PERMALINK

Proceedings of the Australasian Association of Clinical Biochemistry and Laboratory Medicine’s 2021 Virtual Scientific Conference

A1 DYNAMIC SPECIMEN VALIDITY ANALYSIS IN URINE DRUG SCREENING UTILISING LC-QTOF-MS

D Allen

C Warnholtz

C Bachmeier

B McWhinney

Introduction

Methods

Results

Conclusion

A2 COVID-19 MONITORING APP: A TRIAL IN MOUNT GAMBIER’S MEDI-HOTEL

C Phay

E Kararmaz

G Wilson

R Tirimacco

S Carrison

E Pretorius

Introduction

Methods

Results

Conclusion

A3 VALIDATION OF CEREBROSPINAL FLUID FOR PORPHOBILINOGEN TESTING USING MAUZERALL AND GRANICK METHOD

YC Leong

C Chiang

J Smith

Introduction

Methods

Results

Conclusion

A4 THE ACCURACY OF VITAMIN A AND E MEASUREMENT IN SERUM IS AFFECTED BY THE PRESENCE OF VITAMIN E ANALOGUES

M Fitzpatrick

T Andersen

J Bush

R Greaves

G Woollard

K Hoad

J Collie

S Browning

T Harrower

W Punyalack

Introduction

Methods

Results

Conclusions

A5 CONGENITAL ADRENAL HYPERPLASIA DETECTION AND MONITORING IN DRY BLOOD SPOTS BY LIQUID CHROMATOGRAPHY TANDEM MASS SPECTOMETRY (LC-MS/MS)

AL Schick

BC McWhinney

R Swenson

JPJ Ungerer

Introduction

Method

Results

Conclusion

A6 PREDICTIVE UTILITY OF BIOMARKERS FOR MAJOR CARDIOVASCULAR EVENTS IN PATIENTS WITH PERIPHERAL ARTERY DISEASE

TP Singh

R Velu

F Quigley

J Jenkins

J Golledge

Background

Methods

Results

Conclusions

A7 EVALUATION OF FACTORS IN THE POLYETHYLENE GLYCOL (PEG) PROCEDURE FOR MACROPROLACTIN INVESTIGATION

K Tran

C Ariyawansa

MJ Tanner

Introduction

Method

Results

Conclusion

A8 RETROSPECTIVE REVIEW OF INCREASING VITAMIN B6 WHOLE BLOOD LEVELS IN AUSTRALIA

S Browning

B Teis

M Henman

R Bain

K De Silva

T Harrower

D Holman