Annual meeting of the Royal Belgian Society of Laboratory Medicine: “Symphony of the Heart”

AMI: THE IMPACT OF THE USE OF THE 0/1H ALGORITHM IN A ROUTINE CLINICAL LABORATORY

C. Le Goff ¹

¹Clinical chemistry department, University hospital of Liege, University of Liege and CIRM, Liege, Belgium

Cardiovascular disease is the leading cause of death in industrialized countries, and now also in emerging countries. Chest pain is a common presentation for hospital admissions. Emergency department (ED) admissions for chest pain represent approximately 20% of acute hospital admissions, and delays in the investigation and management of these patients increase the pressure on EDs and medical departments. Given that the majority of chest pain cases do not involve acute myocardial infarction, clinicians rely on troponins as the gold standard for diagnosing this condition. Notably, if the clinical presentation indicates myocardial ischemia, a dynamic elevation of cTn beyond the 99th percentile of healthy individuals serves as a strong indicator of myocardial infarction.

Recently, early rule-out pathways utilizing high-sensitive cardiac troponin (hs-cTn) have garnered recognition from the European Society of Cardiology (ESC) for patients with suspected myocardial infarction. Specifically, the ESC 0/1h algorithm has gained endorsement as the preferred diagnostic approach (Class I B). Although the utilization of the 0/1h algorithm is occasionally met with controversy due to its brevity—merely one hour—the benefits it offers are significant. The algorithm’s rapidity reduces the duration of hospital stays without compromising patient safety for individuals presenting with suspected acute coronary syndrome.

In conclusion, as cardiovascular disease continues to exert a substantial toll on global health, efficient and accurate diagnostic strategies become crucial. The ESC 0/1h algorithm, based on high-sensitive cardiac troponin assays, offers a compelling solution for promptly identifying acute myocardial infarction cases among patients presenting with chest pain. Despite initial concerns surrounding the short time frame, empirical evidence has shown that this approach is safe and effective, leading to reduced hospital stays and optimized patient outcomes.

MANAGING CHRONIC HEART FAILURE: GPS AND NATRIURETIC PEPTIDES AT THE POINT OF CARE

J. Verbakel ^1,2

¹Department of Public Health and Primary Care, KU Leuven, Leuven, Belgium, ²Nuffield Department of Primary Care Health Sciences, University of Oxford, Oxford, United Kingdom

The aging population and improved survival of people with ischaemic heart disease are likely to lead to a continuing increase in the prevalence of heart failure. Overall, a general practitioner with a patient population of 2000 will care for about 40-50 patients with heart failure and see 2-3 new cases each year. Since heart failure may be reversible with appropriate treatment in the early stages of disease, early diagnosis is important.

Considering the low prevalence of heart failure, however, general practitioners are unlikely to have enough experience to identify more subtle signs. Although heart failure is often diagnosed by general practitioners, on the basis of clinical signs, symptoms, and the results of 12 lead electrocardiography, the diagnosis is only confirmed by echocardiography in about 1/3 of cases.

Guidelines from the National Institute for Health and Clinical Excellence, and the European Society for Cardiology (ESC) on the initial diagnosis of CHF and referral for echocardiography recommend the use of B-type natriuretic peptide (BNP) tests in combination with clinical assessment. The use of point-of-care (POC) devices in ambulatory care settings allows BNP results to be available when acute management decisions are needed. As well as reducing turnaround time, POC testing by general practitioners could lead to a quicker investigation of dyspnoea, more timely referral, earlier initial treatment, and less uncertainty and anxiety for patients. The development of POC natriuretic peptide testing services in community settings is part of a general effort to move care from hospital settings to the community and make more point-of-care tests available for a range of conditions. Several POC devices that test for BNP or NTproBNP are available.

In this talk, I will discuss the prevalence of chronic heart failure, the added value of POC natriuretic peptide tests, their diagnostic accuracy and clinical effectiveness, with a focus on ambulatory care settings.

ADULT CARDIOMYOPATHY DUE TO INBORN ERRORS OF METABOLISM

W. Meersseman ¹

¹University Hospitals Leuven, Department of Metabolic diseases, Leuven, Belgium

A normal cardiac function requires an uninterrupted and adaptable supply of energy. Disorders that affect the metabolism of the primary sources of energy (fatty acids, glycogen), or the generation of high energy phosphates in the mitochondria, are often characterized by cardiovascular involvement. While most inherited metabolic diseases are rare (1 per 100000 births), collectively they may affect 1 in 800 to 2500 births. These diseases may present with severe and often fatal manifestations in neonates and children (e.g. Pompe disease, fatty oxidation defects, mitochondrial diseases). Milder forms can be diagnosed in adolescence and adulthood during investigation of multisystem diseases (e.g. mucopolysaccharidoses, organic acidemias, and some glycogen storage disorders). In rare cases the heart is the only affected organ, as for example in cardiac phosphorylase kinase deficiency, lysosomal storage disease (e.g. Fabry’s disease) and some mitochondrial disorders. Early diagnosis is important, since many disorders respond to dietary modification and, in some cases, specific therapy such as enzyme replacement, gene therapy or bone marrow transplantation. Diagnosis also facilitates genetic counseling, prenatal screening, and cascade screening of relatives.

LABORATORY DIAGNOSIS IN ADULTS WITH UNEXPLAINED CARDIOMYOPATHY

P. Vermeersch ¹

¹Clinical Department of Laboratory Medicine, University Hospitals Leuven, Leuven, Belgium

The human heart is the organ with the highest metabolic rate and many inborn errors of metabolism (IEM) cause cardiac disease, in particular cardiomyopathy and cardiac rhythm abnormalities. It is estimated that up to 20% of pediatric and up to 5 % adult cases of hypertrophic cardiomyopathy are caused by inborn errors of metabolism. Evaluation for a possible IEM should therefore be performed in all cases of unexplained cardiomyopathy and cardiac rhythm disorders.

From a clinical perspective, a distinction has to be made between cardiomyopathy in adult patients previously diagnosed with an IEM and adult cardiomyopathy as presenting feature of a previously undiagnosed IEM. Causes of the latter include fatty acid oxidation defects, mitochondrial disorders and lysosomal storage disorders including Pompe and Gaucher, and dolichol kinase deficiency (DOLK-CDG). Symptoms that can point to a possible IEM include hypoglycemia, hepatomegaly, presentation during or after an acute metabolic decompensation, hypertrophic cardiomyopathy with ventricular pre-excitation (VPE) or Wolff-Parkinson-White (WPW syndrome).

In additional to routine laboratory testing (including liver function, ferritin, glucose and lactate) and genetic testing, measurement of acylcarnitines and amino acids in plasma, organic acids and mucopolysaccharides in urine, enzyme testing for Pompe, and transferrin isoforms for CDG can be performed in patients with an unexplained adult cardiomyopathy.

INFECTIOUS MYOCARDITIS

B. Gerber ¹

¹Department Cardiovascular Diseases, University Hospital Saint Luc, UC Louvain, Woluwe, Belgium

Myocarditis is an inflammation of the heart muscle that can be caused by a variety of factors, including viruses, bacteria, toxins, autoimmune diseases or after vaccinations and during cancer immune therapy. It can present with a wide range of symptoms, from chest pain resembling acute coronary syndromes to heart failure and arrhythmias. The clinical course of myocarditis can be acute, fulminant with life-threatening heart failure or arrythmia or chronic. Myocarditis can be suspected clinically based on symptoms, ECG and echocardiographic changes and blood biomarkers (ie Troponin). Definite diagnosis of myocarditis requires endomyocardial biopsy and demonstration of myocardial infiltration by acute inflammatory cells with our without necrosis or fibrosis (WHO-Marburg classification), but because of the invasiveness of this procedure it is only rarely performed. Cardiac MRI (CMR) is the most sensitive and specific non-invasive imaging modality for the diagnosis of myocarditis. CMR can detect myocardial inflammation, edema, and scarring, even in the early stages of the disease. In particular it allows to distinguish myocarditis from mimicking diseases, such as MINOCA (myocardial infarction with normal coronary arteries), Non-ST myocardial infarction, Takutsubo disease, other inflammatory diseases such as sarcoidosis and cardiomyopathies. CMR can also be used to assess the severity and extent of myocardial involvement. CMR is recommended as class I indication for patients with suspected myocarditis by the lasts ESC guidelines, but its diagnostic accuracy is better if it is performed early after presentation and depends on the magnitude of troponin elevation. It is also useful in the follow-up of patients with myocarditis, since it can assess the resolution of inflammation, the development of myocardial scarring, and the development of complications such as heart failure. Cardiac MRI extent and late-gadolinium enhancement patterns were also shown to have important prognostic value in patients with suspected myocarditis. Most patients with acute myocarditis have benign course with complete healing, however some patients progress to chronic heart failure. Currently no specific treatment for myocarditis is available besides heart failure medications.

PREANALYTICAL VARIABLES IN THE MEASUREMENT OF IONIZED CALCIUM IN OUT-OF-HOSPITAL BLOOD SAMPLES

K. Claesen ¹, N. Callewaert¹

¹Clinical Laboratory, AZ Groeninge, Kortrijk, Belgium

Objectives: Clinical demand has increased for ionized calcium (iCa²⁺) measurement in settings where rapid testing is not available and where blood is collected in tubes rather than in syringes. However, little information is available on which matrix (serum or lithium heparin whole blood) is preferred as an alternative for a blood gas syringe containing electrolyte-balanced heparin and on the time span within which iCa²⁺ can be measured reliably in these matrices. The present study aimed to investigate these preanalytical variables, as well as to determine the degree of haemolysis above which there is a significant effect on iCa²⁺ concentration.

Material and Methods: Blood from volunteers was collected into multiple serum (BD Vacutainer SST II Advance) and lithium-heparin tubes (BD Vacutainer LH 68 I.U.) and into one blood gas syringe containing electrolyte-balanced heparin (BD Present Critical Care Sampling kits, Ca²⁺ LH 80 I.U.). iCa²⁺ was measured (Radiometer ABL90 FLEX blood gas analyser) immediately and 1h, 2h and 3h after collection. Haemolyzed samples were obtained by spiking different amounts of concentrated hemolysate (0 – 1500 mg/dL) to serum. H-index was assessed on Cobas 6000 c501 (Roche Diagnostics) and iCa²⁺ measured using a blood gas analyser.

Results: iCa²⁺ measured in a lithium-heparin tube immediately after collection was significantly lower (0.03 mmol/L) than in a blood gas syringe. During storage of lithium-heparin tubes at room temperature, a time-dependent increase in iCa²⁺ was observed, which was the result of ongoing anaerobic glycolysis. In a serum tube, iCa²⁺ was significantly higher (0.03 mmol/L) after collection than in a blood gas syringe and remained unchanged up to 3h30 after collection. In addition, a significant change in iCa²⁺ occurred at an H-index of 338 mg/mL (95% CI 294 – 392 mg/mL).

Conclusion: Despite the unexplained higher iCa²⁺ in serum tubes, the use of this matrix for delayed measurement of iCa²⁺ seems the most suitable/appropriate given the absence of time-dependent instability and the absence of heparin that can bind Ca²⁺. Furthermore, as long as the H-index of a serum sample is below 338 mg/dL, there is no significant haemolytic interference on the measurement of iCa²⁺.

CRITICAL RESULTS IN THE CLINICAL LABORATORY: AN EVIDENCE- AND CONSENSUS-BASED APPROACH

K. Truijens ¹, S. Wijnants¹, G. Frans¹, P. Vermeersch¹

¹UZ Leuven, Leuven, Belgium

Objective: While the concept of critical results has been known for a long time and is a requirement for various accreditation standards (including ISO15189), there is a wide diversity in the critical results used by clinical laboratories. Due to the lack of guidelines, clinical laboratories typically rely on a combination of in-house agreed and/or literature-based critical values. The aim of this study was to create an alarm list by summarizing the best available literature evidence and involving the clinician’s opinion to account for the local setting and the needs of different patient populations.

Material and Methods: A systematic literature review was conducted from 2014 to 2022. Selected articles were ranked based on the level of evidence using an adaptation of the Stockholm Hierarchy (Sikaris K. Clin Biochem Rev. 2012). The current critical results in UZ Leuven (a tertiary care hospital with an ISO15189-accredited clinical laboratory) were reviewed and a new alarm list of 16 routine hematology and chemistry parameters was proposed based on the best available evidence. To evaluate the clinician’s opinions on this list, we conducted a survey for all physicians at UZ Leuven using Qualtrics (Qualtrics.com). Survey results were analyzed using Power BI (Microsoft).

Results: We selected 14 articles that met the predetermined inclusion criteria. Unlike our current alarm list, mainly based on grade 4 evidence (67%), the newly compiled alarm list was only based on grade 1, 2a and 2c evidence. The survey achieved a 30% response rate (163/541). Generally, there was a clear preference to transition to the proposed alarm list (average: 68.84%). Additionally, most respondents favored a single threshold across patient contacts (average: 87.75%).

Conclusion: To address the current lack of evidence on critical value policy, we compiled an alarm list based on the best current evidence and demonstrated that local clinicians had a favorable opinion towards the newly proposed critical results.

PSEUDOHYPERPHOSPHATEMIA IN A PATIENT WITH MULTIPLE MYELOMA

M. De Bruyn ¹, M. Cuykx¹

¹Universitair Ziekenhuis Antwerpen, Edegem, Belgium

Objective: This work highlights the existence of interfering paraproteins in the measurement of serum phosphate resulting in falsely elevated phosphate levels and aims to provide accessible solutions to eliminate this interference.

Material and Methods: In a patient known with IgG multiple myeloma and unexplained hyperphosphatemia, serum phosphate (phosphomolybdate UV assay) and IgG concentrations (immunoturbidimetry) were measured on Atellica Solution (Siemens-Healthineers). To investigate the effect of the paraprotein on the phosphate levels, phosphate was measured in one serum sample before and after protein removal by either dilution, protein precipitation (20% sulphosalicylic acid, or zinc sulphate in methanol) or ultrafiltration (Microcon 30 kDa, Sigma-Aldrich).

Results: A patient with multiple myeloma presented with an unexplained hyperphosphatemia which correlated positively with serum IgG concentrations. As serum dilution normalised the phosphate level, it was hypothesised that precipitation of the paraprotein during the assay reaction interfered with the measurement and resulted in pseudohyperphosphatemia. However, dilution is not optimal as it increases the measurement error.

Protein removal by precipitation with 20% sulphosalicylic acid efficiently reduced the IgG level below the detection limit but increased the phosphate level even further. This indicates efficient protein removal, but suggests a different interference in the phosphate assay. Similarly, protein precipitation with zinc sulphate interfered with the assay as it resulted in an unmeasurably low phosphate level. Successful removal of protein and a correct phosphate measurement were obtained by ultrafiltration. IgG levels were below the detection limit and phosphate was normalised from 2.45 mmol/L to 1.07 mmol/L.

Conclusion: Paraproteins can interfere with the reaction components in the phosphomolybdate UV assay and result in pseudohyperphosphatemia. It is important to be aware of this phenomenon to avoid unnecessary testing and therapy.

A CASE OF MESALAZINE CRYSTALLURIA

M. Janssen ¹, T. Yilmaz¹, R. Gadisseur¹

¹Clinical Chemistry Department, University Hospital of Liege, Liege, Belgium

Objective: Crystalluria is sometimes associated with a pathological phenomenon leading to kidney failure. Identifying drug crystals in urine is quite challenging. Indeed, drug crystals are quite rare and can be present in different shapes. Mesalazine, also called 5-aminosalicylic acid (5-ASA), is indicated for Crohn’s disease and is one of the rare crystallogenic drugs. Few articles report the precipitation of mesalazine crystals using automated or polarization microscopy. The aim of our study is to describe a method to identify mesalazine crystals in urine.

Material and Methods: We received urine from a 47-year-old woman with Crohn’s disease who was treated with four grams of mesalazine per day. This urine showed pH = 6, proteinuria > 3 g/L and atypical crystals on the urine sediment from an automated microscopy system. We therefore decided to analyze this homogenized urine by polarized light microscopy. It was not possible to identify the nature of the crystals with certainty. A dry urine pellet was analyzed by Fourier transform infrared spectrometry (alpha-T1, Bruker®, Germany).

Results: The crystals observed by polarization microscopy were highly polarizing, polychromatic needles organized in a “feather fan” or in circular clusters. Infrared analysis of the dry urine pellet identified mesalazine by comparing the spectrum obtained with a reference spectrum library. The spectrum of mesalazine showed two typical peaks around 2900 cm^-1 and numerous fine peaks between 1600 and 500 cm^-1. This pathologic result led the patient to discontinue treatment to avoid a future renal failure or kidney stone.

Conclusion: We have shown for the first time the appearance of mesalazine crystals in urine using polarization microscopy. Recognition of the crystals requires some expertise due to the different shapes that the crystals can assume. Infrared analysis is usually essential to identify these crystals with certainty.

STUDY OF THE EXPRESSION OF THE MTBI BIOMARKERS, GFAP AND UCH-L1, IN A NORMAL AGING POPULATION

E. Calluy ¹, C. Beaudart², O. Bruyère³, J.-Y. Reginster⁴, E. Cavalier¹, A. Ladang¹

¹Clinical Chemistry / CHU de Liège, Liège, Belgium, ²Research institue for life science / Université de Namur, Namur, Belgium, ³WHO Collaborating Center for Public Health Aspects of Musculoskeletal Health and Aging, Division of Public Health, Epidemiology and Health Economics /Université de Liège, Liège, Belgium, ⁴WHO Collaborating Center for Public Health Aspects of Musculoskeletal Health and Aging, Division of Public Health, Epidemiology and Health Economics, Université de Liège, Liège, Belgium

Introduction: Mild traumatic brain injury (mTBI) is one of the most common conditions found in emergency department and can lead to significant long-term consequences. Since older adults are the most at risk of suffering from comorbidities such as bleeding, aging subjects are also the most at risk of mTBI. Nowadays, the gold standard for the diagnostic of mTBI is the computed tomography scan (CT scan) but recently new biomarkers such as GFAP and UCH-L1, have been proposed for the rule out of mTBI.

Objectives: The aim of this study was to evaluate confounding factors that influence the concentration of Glial Fibrillary Acidic Protein (GFAP), and Ubiquitin carboxyl-Terminal Hydrolase L-1 (UCH-L1) to evaluate their specificity in non-suspected cases.

Methods: We measured, on an Alinity I, GFAP and UCH-L1 concentrations (mTBI test) in 341 individuals from the SarcoPhAge cohort, a long-term prospective study in community-dwelling subjects over 65 years old.

Results: The major confounding factors of GFAP and UCH-L1 expression were age, BMI and renal function. According to cut-offs provided by Abbott, the mTBI test was positive in 67% of normal aging participants and the positivity rate increases to 90% when the cystatin C concentration is over 1.55 (n=20) and to 97.92% in participants older than 80 years (n=48). Additionally, positivity rate was solely driven by GFAP confounding factors.

Conclusions: Our data show that the specificity of the mTBI test is low in aging people. Therefore, its added-value to rule-out mTBI in aging people in clinical practice will be limited.

MILD TRAUMATIC BRAIN INJURIES CAN BE EFFECTIVELY RULED-OUT BY MTBI TEST FROM ABBOTT BUT TEST SPECIFICITY CAN BE INCREASED BY AGED-DEPENDENT CUT-OFFS

A. Ladang ¹, I. Trifonidi², G. Vavoulis³, C. Leens¹, A. Perot⁴, L. Vranken¹, K. Karagianni³, A. Mitropoulos³, E. Cavalier¹, K. Makris²

¹Clinical Chemistry / CHU de Liège, Liège, Belgium, ²Clinical Biochemistry Department / KAT General Hospital, Athens, Greece, ³Neurosurgery Department / KAT General Hospital, Athens, Greece, ⁴Emergency department / CHU de Liège, Liège, Belgium

Background: mild traumatic brain injury (mTBI) is a common cause of emergency department admission. Today, its diagnosis is based on clinical symptoms and imaging techniques such as CT-scan. Abbott has recently developed the mTBI test, based on the dosage of GFAP and UCH-L1, to rule-out the presence of mTBI. The aim of the study was to evaluate the diagnostic properties of this test in a Greek and a Belgian cohort.

Methods: To look for confounding factors of GFAP and UCH-L1 expression, 163 Greek and 20 Belgian healthy subjects were selected as reference individuals. To evaluate diagnostic properties of the kit, 246 Greek and 22 Belgian suspected to have a mTBI were recruited. All subjects were submitted to a CT-scan to establish the diagnosis of mTBI. GFAP and UCH-L1 were measured on Alinity (Abbott).

Results: Among reference individuals, GFAP and UCH-L1 expression were found to be associated with age in a quadratic manner (r²: 0.2985 and 0.2562, respectively; both p-value: 0.0001). In the whole mTBI Greek cohort, although the percentage of positive CT-scans remains constant, the percentage of mTBI positive test increased with age ranging from 51% in 18 to 29 years old patients to reach 96% in patients >80 years old, showing that cut-offs should be adjusted for age. Indeed, although the sensitivity of the test is 100%, the specificity is 39.6% for the whole cohort but only 16.5% in patients >60 years old. Thus, we have calculated new cut-offs for which sensitivity remains of 100% but with an increased specificity. By applying a cut-off of 107 pg/mL for GFAP and 335 pg/mL, the specificity was increased to 49 % in patients >60 years old without modifying sensitivity. These new-cut-offs were validated in the Belgian cohort where specificity increased from 8 % to 17% with a 100% sensitivity.

Conclusions: mTBI test is an efficient rule-out test to exclude patients suffering from concussion. However, the rule-out rate in elderly patients is very low and age-dependent cut-offs could increase the attractiveness of this test.

RESOLVING ANALYTICAL CHALLENGES OF SERUM IMMUNOFIXATION ELECTROPHORESIS IN IGD MULTIPLE MYELOMA BY USING MASS SPECTROMETRY

L. Nevejan ¹, M. Perkins², T. Vanhove³, M. Delforge⁴, M. Vercammen¹, X. Bossuyt³

¹Department of Laboratory Medicine, AZ Sint-Jan Hospital Brugge-Oostende AV, Brugge, Belgium, ²Binding Site - part of Thermo Fisher Scientific, Birmingham, United Kingdom, ³Department of Laboratory Medicine, University Hospitals Leuven, Leuven, Belgium, ⁴Division of Haematology, University Hospitals Leuven, Leuven, Belgium

Objective: Serum immunofixation electrophoresis (sIFE) from a multiple myeloma (MM) patient in biochemical progression showed free monoclonal λ light chains (LC) and monoclonal IgD heavy chains (HC), the latter without corresponding LC (Figure 1A). To exclude a theorical δ-HC disease, novel mass spectrometry (MS) assays were performed as an alternative method to detect M-proteins in serum.

Material & Methods: Two serum samples of the index patient, drawn three weeks apart, together with seven stored serum samples of different IgD λ MM patients were analysed with two MS approaches. First, immunoglobulins were purified using beads coated with antibodies specific for human IgG -, IgA -, IgM HC and total κ -, total λ -, free κ - and free λ LC. Immunopurified samples were analysed by MALDI-TOF MS (Intact LC MS, EXENT® solution). Second, to confirm IgD specificity, total λ immunoprecipitation samples were digested using trypsin and analysed by liquid-chromatography MS (LC-MS) (BioAccord, Waters). Fragmentation data was screened for unique IgD – and IgE HC sequences.

Results: In all samples (n=9), EXENT® identified a peak in total λ spectra which had an identical mass-to-charge ratio (m/z) as the peak observed in free λ spectra of the respective sample (Figure 1B). In the index patient, the peaks in total λ and free λ spectra were identical in both samples. Tryptic digest of total λ immunoprecipitation followed by LC-MS identified peptides associated with IgD HC sequence in all samples (range: 29-70% homology). No hits were identified for IgE HC in any sample.

Conclusions: Conventional sIFE failed to unequivocally detect intact monoclonal IgD λ. By using two MS approaches, intact IgD λ M-proteins with associated monoclonal free λ LC of the same molecular mass as the LC derived from the intact IgD M-protein are likely identified in all samples. Intact LC MS is a promising successor to conventional sIFE.

Figure 1. M-proteins detected in the index patient by using (A) sIFE and (B) MS

SIMULTANEOUS DETECTION AND QUANTIFICATION OF GASTRIN 17 AND 34 SULFATED AND NON-SULFATED FORMS BY LC-MS/MS IN HUMAN PLASMA

L. Huyghebaert¹, P. Massonnet¹, E. Grifnée¹, J. Demeuse², T. Dubrowski¹, M. Schoumacher¹, S. Peeters¹, E. Cavalier³, C. Le Goff ³

¹CHU de Liège, Liège, Belgium, ²University of Liège, Liège, Belgium, ³University of Liège, CHU de Liège, Liège, Belgium

Objective: Gastrin, secreted by G cells, plays a crucial role in digestion and has diverse functions including regulation of the intestinal epithelium and stomach growth.

Gastrin peptides are derived from progastrin. Peptides G17 and G34 are the most abundant in the blood. Both of them may be sulfated.

Current gastrin measurement relies on the DIAsource RIA kit, however it displays cross-reactivity issues. Therefore, we developed a LC-MS/MS method to quantify both sulfated and non-sulfated G17 and G34 forms.

Materials and Method: The samples were extracted by solid-phase extraction and then concentrated. We assessed both ESI- and ESI+ modes. For ESI - mode, prepared samples were directly injected whereas analysis in ESI + mode required an additional derivatization step.

Quantification was achieved on a Nexera X2 UPLC (Shimadzu Corporation, Kyoto, Japan) coupled to a QT6500 mass spectrometer (Sciex, CA, USA). Regarding ESI- mode, chromatographic separation was performed on a Acquity UPLC BEH C18 130Å column (100 × 2,1 mm, 1.7 µm) (Waters, Massachusetts, USA). The mobile phases composition was H₂O and acetonitrile both containing 6mM NH₄OH. On the other hand, for ESI + mode, chromatographic separation was performed on a Acquity UPLC Protein BEH C4 300Å column (100 × 2,1 mm, 1.7 µm) (Waters, Massachusetts, USA). The mobile phases composition was H₂O and acetonitrile both containing 0,1% formic acid (FA).

Results: Given the number of carboxilic acids in the sequence, source parameters and transition were first optimized in negative ion mode. Once the inlet and MS method were optimized, the sample preparation was developed. We achieved a successful separation of the four gastrin peptides with recovery rates ranging from 43% to 62% and no matrix effect.

Column lifetime issues likely due to the high pH of the mobile phases led us to explore ESI + mode. It allowed for the use of mobile phases with pH levels better suited for chromatographic columns.

Conclusion: Despite the necessary change in strategy, we are close to finish the development and validate our LC-MS/MS method that will enable the simultaneous quantification of G17 and G34 sulfated and non-sulfated forms.

COMPARISON BETWEEN TWO URINE DIPSTICKS FOR THE DETECTION OF PROTEINURIA AND THE RATIO OF URINE PROTEIN-TO-CREATININE

S. Goes ¹, N. Graindor, T. Schiemsky, J. Penders

¹Ziekenhuis Oost-limburg, Genk, Belgium

Objective: The aim of this study was to compare the analytical performance of two different urine sticks for the detection of proteinuria.

Material and Methods: Repeatability and reproducibility were assessed using a positive control for both the Meditape UC-11A and the FUS-10II strips. For method comparison, an unpaired retrospective study was conducted from January 2022 to June 2022 of 7654 Meditape and 1628 FUS strip results, performed on the UC-3500 (Sysmex, Japan) and FUS 3000 (Dirui, China), respectively. To evaluate the performance of the strips, the quantitative protein and protein-to-creatinine ratio (<0.2) determination on the cobas c502 analyzer (Roche Diagnostics, Germany) was used as a reference standard. Strips are considered positive according to the manufacturer if the result is ≥1+ and were classified as discordant in this study if the total protein was negative (<0.15 mg/dL).

Results: Both strips repeatability (n=10) and reproducibility (n=10) for protein had an agreement above 90%, assessed according the Clinical and Laboratory Standards Institute (CLSI) guideline EP12-A2 (1). A total of 10% and 7% of discordant strips relative to the quantitative total protein measurement were found for the UC-3500 and FUS 3000 strips respectively. A false positive and negative rate were for both strips below 10% and 5% respectively. The sensitivity and specificity of the Meditape and FUS strips did not differ significantly, being 99% and 94% (p=.11) with specificities of 79% and 91% (p=.18), respectively. Relative to the protein-to-creatinine ratio, the specificity and sensitivity of the Meditape strips (n=3636) were 80% and 85%, respectively. The specificity but not the sensitivity significantly differed from the FUS strips (n=583) with 91% (p=.048) and 76% (p=.35) respectively.

Conclusion: Both urine dipsticks are an equally good screening method for the detection of proteinuria when compared to the quantitative detection of protein-to-creatinine in urine.

References:

(1) Clinical and laboratory standards institute (CLSI). User Protocol for Evaluation of Qualitative Test Performance; Approved Guideline - Second Edition. CLSI Document EP12-A2. Wayne, PA:CLSI,2008.

VALIDATION OF A LC-MS/MS METHOD FOR THE SIMULTANEOUS QUANTITATION OF ANGIOTENSIN (1-7), (1-8), (1-9) AND (1-10)

J. Demeuse¹, L. Huyghebaert², W. Determe¹, M. Schoumacher², E. Grifnée², P. Massonnet², T. Dubrowski², S. Peeters², E. Cavalier³, C. Le Goff ³

¹University of Liège, Liège, Belgium, ²CHU de Liège, Liège, Belgium, ³University of Liège, CHU de Liège, Liège, Belgium

Objective: Cardiovascular diseases have cast a significant negative impact on the lives of millions worldwide. Over the years, extensive efforts have been dedicated to enhancing diagnostic and prognostic tools for these diseases. A growing body of evidence indicates that the angiotensin convertase enzyme and the angiotensin convertase enzyme 2, and angiotensin peptide levels could hold a pivotal role in assisting clinicians with the management of cardiovascular conditions, notably hypertension and heart failure. However, despite the considerable body of knowledge in this domain, a void remains in the field of analytical methodologies for these molecules. In this study, we present a fully validated LC-MS/MS method for the precise quantitation of plasma angiotensin (1-7), (1-8), (1-9), and (1-10), following the guidelines set by the Clinical and Laboratory Standards Institute (CLSI).

Method: Our LC-MS/MS method was rigorously validated with 8 calibration standards (5-5000 pg/mL), analyzed in quadruplicate over 5 days. Validation covered selectivity, linearity, calibration curve, sample preparation recovery, ion suppression, matrix effect, carryover, lower limit of the measurement interval (LLMI) set at 5 pg/mL, and measurement uncertainty, all following CLSI guidelines.

Results: All validation parameters met CLSI criteria. Our method effectively separated angiotensin peptides and displayed linear calibration within our measurement range. Sample preparation was robust against matrix effects and ion suppression, with high recovery rates for all peptides. Accuracy and precision were satisfactory, even at LLMI.

Conclusion: Our validated LC-MS/MS method enables precise quantification of angiotensin peptides, offering a valuable tool for cardiovascular disease research.

DEVELOPMENT OF A QUANTITATION METHOD FOR PARATHORMONE RELATED PROTEIN (PTHRP) BY LC-MS/MS

P. Massonnet¹, L. Huyghebaert¹, E. Grifnée¹, J. Demeuse¹, W. Determe¹, M. Schoumacher¹, D. Dubrowski¹, C. Cavalier¹, C. Le Goff ¹

¹Department of Clinical Chemistry - CHU de Liège, Liège, Belgium

Objective(s): Parathyroid hormone-related protein (PTHrp) is a protein that has a high sequence similarity with the parathyroid hormone (PTH)(1). In humans, 3 main species are reported : 1-139 ; 1-173 and 1-141. Because of their sequence similarities, PTH and PTHrp act on the same biological receptor. By opposition to PTH, that has one endocrine activity, PTHrp has one paracrine activity.

Produced in high amount by cancer cells, PTHrp is known to be at the origin of the tumor hypercalcemia syndrome(2). For now, PTHrp is mainly quantified using radio-immuno assay (RIA)(3). However, these assays shows a lot of cross-reactivities, making difficult a good interpretation of the results. In order to overcome these problems, the goal of this study is to develop one LC-MS/MS method for the quantitation of PTHrp in human plasma.

Method: For the LC-MS/MS method, quantification was achieved on a Nexera X2 UPLC (Shimadzu Corporation, Kyoto, Japan) coupled to a QT6500 mass spectrometer (Sciex, CA, USA). For the LC method, a XSelect PRM PST HSST3 2.5μm 2.1x150mm column (Waters, UK) has been used. Water and acetonitrile were used as mobile phases (each with 0.01% formic acid). Standards of PTHrp were digested by trypsin and signature peptides were searched.

Results: Two transitions of a signature peptide after tryptic digestion have been found and a first attempt of sample preparation has been tried. Interferences are still found in chromatogram and spikes of 5 ng/ml of PTHrp in human plasma can for now be detected.

Conclusion(s): Signature peptide, source parameters, LC gradients and transitions have been found. However, improvements are still needed in term of sample preparation for an accurate and robust quantitation of PTHrp in human plasma.

1. McCauley LK, Martin TJ. Twenty-five years of PTHrP progress: From cancer hormone to multifunctional cytokine. Journal of Bone and Mineral Research. John Wiley & Sons, Ltd; 2012;27:1231–9.

2. Gensure RC, Gardella TJ, Jüppner H. Parathyroid hormone and parathyroid hormone-related peptide, and their receptors. Biochemical and Biophysical Research Communications. Academic Press; 2005;328:666–78.

3. Fraser WD, Robinson J, Lawton R, Durham B, Gallacher SJ, Boyle IT, et al. Clinical and laboratory studies of a new immunoradiometric assay of parathyroid hormone-related protein. Clinical Chemistry. Oxford Academic; 1993;39:414–9.

ANALYTICAL AND CLINICAL EVALUATION OF A NEW POC : AFIAS TN-I PLUS ASSAY

G. Pittie ¹, P. Lukas¹, M. Massart¹, E. Cavalier¹, C. Le Goff¹

¹Department of Clinical Chemistry Department, CHU de Liège, CIRM, University of Liège, Liège, Belgium, Liège, Belgium

Background: In emergency departments (ED) and clinical laboratories, point-of-care (POC) systems have already proven their worth. The objective of this evaluation was to determine the clinical performance of the AFIAS POC Tn-I Plus assay (Boditech Med, South Korea) to firstly demonstrate the performance of the assay for the detection of cardiac troponin I (cTn-I) in serum samples and secondly to show that the product is safe and effective for its intended use but also suitable for routine use in clinical laboratories.

Design and methods: Repeatability, reproducibility, limit of detection (LOD) and limit of quantification (LOQ) were assessed according to CLSI guidelines. In addition, the 99th percentile upper reference limit (URL) of cTn-I in an apparently healthy population was estimated. Clinical concordance was checked in 517 sample patients for whom high-sensitivity cTn-I was requested.

Results: The CVs for repeatability and reproducibility were 6.7-8.5% and 7.5-7.6%, respectively. The LOD and LOQ were in accordance to the manufacturer’s specifications of 0.010 ng/mL and 0.030 ng/mL, respectively. The 99^th percentile URLs were 0.0300 ng/mL (7.8% CV) and 0.0239 ng/mL (9.4% CV) for males (18-75 years) and females (17-65 years), respectively. The overall 99th percentile URL was 0.0296 ng/mL (8.2% CV). In the total apparently healthy population, the percentage of measurable cardiac troponin I (cTn-I) values below the 99th percentile (i.e., 0.0296 ng/mL) was 47.7% (391/820 samples). No significant clinical differences were observed between AFIAS Tn-I plus and Abbott ARCHITECT High Sensitive Troponin-I for the diagnosis of acute myocardial infarction (AMI).

Conclusion: The results of our analytical and clinical performance evaluation demonstrate that the AFIAS Tn-I Plus Assay for AMI is comparable to Abbott’s ALINITY STAT High Sensitive Troponin-I. The assay can be used for routine clinical laboratory use.

ASSOCIATIONS BETWEEN ENVIRONMENTAL POLLUTANTS AND THYROID CANCER, A MULTIPOLLUTANTS APPROACH

P. Dufour ¹, V. Cirello², M. Lugaresi³, C. Moneta³, A. Manzo², E. Carbone², C. Colombo², L. Fugazzola², C. Charlier¹, C. Pirard¹

¹Department of Clinical, Forensic and Environmental Toxicology, University hospital of Liege (CHU Liège), Liège, Belgium, ²Department of Endocrine and Metabolic Diseases, Istituto Auxologico Italiano IRCCS, Milan, Italy, ³Department of Biotechnology and Translational Medicine, University of Milan, Milan, Italy

Objectives: In recent decades, the incidence of thyroid cancer (TC) has been increasing worldwide. This increase is partly explained by improved diagnostic procedures. However, an increase in the incidence of more aggressive forms of TC or TC in young people has also been observed, which cannot be explained by better diagnostic efficiency. The effect of environmental pollution on this rising incidence has therefore been suspected, and many studies have investigated the association between pollutants and TC. However, few of them have evaluated the effect induced by simultaneous exposure to several compounds (mixture effect). Therefore, the aim of the present study was to evaluate the associations between TC incidence and exposure to a mixture of pollutants.

Material and methods: 112 patients with TC and 112 controls were recruited at the Istituto Auxologico Italiano Hospital of Milan (Italy). In the serum of the subjects we measured 9 perfluoroalkyl substances, 12 polychlorobiphenyls and 2 pesticides. We calculated individual logistic regressions for each pollutant. In addition, to assess the mixture effect, we computed statistical models using weighted quantile sum regression and Bayesian kernel machine regression methods.

Results: Monopollutant models showed a positive association between TC and PFDA (detected vs not detected) (OR=2.03, 95% CI 1.10-3.75, p=0.02) and a negative association with PFHxS (OR=0.63, 95% CI 0.41-0.98, p=0.04). Both WQS and BKMR models showed a negative association between TC and the pollutant mixture. This negative association was mainly due to the contribution of PFHxS, PFOA, PCB-118 and PCB-180.

Conclusions: The negative association between pollutants and TC is quite surprising, but an inverse causality is suspected. Indeed, alterations of the thyroid axis due to cancer or treatment could increase the metabolism of pollutants and thus decrease their serum concentrations. This underlines the need for longitudinal studies.

COMPARISON RBA AND ELISA FOR THE DETECTION OF ANTI-GAD AND ANTI-IA2 IN TYPE 1 DIABETES

Z. Janssen ¹, A. Evenepoel¹, K. Verhaeghen¹, B. Blomme¹, I. Weets¹

¹Vrije Universiteit Brussel (VUB), Universitair Ziekenhuis Brussel (UZ Brussel), Department of Laboratory Medicine, Bruxelles, Belgium

Objective: Comparison of routinely used radio-binding assay (RBA) and enzyme linked immunosorbent assay (ELISA) for detection of glutamic acid decarboxylase antibody (GADA) and insulinoma-associated protein 2 autoantibody (IA-2A) in type-1 diabetes (T1D).

Materials and methods: The analytical performance and diagnostic advantages of both assays are compared. The cut-off (<10 iU/mL), set by manufacturer, was obtained by analyzing negative controls (GADA n = 300; IA-2A n = 153). These values were evaluated by analyzing control samples for GADA (n = 545) and IA-2A (n = 475) on the Euroimmun analyser I. Sensitivity and specificity were determined by analyzing Islet Autoantibody Standardization Program (IASP) samples from 2023.

Results: Following results were found for analytical performance. The cut-off for GADA - ELISA is set at <7.50 iU/mL. The IA-2A – ELISA cut-off remains <10 iU/mL.

Conclusion: Insulin-dependent diabetes mellitus or T1D is an autoimmune disease, characterized by an asymptomatic period (prediabetes) with destruction of the pancreatic beta cells. The presence of autoantibodies such as GADA and IA-2A are used for detection, prediction and classification of T1D. RBA is the golden standard, but has shortcomings. In comparison, ELISA is more rapid, uses less sample volume and is non-radioactive. Both assays have good performance characteristics. However, higher specificity (84%) was achieved for GADA – ELISA and can therefore enhance the detectability of positive samples. Both sensitivity and specificity were similar when comparing IA-2A assays. However, the ELISA will be implemented as it is more rapid and a fully automated process.

ANALYTICAL PERFORMANCE OF THE LUMIPULSE® G NFL RUO ASSAY FOR BLOOD

A. Ladang ¹, S. Kovacs¹, F. Watar¹, E. Cavalier¹

¹Clinical Chemistry / CHU de Liège, Liège, Belgium

Background: Neurofilament light chains (NfL) provide structural support to neurons and are recognized to be modified in a broad spectrum of neurological diseases and traumatic brain injuries. NfL are strongly expressed in myelinated axons, secreted in the cerebrospinal fluid (CSF) and the blood. NfL blood and CSF are tightly correlated. Thus, this new biomarker offers a wide range of clinical applications. However, only a limited number of fully automated assays are currently available for the determination of blood and CSF NfL levels. The aim of the study is to validate the newly developed Lumipulse G NfL blood RUO assay.

Methods: The LUMIPULSE G SYSTEM is a chemiluminescent enzyme immunoassay platform enabling automated processing of samples using ready-to-use immunoreaction cartridges. The parameters analyzed were: precision for both serum and plasma, sensitivity, proportional linearity and a method comparison with the Simoa NF-light Advantage kit from Quanterix (USA). Comparison between serum and plasma was also performed.

Results: Regarding precision for both plasma EDTA and serum, median analytical CV was below 5% with-in day and below 8% between days. There were no bias between serum, plasma EDTA and plasma heparinate. The limit of quantification was established at 4 pg/mL. Dilution linearity up-to 5 time gave a linear regression. Method comparison with NfL V1 Quanterix assay gave an acceptable bias but 1 sample gave clinically different results.

Conclusion: The new Lumipulse G NfL RUO assay offers suitable analytical performance for use in clinics.

CONCORDANCE BETWEEN APOE GENOTYPING AND QUANTITATIVE MEASUREMENT OF APO E4 ISOFORM FROM FUJIREBIO°

A. Ladang ¹, F. Meyer², S. Kovacs¹, E. Cavalier¹

¹Clinical Chemistry / CHU de Liège, Liège, Belgium, ²Neurology department/ CHU de Liège, Liège, Belgium

Background: Apolipoprotein E (Apo E) is a glycoprotein encoded by the APOE gene. The APOE gene has three different alleles that encode for three different isoforms (Apo E2, Apo E3, Apo E4). Depending on whether patients are homozygote or heterozygote, the level of expression of each isoform will change. Bearing the Apo E4 allele is known to be linked to higher risk of Alzheimer’s disease whereas the presence of Apo E2 allele is known to reduce this risk. In this study, we evaluated whether the APO E phenotype evaluated by the newly launched Lumipulse° G APOE kits was in concordance with the APOE genotyping.

Methods: APOE genotyping and phenotyping was evaluated in 34 patients from an observational Alzheimer cohort. After DNA extraction from whole blood, APOE allele were assessed by PCR amplification after HhaI restriction. Phenotype was determined by the combined measurement of APO E4 and total APO E thanks to Lumipulse° G APOE4 kit and Lumipulse° G Pan-ApoE kit, according to manufacturer’s instruction.

Results: Of the 34 patients, 14 were homozygote for APO E3, 6 were E2/E3 heterozygote and 14 were carrying one E4 allele. Median of Apo E4 / Pan ApoE ratio was 0.2% for E2/E3 heterozygotes, 0.0 % for E3 homozygotes, 20.0 % for E2/E4 heterozygote and 27.2 for E3/E4 heterozygote meaning that all subjects with the Apo E4 allele had a measurable Apo E4 and an increased Apo E4 / Pan ApoE ratio.

Conclusion: Our results show a 100% concordance between ApoE genotyping and Apo E4 phenotyping established through Apo E4 quantitative measurement.

ZINC TRANSPORTER 8 AUTOANTIBODY DETECTION IN TYPE-1-DIABETES: COMPARISON AND VALIDATION OF THREE ASSAYS

A. Evenepoel¹, Z. Janssen ¹, K. Verhaeghen¹, B. Blomme¹, I. Weets¹

¹Vrije Universiteit Brussel (VUB), Universitair Ziekenhuis Brussel (UZ Brussel), Department of Laboratory Medicine, Laarbeeklaan 101, 1090 Brussels, Belgium

Objective: To evaluate the use of a fast and automated zinc transporter 8 autoantibody (fZnT8A) ELISA (RSR, United Kingdom), in comparison to the automated (sZnT8A) ELISA (EUROIMMUN, GERMANY) and current in-house radio bindings assay (RBA) for routine use.

Materials and methods: The analytical performance of the fZnT8A kit was determined using IQC material and control serum. The clinical cut-off was determined on ZnT8A negative control samples (n = 300). The clinical validation was performed by measuring the IASP (Islet Autoantibody Standardization Program) samples of 2020 and 2023. These results were used to compare the three assays

Results: We found following results for the clinical and analytical performance (table 1). The fZnT8A performed better for all the different evaluated parameters except for the sensitivity compared to the sZnT8A. However, this can be explained due to the difference between IASP sample sets.

Conclusions: Type-1-Diabetes (T1D) is an auto-immune disorder presumably caused by an interplay of different factors. Specific autoantibodies are used for risk-assessment, diagnosis, and disease progression. ZnT8A usually appears later in the preclinical phase of T1D. If ZnT8A is present, in combination with another autoantibody, patients are at high risk of developing T1D within a period of 5 years. Ziegler et al. states that an early detection of antibodies could mean a more effective intervention. We found acceptable analytical and clinical performance characteristics for the fZnT8A. This leaves us to believe that this new fast and automated ZnT8A ELISA is a promising assay to be used in T1D work-up.

DEVELOPMENT OF A LIQUID CHROMATOGRAPHY – TANDEM MASS SPECTROMETRY METHOD FOR THE QUANTIFICATION OF OXYTOCIN

E. Grifnée¹, J. Demeuse¹, L. Huyghebaert¹, P. Massonnet¹, M. Schoumacher¹, T. Dubrowski¹, S. Peeters¹, E. Cavalier¹, C. Le Goff ¹

¹Department of Clinical Chemistry, University and CHU of Liège, Liège, Belgium

Objective: Oxytocin is a 9 amino acids peptide that serves as neuromodulator in the human central nervous system. This peptide is implicated in the control of divers social behaviors and plays a role in positive social interaction by decreasing stress and increasing self-confidence. Patients suffering from psychiatric troubles such as depression present important deficit in terms of social and emotional competence. A reliable and sensitive assay for oxytocin is essential to reach relevant conclusions and to support clinical interventions. The purpose of this work was to develop a liquid chromatography coupled to mass spectrometry method to measured oxytocin.

Materials and Methods: Plasma samples have been extracted by solid phase extraction with an OASIS HLB 96-well elution plate, evaporated and then reconstituted. Separation and quantification were achieved on an ACQUITY UPLC H-Class PLUS (Waters, Mildford, USA) coupled to a Xevo TQ-XS (Waters, Mildford, USA). Chromatographic separation was performed on an ACQUITY UPLC BEH C18 column (130 Ǻ, 3 µm, 2.1 mm x 100 mm) from Waters (Mildford, USA). The mobile phase composition was water and acetonitrile both containing 0.1% formic acid.

Results: First, the parent ion and the corresponding transitions are defined and the MS and LC parameters were optimized. The sample preparation was then developed. A recovery rate of 91 % was obtained after optimization of the wash and elution solutions. A calibration curve was performed in cow plasma with calibrators going from 1 ng/mL to 1 pg/mL. At a 1 pg/mL concentration of oxytocin, the peaks were intense enough with a good and symmetrical peak shape.

Conclusion: As the method development is finished, the method needs to be validated.

VALIDATION OF A HIGHLY SENSITIVE IN-HOUSE LC-MS/MS METHOD FOR THE DETERMINATION OF PLASMA CATECHOLAMINES

C. Le Goff ¹, S. Peeters¹, M. Rechchad¹, J. Deleersnyder¹, E. Cavalier¹

¹Department of Clinical Chemistry, University Hospital of Liege, Liege, Belgium

Objectives: The determination of epinephrine (E), norepinephrine (N), dopamine (D) help to diagnose pheochromocytoma, a rare but potentially fatal tumor arising primarily from the chromaffin cells of the adrenal medulla. However, it is not easy to measure due to the low concentrations in plasma. The aim of this work was to validate an in house method for the determination of plasmatic catecholamines.

Materials and methods: Plasma samples were extracted by solid phase extraction and analyzed by an in sample ion-pairing chromatography LC-MS/MS on a ACQUITY UPLC H-Class PLUS (Waters, Mildford, USA) coupled to a Xevo TQ-XS (Waters, Mildford, USA). Chromatographic separation was performed on a Kinetex Biphenyl column (2.6 µm, 2.1 mm x 100 mm) from Phenomenex (Torrance, California, USA). The mobile phases composition was water and methanol both containing 0.1% formic acid.

Validation of the method was realized according the European Medicines Agency (EMA) guidelines. We determined the repeatability, reproducibility, uncertainty, linearity and LOQ on free catecholamine plasma spiked with different known concentrations of the different compounds (E, N, D). We carried out also matrix effect, linearity of dilution, carry-over and recovery. The validation concentration range was 4.2-2083 µg/L for the 3 compounds. Each level was analysed in fourfold in 5 run. A calibration curve (8 points) was systematically carried out with each run validation.

Results: The compounds showed a good linearity for the range tested. The LOQ were: 37, 23.5 and 51.9 ng/L for E, N, D respectively. The repeatability did not exceed 6.8 % for E, 4.7% for N and 4.1% for D. The total precision did not exceed 9.4%, 8.8% and 7.6% for E, N and D. The mean maximum relative bias were 6.3%, 5.9% and 7.5% for E, N and D respectively. For the linearity of dilution, the dilution up to 10 meets the NCCLS recommendations. After experiments, recovery rate was above 77 % while matrix factor for the analyte was above 86%. No carry-over was demonstrated.

Conclusion: Our in-house method was successfully validated. It represents a convincing alternative to the HPLC method for a faster and reliable measurement of plasmatic catecholamines especially as there is no commercial alternative method.

THE AFIAS 10: A NEW POINT-OF –CARE FOR CTN-T AND ST2

C. Le Goff ¹, E. Brevers¹, M. Massart¹, P. Lukas¹, S. Peeters¹, E. Cavalier¹

¹Department of Clinical Chemistry, University Hospital of Liege, Liege, Belgium

Objectives: There have been concerns regarding results accuracy between point-of-care (POC) testing performed in an emergency department (ED) setting and laboratory testing.This study aimed to analytically validate the AFIAS 10 TroponinT (cTn-T) and Suppression of Tumoreginicity ST2 tests and compare them to routine methods to demonstrate interchangeability.

Materials and methods: We evaluated the AFIAS 10 POC device, distributed by Menarini in Belgium, for cTn-T and ST2 levels using fluorescence immunoassays (FIA). Analytical evaluation included intra and inter-assay variation assessments at three concentration levels. Additionally, we compared the results using 46 and 16 remnant samples for cTn-T and ST2, respectively. The reference method for cTn-T was Roche’s highly sensitive electrochemiluminescence immunoassay (ECLIA) on the Cobas E411. ST2 levels were measured with a ST2 Rapid Test on the POC "Aspect-PLUS" reader (Critical Care). Comparison was conducted using Passing-Bablok regression and Bland-Altman tests, and analytical validation was performed with Enoval (Arlenda).

Results: The AFIAS-10 demonstrated maximum intra and inter-assay CVs of 22.22% and 12.25% for cTn-T and ST2, respectively. For cTn-T across the entire measurement range (n=26, 4-9234 ng/L), the two methods exhibited a good correlation. The regression equation was cTn-T AFIAS10 = -7.27 + 0.91 cTn-T hs Roche (95% CI of the intercept: (-63.3;15.2) and 95% CI of the slope (0.77;0.98)). A minor average proportional difference of 1.2% was observed between the methods. For ST2, the regression equation was ST2 AFIAS10 = -7.17 + 0.77 ST2 Critical Care (95% CI of the intercept: (-11.99;2.31) and 95% CI of the slope (0.66;0.85). A proportional difference of 35.6% was found between the methods. It was explained by a problem of death volume adapted now by the manufacturer. Retrospective analysis of final diagnoses showed concordance for cTn-T in all cases except 1/46 (positive with AFIAS and not with COBAS), but not for ST2: 2/16 were non-concordant. The device is user-friendly from a technical perspective.

Conclusion: POC testing is accurate and correlates well with laboratory testing method. This study demonstrates that POC systems are suitable for promptly assessing ED patients. However, POC devices are costlier, and the cTn-T test lacks of sensitivity compared to the lab test method.

THE MEDCAPTAIN: A POINT-OF –CARE FOR HS-TNI AND NT-PROBNP DETERMINATION

C. Le Goff ¹, E. Brevers¹, P. Lukas¹, G. Pittie¹, S. Peeters¹, E. Cavalier¹

¹Department of Clinical Chemistry, University Hospital of Liege, Liege, Belgium

Objectives: There have been concerns regarding the accuracy of results between point-of-care (POC) tests performed in an emergency department (ED) setting and laboratory testing. The aim of this study was to perform an analytical validation of the Medcaptain Troponin I (hs-TnI) and N-terminal prohormone of brain natriuretic peptide (NT-proBNP) assays and to compare the results with those obtained using routine methods.

Materials and methods: We evaluated a new chemiluminescent immunoassay analyser, the Medcaptain (Analis), for the quantification of hs-TnI and NT-proBNP in human whole blood/serum/plasma. We performed an analytical evaluation with the 3 concentration levels to validate the intra- and inter-assay variation, trueness, measurement uncertainty and we performed a comparison on 23 residual samples for hs-TnI and NT-proBNP. For comparison, the laboratory method for NT-proBNP and hsTnI was a microparticle chemiluminescence immunoassay (CMIA) on the Alinity i analyser. Passing-Bablok regression and Bland-Altman test were used for comparison (Medcalc) and analytical validation was performed with Enoval (Arlenda).

Results: On the Medcaptain device, the intra- and inter-assay CVs obtained were maximum 7% and 8.7% for hsTnI and 11% and 15.74% for NT-proBNP, respectively. The relative bias was maximum 12.8% and 20.43% and the relative expanded uncertainty was maximum 18.9% and 34.84% for hsTnI and NT-proBNP, respectively. The regression equation was NT-proBNP Medcaptain = 104.02+ 1.01NT-proBNP Alinity (95%CI of the intercept: (50.27-280.90) and 95%CI of the slope (0.90-1.06). A small systematic difference of only 2.3% on average is observed between the 2 methods. For hsTnI, the regression equation was: hsTnI Medcaptain = -685.257123 + 1.833114 hsTnI Alinity i (95%CI of the intercept: (-1363.78-215.96) and 95% CI of the slope (1.66-2.07). A proportional difference was found between the 2 methods. The difference is 42.2 %. It is well known that the antibodies used are different between two kits even of hsTnI.

Conclusion: POC testing is accurate and correlates well with laboratory testing methods. However, with troponins there is always a lack of standardisation, which leads to a large difference when we compare the results. What is more important is that the diagnosis is the same. This device can help the emergency physician to quickly identify evidence of cardiac injury and be confident that the results are accurate.

CORRECTION OF TOTAL PROTEIN DETERMINATION IN HEMOLYTIC SAMPLES ON THE COBAS C501 MODULE

A. M. Nguyen¹, G. Mutluoglu ², N. Callewaert¹

¹AZ Groeninge Hospital, Kortrijk, Belgium, ²LKO-LMC Medical Laboratory, Sint-Truiden, Belgium

Objective: Total protein (TP) measurement is a clinical chemistry assay that quantifies protein concentration in both plasma and serum. In the case of haemolytic specimens, an erroneously elevated total protein concentration is observed when employing the biuret method (Lubran, 1978). The laboratory of AZ Groeninge aims to address this haemoglobin interference in TP measurements by introducing a correction factor. This is based on specific calculations, in which the haemoglobin concentration is multiplied by the slope, obtained through the linear regression between the bias of the analyt and the haemoglobin concentration in plasma or serum (Lippi et al., 2008). If the slope remains unaffected by the total protein concentration and haemolysate, a reliable correction factor can be determined.

Methods: We investigated the impact of TP concentration and haemolysate on the slope and described the experimental process for determining the correction factor. This investigation involved two experiments using the same setup. The setup involved creating dilution series of ’blank’ serum with increasing haemoglobin concentrations. In experiment 1, dilution series were prepared for three serum samples with different TP concentrations (low, normal, high) using the same haemolysate. In experiment 2, dilution series were created for three serum samples with normal TP concentrations, each employing a different haemolysate (A,B & C). The results of each experiment were assessed using Pearson correlation and linear regression.

Results: Strong linear correlation between free haemoglobin concentration (mg/dL) and falsely elevated TP (g/L) was observed in all dilution series (P < .001). The slope and its corresponding 95% confidence interval from each dilution series were of a similar magnitude, except for the high TP concentration and haemolysate C. Despite variations in slope in both experiments, the slope remained quite constant, allowing for the calculation of an average slope based on the obtained values from both experiments. The average slope was determined to be 1,488. To assess its reliability as a correction factor, the %bias was calculated after correction and compared to the %biasmin criteria from the EFLM database (The European Federation of Clinical Chemistry and Laboratory Medicine Biological Variation Database). The %bias slightly exceeded the criteria for only two of the 40 data points.

Conclusion: It can be concluded that the TP concentration and haemolysate have impact on the slope, although the observed effect is practically negligible. Based on the criteria provided by the EFLM database, the determined slope can correct the haemoglobin interference. The correction factor can be implemented on the Cobas c501-analyzer, thereby allowing a more accurate representation of TP in haemolytic specimens.

References

Lubran MM. The measurement of total serum proteins by the Biuret method. Ann Clin Lab Sci. 1978 Mar-Apr;8(2):106-10. PMID: 345944.

Lippi G, Blanckaert N, Bonini P, Green S, Kitchen S, Palicka V, Vassault AJ, Plebani M. Haemolysis: an overview of the leading cause of unsuitable specimens in clinical laboratories. Clin Chem Lab Med. 2008;46(6):764-72. doi: 10.1515/CCLM.2008.170. PMID: 18601596.

DISCRIMINATING SIGNAL FROM NOISE: THE BIOLOGICAL VARIATION OF CIRCULATING CALPROTECTIN

M. Briers ^1,2, B. Vander Cruyssen³, B. Massa^1,2, S. Van Den Bremt², L. Hofman², A. Belmans⁴, B. Hoermann⁵, X. Bossuyt^1,6, L. Van Hoovels^2,6

¹Department of Laboratory Medicine, University Hospital Leuven, Leuven, Belgium, ²Department of Laboratory Medicine, OLV Hospital, Aalst, Belgium, ³Department of Rheumatology, OLV Hospital, Aalst, Belgium, Aalst, Belgium, ⁴I-BioStat, KU Leuven, Leuven, Belgium, ⁵Thermo Fisher Scientific, Freiburg, Germany, ⁶Department of Microbiology, Immunology and Transplantation, KU Leuven, Leuven, Belgium

Objectives: Circulating calprotectin (cCLP) is a promising systemic biomarker of disease activity in neutrophil-related inflammation. Our study aimed to define the within-subject (CV_I) and between-subject (CV_G) biological variation (BV) of cCLP in healthy individuals (HI).

Materials and methods: Every HI included in the study, voluntarily underwent 10 blood collections, equally spread over 5 consecutive weeks. Each collection comprised a serum, citrate and EDTA sample. Pre-analytical requirements were strictly addressed.

Besides C-reactive protein analysis, cCLP measurement was performed with the EliA Calprotectin 2 assay (Thermo Fisher Scientific, Freiburg, Germany) on the Phadia 2500 instrument, using a serum/plasma protocol for research use only.

Outlier detection was performed on the CV-transformed data using the Cochran’s method. After correction for trends over time, the estimated geometric mean values, CV_I and CV_G were calculated for each sample type. Impact of gender and age on estimated geometric mean cCLP values, CV_I and CV_G was also evaluated.

Results: In total, 55 HIs were included in the study (58.2% female; median age [range] = 39 [21-62] years). After applying the exclusion criteria, 538 serum samples, 534 EDTA plasma samples and 534 citrate plasma samples were statistically analyzed.

cCLP concentrations measured in serum were more than triple the concentration in EDTA and citrate plasma (1.837 µg/mL versus 0.366 µg/mL and 0.471 µg/mL respectively). Furthermore, cCLP levels obtained from male HI were higher than for female HI (p<0.05 on citrate and EDTA plasma).

cCLP CV_I [95%CI] were substantially lower in citrate and EDTA plasma samples (resp. 28.99 [27.25-30.74]% and 23.81 [22.38-25.24]%) than in serum samples (38.60 [36.29-40.91]%). CV_G tended to be lower for cCLP levels in citrate plasma (36.54 [28.60-44.47]%) compared to EDTA plasma (45.77 [35.85-55.70]%) and serum (41.27 [31.76-50.77]%). In contrast to geometric mean values, CV_G was not gender dependent (p>0.05 all matrices). Impact of age on cCLP CV_I or CV_G couldn’t be revealed either.

Conclusion: cCLP CV_I and CV_G results are essential in accurately interpreting cCLP concentrations in the follow-up of neutrophil-related inflammatory diseases. The highest values for cCLP CV_I were observed in serum samples, whereas the lowest values for cCLP CV_G were observed in citrate plasma samples.

ANALYTICAL PERFORMANCE AND USER-FRIENDLINESS OF FOUR COMMERCIALLY AVAILABLE POINT-OF-CARE TESTS FOR C-REACTIVE PROTEIN

L. Van Hoovels ^1,2, B. Massa ^1,3, H. De Meyer¹, P. De Schrijver¹, V. Van Laethem¹, D. Barglazan⁴, D. Gruson⁵, R. Hopstaken⁶, B. Peeters⁷, V. Van Hoof⁸, A. Verdonck³, J. Verbakel^9,10

¹Department of Laboratory Medicine, OLV Hospital Aalst, Aalst, Belgium, ²Department of Microbiology, Immunology and Transplantation, KU Leuven, Leuven, Belgium, ³Department of Laboratory Medicine, University Hospital Leuven, Leuven, Belgium, ⁴Laboratoire Hospitalier Universitaire de Bruxelles, Universitair Laboratorium Brussel (LHUB-ULB), Brussels, Belgium, ⁵Department of Medical Biochemistry, Clinique Saint-Luc, UCLouvain, Woluwe-Saint-Lambert, Belgium, ⁶Star-shl Diagnostic Centers, 3068 Rotterdam, Netherlands, ⁷Department of Laboratory Medicine, Heilig Hart Hospital Lier, Lier, Belgium, ⁸Faculty of Medicine and Health Sciences, University of Antwerp, Antwerp, Belgium, ⁹EPI-Centre, Academisch Centrum Huisartsgeneeskunde, KU Leuven, Leuven, Belgium, ¹⁰NIHR Community Healthcare Medtech and IVD cooperative, Nuffield Department of Primary Care Health Sciences, University of Oxford, Oxford, United Kingdom

Objective: The analytical performance and user-friendliness of four quantitative POCT-CRP assays was evaluated: QuikRead go easy CRP (Aidian, Espoo, Finland), LumiraDx CRP (LumiraDx, Stirling, UK), Cobas b101 (Roche Diagnostics, Mannheim, Germany) and Afinion 2 (Abbott, Oslo, Norway).

Materials and methods: Imprecision was evaluated using patient lithium heparin plasma pools and manufacturer specific internal quality control material. Accuracy was assessed using external quality control material (Sciensano and UKNEQAS) and method comparison towards the central laboratory CRP analysis method (Tina-Quant C-Reactive protein IV, Cobas c503, Roche Diagnostics) using 100 retrospective selected routine lithium heparin plasma samples and 50 prospectively collected capillary blood samples. User-friendliness of all assays was assessed using a questionnaire.

Results: Between-day coefficients of variation varied from 1.7% (Afinion 2) to 8.9% (QuikRead go) for internal quality controls and from 4.5% (LumiraDx) to 11.5% (QuikRead go) for patient pools. Cobas b101 and Afinion 2 achieved the best analytical agreement with the central laboratory method. LumiraDx and QuikRead go revealed a negative mean difference towards the central laboratory method, with LumiraDx violating the performance criterion of a mean difference of ≤20% for lithium heparin plasma samples with CRP-values <60 mg/L and both external quality control materials. All instruments were considered user-friendly with the highest Likert-scores obtained for Afinion 2.

Conclusion: The analytical performance and user-friendliness of POCT-CRP devices varies among manufacturers emphasizing the need for quality assurance supervised by a central laboratory.

Acknowledgement: We thank the manufacturers Abbott, LumiraDx, Aidian and Roche Diagnostics for the in-kind reagents provision and the technical training. We acknowledge UKNEQAS and Sciensano for the cooperation in using their external quality control samples and results.

CIRCULATING CALPROTECTIN IN RHEUMATOID ARTHRITIS: UNRAVELLING THE IMPACT OF (PRE-)ANALYTICAL CONFOUNDERS ON META-ANALYSIS RESULTS

B. Massa ^1,2, M. Mylemans ^1,2, B. Vander Cruyssen³, M. Infantino⁴, X. Bossuyt^2,5, L. Van Hoovels^1,5

¹Department of Laboratory Medicine, OLV Hospital, Aalst, Belgium, ²Department of Laboratory Medicine, University Hospital Leuven, Leuven, Belgium, ³Department of Rheumatology, OLV Hospital, Aalst, Belgium, ⁴Immunology and Allergy Laboratory Unit, San Giovanni di Dio Hospital, Florence, Italy, ⁵Department of Microbiology, Immunology and Transplantation, KU Leuven, Leuven, Belgium

Introduction: Recent meta-analyses, investigating the clinical evidence of cCLP in RA patients versus healthy individuals (HI) described an important inter-study heterogeneity. We investigated the impact of pre-analytical on the inter-study heterogeneity of available publications from January 2000-February 2023.

Materials and methods: A systematic literature search was performed in the electronic databases Medline, Embase and Cochrane databases (March 23^st, 2023). MEDCALC statistical software version 19.3 (MedCalc Software ltd, Ostend, Belgium) was applied for meta-analysis.

Results: Our literature search resulted in a total of 646 titles: 44 titles were retrieved for full-text screening and 18 studies including 21 cCLP datasets were eligible for inclusion in the meta-analysis. Totally, cCLP was measured in 2627 RA and 897 HI and cCLP levels were significantly higher in RA patients and significant (p<0.0001) heterogeneity was observed between the different studies (Q=171.04; I²=88.3%; 95%CI=83.5-91.7%).

cCLP analyses was tested using 14 different cCLP assays. Only in 9 out of the 18 included studies, adherence to pre-analytical conditions (i.e. time to centrifugation and storage temperature) was specified. Subgroup analysis of the datasets (n=12) of these 9 studies still showed significant heterogeneity (Q=96.107; I²=88.6%; 95%CI=81.9-92.8%), mainly caused by the use of Bühlmann cCLP assay (n=5). Excluding Bühlmann assays: datasets compliant to pre-analytical recommendations (n=8), using 8 different assays on samples of 715 RA patients and 284 HI, showed absence of heterogeneity (Q=1.31; I² =0.0; 95%CI=0.0-0.0), whilst in the other subgroup (n=8) significant inter-study heterogeneity persisted (Q=66.0; I² =89.40%; 95%CI= 81.48-93.93).

Conclusion: The inter-study variability of cCLP in RA is, besides analytical issues, mainly caused by pre-analytical confounders. Adherence to pre-analytical recommendations significantly reduced inter-study heterogeneity and provided a reliable SMD, accurately revealing higher cCLP levels in RA versus HI.

METHOD COMPARISON OF TWO CALPROTECTIN IMUNOASSAYS: THE ELIA™ CALPROTECTIN 2 TEST (THERMO FISHER SCIENTIFIC) AND CALIAGOLD® (SENTINEL DIAGNOSTICS)

E. Vermeulen ¹, L. Pouillon², S. van Aelst³, A. Sierens⁴, M. van Blerk⁵, I. Geerts¹, P. Bossuyt²

¹Department of Laboratory Medicine, Imelda Hospital, Bonheiden, Belgium, ²Department of Gastroenterology, Imelda Hospital, Bonheiden, Belgium, ³Department of Laboratory Medicine, Heilig Hartziekenhuis, Lier, Belgium, ⁴Department of Laboratory Medicine, AZ Sint-Maarten, Mechelen, Belgium, ⁵Department of Laboratory Medicine, AZ Jan Portaels, Vilvoorde, Belgium

Introduction: Calprotectin is an important biomarker in the diagnosis and follow-up of IBD patients. Calprotectin immunoassays are not harmonised - despite the common cut-off of 50 mg/kg. Tight control of inflammatory bowel disease (IBD) patients also requires a good knowledge of assay performance in a higher range, especially 100-300 mg/kg^1,2.

Aim: Our aim was to evaluate the performance of the CALiaGold® (Sentinel Diagnostics) calprotectin immunoassay, compared to the established method in our laboratory, the EliA™ Calprotectin 2 test (Thermo Fisher Scientific) including the respective calprotectin extraction methods.

Methods: Study population and design

At the clinical laboratory of Imelda Hospital, which performs the calprotectin assays for four secondary-care hospitals and which has a large IBD-department, routine samples were prospectively included for 2 weeks, resulting in 113 samples.

Preanalytical aspects and extraction methods

Stool samples were stored at -20°C until batch extraction (3 batch extractions/week). Extracts for the reference method were prepared with a weighing-based method (Fecal sample preparation kit®; Roche Diagnostics). Extracts for the CALiaGold® assay were prepared with the CALiaGold® Tube extraction tool (a volume-based dipstick method) as prescribed by the manufacturer. The extract for the study method was prepared at the same time as the standard method, on residual stool material.

Analytical methods

Both the extracts per sample were analyzed with the reference method and the study method on the same day, to minimize the effect of differences in storage conditions on the analytical result. The reference method is the EliA™ Calprotectin 2 test (Thermo Fisher Scientific), performed on the Phadia 250 analyzer (Thermo Fisher Scientific). The study method is CALiaGold® (Sentinel Diagnostics) on the SENTiFIT®270 analyzer (Sentinel Diagnostics). Samples with CALiaGold® were measured in duplicate.

Results: Analytical performance

The duplicate CALiaGold® calprotectin measurements (performed on one CALiaGold® extract) have a good coefficient of variation of 3.96% (calculated for 78 samples with a result in the measuring range of 20-2200 mg/kg).

Concordance at the cut-off of 50 mg/kg

From 113 included samples, 51 samples were considered negative with the cut-off of 50 mg/kg with both the CALiaGold® method and the EliA™ Calprotectin 2 test. 46 samples were considered positive (>50 mg/kg) with both methods. However 16/113 samples (14%) had a different qualitative interpretation (for both methods, 8 samples positive classified negative with the alternative method).

Method comparison

When calculated for the subset of samples with a result in the measuring range for both methods (n=77), the passing bablok fit is y = 20.05 + 0,5549x (x is the reference method). The Pearson correlation coefficient is 0.80 (95%CI 0.70-0.87).

The subset of samples with a result in the clinical relevant range for follow-up (defined as 100-300 mg/kg with the reference method) consists of 19 samples. For this subset, with the EliA™ Calprotectin 2 result (Thermo Fisher Scientific) considered as the reference method, only 10/19 samples measured with CALiaGold® have a bias between -50% and 50% and bias is ranging from -89% up to +363%. The mean total analytical error (TEA%) is 75.1%.

Preanalytical aspects

When the CALiaGold® extracts of a challenging selection of 10 discordant samples are analyzed with Calprotectin 2 test (Thermo Fisher Scientific), performed on the Phadia 250 analyzer, better agreement is obtained, suggesting that the extraction is probably not the most critical factor in the differences obtained, but the analytical method itself.

Conclusions: The concordance in qualitative interpretation of calprotectin at the cut-off of 50 mg/kg is acceptable for CALiaGold® (Sentinel Diagnostics) compared to the EliA™ Calprotectin 2 test (Thermo Fisher Scientific). However, large differences are observed in the quantitative results: methods are not interchangeable in the follow-up of IBD patients. Method validation data communicated to clinicians should not only focus on qualitative interpretation at the cut-off of 50 mg/kg.

References: ¹Colombel JF, et al. Effect of tight control management on Crohn’s disease (CALM): a multicentre, randomised, controlled phase 3 trial. Lancet 2017; 390: 2779–89

²Costa F, et al. Calprotectin is a stronger predictive marker of relapse in ulcerative colitis than in Crohn’s disease. Gut 2005;54:364–368

THE COMPARISON AND OPTIMIZATION OF CALPROTECTIN EXTRACTION IN ROUTINE IBD CARE

E. Vermeulen ¹, L. Pouillon², B. Possemiers¹, A. Goovaerts¹, I. Geerts¹, P. Bossuyt²

¹Department of Laboratory Medicine, Imelda Hospital, Bonheiden, Belgium, ²Department of Gastroenterology, Imelda Hospital, Bonheiden, Belgium

Introduction: The preanalytical phase contributes to variability in calprotectin measurements between laboratories while tight control of patients with IBD relies on accurate and precise calprotectin results. The ’weighing-based’ extraction method is considered to be more accurate for calprotectin than ‘volume-based’ methods. However, weighing-based methods are challenging in routine practice. This procedure is laborintensive and poses a risk for identification errors and other preanalytical concerns in a batch extraction setting - such as prolonged sample storage and freeze-thaw cycles. The consolidation of calprotectin samples from four hospitals in our laboratory network urged us to re-evaluate the preanalytical phase of calprotectin measurement.

Aim: Our aim was to optimize the preanalytical phase in calprotectin analysis, by comparison of three different calprotectin extraction methods combined with a single analytical method, the EliA™ Calprotectin 2 test (Thermo Fisher Scientific).

Methods: From February 2022 onwards, 50 prospective calprotectin samples from the gastroenterology department in our hospital underwent simultaneous calprotectin extraction with 3 methods. The reference method (the current extraction method in our lab), a weighing-based method (Fecal sample preparation kit®; Roche Diagnostics) was compared with two volume-based (dipstick) methods (CALiaGold® pierceTube; Sentinel diagnostics, and Elia™ Stool Extraction Kit plus; Thermo Fischer Scientific). The extracts were measured simultaneously to eliminate difference in sample or extract storage and to minimize analytical (interrun) variability.

Results: Both the volume-based extraction methods correlate well with the reference method (Pearson’s R>0.98). At the diagnostic cut-off of 50 mg/kg, there is a good concordance in terms of positive or negative results, between the extraction methods. Of 26 negative (calprotectin <50 mg/kg) samples obtained with the reference method (Roche), 25/26 samples were also negative with the alternative extraction methods. The discordant samples had low positive values: 54 mg/kg for the discordant Caliagold® extract, compared to 35 mg/kg with the reference method and 42 mg/kg with Elia™ Stool Extraction Kit plus; the discordant Elia™ Stool Extraction Kit plus extract measured 72 mg/kg compared to 41 mg/kg with the reference method and 34 mg/kg with Caliagold extraction.

In 24 positive samples with a calprotectin result >50 mg/kg with the reference method, the concordance for positivity is 23/24 for Elia™ Stool Extraction Kit plus and 21/24 for the Caliagold extraction. All discordances – gaining a negative result with the alternative extraction methods - were borderline results with the reference method (ranging 56-65 mg/kg).

In positive samples >50 mg/kg, the Caliagold extraction gained lower results with a median negative bias of -28.4% (95%CI -45.7 to 0.0), even ranging to -74.3% (267 mg/kg compared to 1040 mg/kg with the reference method) in one sample. The Elia™ Stool Extraction Kit plus showed a median bias of 11.1% (95%CI: -24% to 40%) and less variability – especially in the clinically important range of 100-300 mg/kg, with the exception of some distinct discordant results. However, when these discordant samples were repeated with a good homogenisation of the samples before repeating both extractions, the Elia™ Stool Extraction Kit plus (Thermo Fisher Scientific) and the Fecal sample preparation kit® (Roche) gained comparable results and even confirmed the result obtained with the Elia™ Stool Extraction Kit plus.

Conclusions: The ‘volume-based’ (or ‘dipstick’) calprotectin extraction method Elia™ Stool Extraction Kit plus (Thermo Fischer Scientific) was judged clinically equivalent to the ‘weighing-based’ gold standard Fecal sample preparation kit® (Roche Diagnostics) in combination with the EliA™ Calprotectin 2 test, while another ‘volume-based’ calprotectin extraction method (CALiaGold® pierceTube; Sentinel diagnostics) was not. Since the Elia™ Stool Extraction Kit plus (Thermo Fischer Scientific) can be performed daily - excluding the need to freeze samples, has a lower risk for identification errors and facilitates a good homogenization of the faecal sample, it might be the preferred technique in a large-volume clinical lab.

DE NOVO DIAGNOSIS OF TYPE I HYPERLIPIDEMIA IN AN INFANT

D. Oulkadi ¹, E. Huyghe¹, C. van der Heijden¹, M. Cuykx¹

¹Department of Laboratory medicine, Antwerp University Hospital, Antwerp, Belgium

Objective: To describe a case of a de novo diagnosis of type I hyperlipidemia in an infant, triggered by a pinkish looking serum that caused severe lipemia interference.

Case presentation and methods: A 6-week-old girl born at 39 weeks was referred to the emergency department of the Antwerp University Hospital because of viscous, rosy serum. Symptoms included coughing, fever and vomiting.

During blood sampling, high viscosity and a pinkish appearance of the sample was noticed. With exception of the lipids, biochemistry tests could only be performed after high speed centrifugation (10g for 15 minutes) with removal of the fatty supernatant on the Atellica Solutions (Siemens Healthineers). Hemoglobin (Hb) testing on the Sysmex XN 9100 (Sysmex) failed at first attempt. After centrifugation on 3500 rpm for 10 minutes with replacement of the plasma with Cellpack DCL, reliable results for Hb, hematocrit and MCHC were obtained.

Results: Biochemical results included significantly elevated triglycerides up to 37 268 (< 150 mg/dl) and total cholesterol up to 2108 mg/dL (< 170 mg/dl). Markedly elevated lipase up to 656 U/L (12 – 53 U/L) and elevated total bilirubin up to 3.3 mg/dL (0.3 – 1.2 mg/dl), in combination with widened bile ducts and enlarged gallbladder on ultrasound imaging of the abdomen led to a diagnosis of incipient secondary pancreatitis consequential to an underlying familial hyperchylomicronemia syndrome. The lipidogram showed predominant presence of pre-beta (VLDL) and chylomicrons in serum and lipid fraction, which is consistent with a type I or type V hyperlipidemia. Genetic analysis revealed a homozygote mutation in the LPL gene which causes type I hyperlipidemia.

A lipid-free diet was initiated with a favorable clinical and biochemical evolution (last measured triglycerides and cholesterol concentrations of 387 mg/dL and 321 mg/dL respectively).

Conclusion: We described a case of type I hyperlipidemia in an infant with a clinical presentation of pancreatitis and viscous pinkish blood on blood sampling. With exception of the lipids, biochemistry tests could only be performed after high speed centrifugation. Hb could only be measured after centrifugation on 3500 rpm for 10 minutes with replacement of the plasma with Cellpack DCL.

MARTIN-HOPKINS EQUATION A SUPERIOR ALTERNATIVE TO THE TRADITIONAL FRIEDEWALD FORMULA FOR THE ESTIMATION OF LDL-CHOLESTEROL: COMPARISON WITH A DIRECT LDL ASSAY

B. Vanmechelen ¹, E. Vermeulen¹, P.-J. Briers², M. Langlois², C. Vercammen¹, B. Vets¹, I. Geerts¹

¹Imelda, Bonheiden, Belgium, ²AZ Sint-Jan, Brugge, Belgium

Objective: To compare the concordance of the Martin-Hopkins, Sampson and Friedewald equations for calculated LDL- with direct LDL-cholesterol (LDL-D).

Material and Methods: Lipid profiles extracted between January 2016 to December 2022 from the laboratory information system of the clinical laboratory of Imelda Hospital. Triglycerides and HDL-, LDL- and total cholesterol concentrations were determined by colorimetric enzymatic reactions using Roche Cobas 6000 c501. Data from patient samples with triglycerides >400 mg/dL were excluded. Analytical performance was evaluated between the Martin-Hopkins (LDL-MH), Sampson (LDL-S), Friedewald (LDL-F) equations and LDL-D through Pearson correlation, Bland-Altman and concordance analysis based on the ESC/EAS guidelines 2019 risk thresholds.

Results: A total of 22 661 samples were included. All formulae showed a good linear relationship against LDL-D. The correlation coefficients were significantly higher with LDL-MH and LDL-S formulae in the different subgroups. In general, higher LDL-cholesterol concentrations were obtained using the LDL-MH and LDL-S equations compared to LDL-F. Of the three formulae, LDL-MH showed the lowest bias vs. LDL-D. In the subgroup of LDL-D <70 mg/dL (N=4566) a mean difference of 3.8 mg/dL was observed. LDL-MH was also found to result in the lowest number of discordant risk classifications based on a LDL-D threshold of 70 mg/dL. Thirty percent of samples with LDL-F <70 mg/dL had LDL-cholesterol concordantly reclassified to >70 mg/dL by the Martin-Hopkins equation.

Conclusion: The Martin-Hopkins equation is more concordant with Roche LDL-D than the Friedewald equation. The implementation could improve the identification of patients who may benefit from more intensive LDL lowering treatment without an extra cost.

CLINICAL VALIDATION OF CAPILLARY SAMPLING USING VAMS FOR APPLICATION IN POPULATION STUDIES ON PFAS

L. Delahaye¹, D. Bernard ², T. Demoor¹

¹Eurofins Forensics, Brugge, Belgium, ²Eurofins Clinical Diagnotics Kortrijk, Kortrijk, Belgium

Objective: In this study PFAS concentrations determined on capillary finger prick samples, using a volumetric absorptive micro sampling device (VAMS) are compared with serum samples (reference), for 19 compounds. The aim was to examine the feasibility of widespread use of capillary samples in a blood study on PFAS, as capillary sampling is expected to improve participation in specific subpopulations, such as children. Data analysis focused on determining a conversion factor to convert capillary blood sample concentrations into serum equivalent concentrations.

Materials and methods: Capillary blood samples were collected by Mitra® VAMS, Neoteryx. Venous serum was collected in BD Vacutainer SST II Advance Plus tubes. Both sample types were analysed with a validated LC-MSMS method, with limit of quantitation (LOQ) 0,1 ng/mL for all compounds.

Samples were obtained from participants of the large scale blood study on PFAS contamination in the population residing <5 KM area around the 3M-plant in Zwijndrecht, Belgium. A separated informed consent was obtained for this substudy. Statistical analysis was performed using MedCalc®, MedCalc Software Ltd.

Results and conclusions: Paired samples from 293 males and 310 females were obtained.

Unfortunately, insufficient sample pairs (N<40) were observed with concentrations>LOQ for PFBA, PFPeA, PFHxA, PFHpA, PFOA branched, PFDoDA, PFBS, PFHxS branched.

Following conversion factors could be determined: PFOA lin: 2,12; PFOA total: 2,12; PFNA: 2,12; PFDA: 1,83; PFHxS lin: 2,04; PFHxS tot: 2,04; PFHPS: 2,20; PFOS lin: 1,90; PFOS branched: 2,06; PFOS tot: 1,90. For PFUnDA a factor of 1,54 is proposed as more data are needed for a smaller confidence interval. For total PFOA and total PFOS the utility of gender specific conversion factors is under study.

Disclosures: This abstract is based on research commissioned and funded by the Department of Care of the Flemish government. The views expressed herein are those of the authors and are not necessarily endorsed by the Flemish government. The study was approved by the Ethical Committee of the UIA.

GENOTYPIC DIAGNOSIS OF PORPHYRIA IN BELGIUM

D. C. Nyamessameye ¹, M. Delaunoy¹, A. Gilles¹, F. Cotton¹

¹LHUB-ULB/Department of Clinical Chemistry/Université Libre de Bruxelles, Bruxelles, Belgium

Objective: Porphyrias are rare but serious metabolic disorders caused by enzymatic defects in the heme biosynthesis pathway. The diagnosis of symptomatic individuals is based on specific biochemical profiles of porphyrins and their precursors. The family screening by molecular analysis of genetic mutations is essential for identifying asymptomatic individuals. Since 2019, the Centre Belge des Porphyries (CBP) (ULB Genetics Center – Erasme Hospital – LHUB-ULB) has been providing genotypic diagnosis of these diseases, and in 2022 it was approved by the European Porphyria Network (EPNET) as a Porphyria Expert Clinical Centre (PECC). In order to determine the profile of porphyrias in Belgium, we aim to make an inventory of genotypic diagnoses performed by the centre and to correlate the genotype with the phenotype of each case.

Methods: We selected all patients for whom samples were sent to the laboratory between 1 January 2020 and 17 January 2023 for porphyria genotyping. Patient data were collected from their medical and laboratory records, when available.

Results: Among the 108 requests sent to the CBP, 35 cases of porphyrias were confirmed, including 12 cases of acute intermittent porphyria (11 phenotype-genotype agreement and 1 discordance); 12 cases of erythropoietic protoporphyria (11 agreements and 1 weak discordance); 6 cases of variegate porphyria (4 agreements and 2 discordances), 3 cases porphyria cutanea tarda (PCT) and 2 cases of hereditary coproporphyria (all in agreement with the phenotype). These results are consistent with the respective prevalence of porphyrias, with an expected under-representation of PCT given its moderate severity. All patients diagnosed were heterozygous for the mutation identified. No mutation was identified in 54 cases. Five mutations never described in the literature were reported: HMBS gene (c.718_721dup, c.799G>C, c.246delA), UROD gene (c.847G>C) and CPOX gene (c.218G>A).

Conclusion: The different porphyrias diagnosed at the CBP at the end of this study are representative of the porphyrias epidemiology in Belgium. However, this study presents certain biases such as the non-availability of certain biochemical test results and the absence of samples allowing these tests to be performed at the CBP in order to confirm the patient’s phenotype. A study with stricter selection criteria and without bias should be considered.

INTERVENTIONS TO IMPROVE APPROPRIATENESS OF LABORATORY TESTING IN THE INTENSIVE CARE UNIT

L. Devis ¹, E. Catry¹, P. M. Honore¹, A. Mansour², G. Lippi³, F. Mullier¹, M. Closset¹

¹CHU UCL Namur, Yvoir, Belgium, ²Pontchaillou University Hospital of Rennes, Rennes, France, ³University Hospital of Verona, Verona, Italy

Healthcare expenses are increasing, as is the utilization of laboratory resources. Despite this, between 20% and 40% of requested tests are deemed inappropriate. Improper use of laboratory resources leads to unwanted consequences such as hospital-acquired anemia, infections, increased costs, staff workload and patient stress and discomfort. In this context, several interventions have been carried out to improve the appropriateness of laboratory testing. To date, there have been few published assessments of interventions specific to the intensive care unit. We reviewed the literature for interventions implemented in the ICU to improve the appropriateness of laboratory testing. We searched literature from 2008 to 2023 in PubMed, Embase, Scopus, and Google Scholar databases between April and June 2023. Six intervention categories were identified: education, guidance, audit and feedback, gatekeeping, computerized physician order entry, and multifaceted interventions (MFI). We included a seventh category exploring the potential role of artificial intelligence-based assisting tools in such interventions. Education-based interventions and MFI are the most frequently used approaches. MFI is the most effective type of intervention, and shows the strongest persistence of effect over time. AI-based tools may offer valuable assistance to the improvement of appropriate laboratory testing in the near future. Patient safety outcomes are not impaired by interventions to reduce inappropriate testing. The literature focuses mainly on reducing overuse of laboratory tests, with only one intervention mentioning underuse. We highlight an overall poor quality of methodological design and reporting and argue for standardization of intervention methods. Collaboration between clinicians and laboratory staff is key to improve appropriate laboratory utilization.

AN AI-DRIVEN SUPPORT TOOL FOR PREDICTION OF URINE CULTURE TEST RESULTS

L. Dedeene ¹, J. van Elslande¹, J. Dewitte¹, G. Martens¹, E. De Laere¹, D. De Smet¹

¹Department of Laboratory Medicine, AZ Delta General Hospital, Roeselare, Belgium, Roeselare, Belgium

Objective: We aimed to create a readily deployable artificial intelligence (AI)-driven model to rapidly predict urine culture test results.

Material and Methods: The training dataset encompassed urine samples collected from five hospital sites over the course of one year (n=34,584 samples). A time-separated, unseen test set of urine samples (n=10,083) was used to estimate the out-of sample performance of the prediction models. Urine samples were classified as “positive” if at least one uropathogen was present counting ≥10⁴ colony forming units/mL without overgrowth of urogenital flora. We compared the diagnostic performance of various machine learning models, including multiple logistic regression (MLR), Random Forest classifier (RF), Extreme Gradient Boosting algorithm (XGBoost), K-nearest neighbours algorithm (KNN), multilayer perceptron (MLP), and ensemble voting classifier combining all aforementioned models. Predictive parameters integrated into the models included readily accessible urinalysis results (from the UF-4000 analyzer and dipstick), patient demographics (age and gender), and collection method. To provide interpretable results, we categorized probability scores based on increasing likelihood ratios (LR) for positive urine culture results.

Results: While the more complex models yielded the highest AUCs for prediction of positive cultures (highest: MLP with AUC of 0.884, 95% CI 0.878-0.89), a very good AUC (0.858, 95% CI 0.852-0.865) was reached by MLR using just the six cell count parameters generated by the UF-4000 (WBC, RBC, bacteria, yeast, epithelial and tubular cells). As this model is easily programmable in the laboratory information system, this MLR model was chosen to further categorize the prediction results into five categories based on the LR for positivity: highly unlikely (LR 0.1), unlikely (LR 0.3), grey zone (LR 0.9), likely (LR 5), and very likely (LR 40). When used on the test set, this resulted in 17%, 28%, 32%, 10% and 13% of samples in each respective category. In the highly unlikely and unlikely group, the negative predictive value was 95.2% and 86.7%, respectively. In the highly likely and likely group, the positive predictive value was 95.4% and 72.0% respectively.

Conclusions: This tool has the potential to quickly (0.5-1h) assist clinicians in their clinical decision-making process by offering insights prior to the availability of urine culture results (turnaround time of typically 1-2 days) in a significant portion of samples (68%). Additionally, in the clinical laboratory, it allows for the prioritization of the interpretation of urine cultures with a high likelihood of being positive. This approach may facilitate an activity-driven deployment of laboratory technicians. It is important to note that, to ensure the detection of significant growth in vulnerable patient groups and maintain applicability across the entire hospital, all urine samples will continue to undergo culture.

EVALUATION OF THE PERFORMANCE OF TWO RAPID TESTS FOR THE DOSAGE OF C-REACTIVE PROTEIN

P. Mbouche ¹, N. Riahi¹, M. Tré-Hardy¹, R. Cupaiolo¹, L. Blairon¹

¹Iris Hospitals South, Brussels, Belgium

Objective: C-reactive protein (CRP) is a widely used marker of inflammation and infection. The use of point-of-care testing (POCT) methods for CRP measurement provides rapid and relevant results enabling clinicians to make appropriate clinical decisions and manage their patients accordingly. The aim of this study is to evaluate the performance of two POCT devices, LumiraDx (LumiraDx UK Ltd, Tullibody, UK) and QuikReadGo (Aidian Oy, Espoo, Finland) for the measurement of CRP.

Material and Methods: For each level of low and high internal quality control (IQC), repeatability was performed over 20 runs on the same day for both tests, and intermediate precision was performed twice a day for 10 days for LumiraDx and 4 times a day for 20 days for QuikReadGo. The coefficient of variation was also determined. Method comparison was performed for each assay with the Roche cobas 8000 analyzer as the gold standard, using 51 and 200 patient whole blood samples for LumiraDx and QuikReadGo respectively. Passing-Balock regression and Bland-Altman were used for the correlation.

Results: The table 1 summarize the main results.

Conclusion: Our results indicate that LumiraDx and QuikReadGo have good accuracy and high correlation with the reference method for CRP measurement, and can be used as a rapid test to guide clinicians in their clinical and therapeutic decisions.

EVALUATION OF THE ANALYTICAL PERFORMANCES OF THE ARK PREGABALIN II ASSAY

P. Mbouche ¹, F. Ponthieux¹, F. Cotton¹

¹LHUB-ULB, Brussels, Belgium

Objectives: Pregabalin is an analog of gamma-aminobutyric acid (GABA), the predominant inhibitory neurotransmitter. It is used in the treatment of epilepsy, peripheral neuropathic pain and anxiety. There are growing concerns about the misuse and/or abuse of pregabalin. Our objective was to perform an analytical verification of the ARK Pregabalin II Assay on a Cobas c 502 analyzer in comparison with a reference method on urine samples.

Material and Methods: For each low and high internal quality control (IQC) level, repeatability was assessed on 10 runs on the same day, and intermediate precision twice a day for 10 days. Linearity was verified by serial dilutions of the highest calibrator (6 points in triplicate). The lower limit of quantification (LLOQ) was estimated the same way with the low IQC level. The European Medicines Agency (EMEA) performance characteristics were used with an acceptable coefficient of variation (CV) ≤ 20%. A comparison of qualitative results was performed with liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) (MassTox® Drugs of Abuse Testing in Urine) on 61 samples positively screened by LC-MS/MS (Chromsystems Diagnostics, Germany).

Results: For low and high IQC levels, the repeatability was 4.6% and 3.9%, and the intermediate precision was 6.0% and 2.8%, respectively. The linearity range extended from 31 to 2000 ng/mL. A discordance with LC-MS/MS was observed for 4 samples (6.5%), giving a sensitivity of 87% and a specificity of 100%. The Cohen’s Kappa concordance coefficient (95%CI) was 0.87 (0.75 to 0.99).

Conclusion: The ARK Pregabalin II Assay meets EMEA acceptance criteria and displays an acceptable agreement with LC-MS/MS. This semi-quantitative automated assay can be routinely implemented as a screening assay, allowing a short turnaround time that could help improving the patient care in a toxicological context.

APPLICATION OF ANALYTICAL PERFORMANCE SPECIFICATIONS FOR URINE TEST STRIP METHODS: IMPORTANCE OF REFLECTANCE SIGNALS

J. Cabo ¹, J. Favresse¹

¹Clinique Saint-Luc Bouge, Bouge, Belgium

Objective: Urinalysis is essential for diagnosing kidney-related medical conditions. The objective was to evaluate if urine test strip analysis might serve as an initial and efficient screening method for reflex testing with accurate quantitative methods.

Materials and Methods: Freshly voided urines (n=206) were analysed using two urine test strip brands on UC-MAX (Menarini) and cobas u 601 (Roche Diagnostics) instruments. Ordinal scale categories and reflectance signals (if available) were both used for the comparison with reference quantitative methods for glucose, proteins and albumin (cobas 503). Samples were considered positive when glucose >15 or ≥54 mg/dL, proteins ≥200 mg/L and albumin ≥10 mg/L. Optimized reflectance thresholds were calculated by ROC curve analysis. Analytical performance specifications (APS) for trueness of test strip were gathered from EFLM guideline (FP_D, FN_G, FN_C).

Results: Reflectance signals were significantly lower in urine samples considered positive by the reference method (p<0.0001) (Figure 1). Reflectance signals were also correlated with quantitative measurements, showing strong correlation (0.754 to 0.969). Only the use of optimized reflectance thresholds on cobas u 601 achieved at least the minimum EFLM APS (FP_D <20%, FN_G <50% and FN_C <10%).

Conclusion: The use of reflectance signals from urine test strips enhanced accuracy for glucose, proteins, and albumin measurement and may contribute to improve diagnosis of diverse kidney-related conditions.

Figure 1: Comparison of reflectance data to quantitative measurements. Reflectance data compared to positive and negative cut-off values. A = glucose; B = proteins; C = albumin.

SALIVARY BIOMARKERS: PROOF OF CONCEPT IN REFLUX PATIENTS

N. De Vos ¹, M. Antoine¹, H. Dahma², S. Saussez³, J. Lechien³

¹Department of Clinical Chemistry, LHUB-ULB, Université Libre de Bruxelles and CHU Saint-Pierre, Brussels, Belgium, ²Department of Infectious serology, LHUB-ULB, Université Libre de Bruxelles and CHU Saint-Pierre, Brussels, Belgium, ³Department of Human Anatomy and Experimental Oncology, Faculty of Medicine, UMONS Research Institute for Health Sciences and Technology, University of Mons, Mons, Belgium

Objectives: There is a lack of non-invasive methods in medical diagnosis. Some patients are resistant to needle prick for blood sampling. Saliva is an interesting alternative matrix, solving the problem of an invasive medical procedure. In this study, we investigated saliva of laryngopharyngeal reflux (LPR) patients. The clinically relevant salivary biomarkers are presented here with decisional cut-off.

Methods: Saliva of patients with LPR at the hypopharyngeal-esophageal multichannel intraluminal impedance pH-monitoring (HEMII-pH) was prospectively collected through a Salivette device. The saliva biomarkers elastase, pH, total cholesterol and total bilirubin were assessed. Cut-off was obtained through Receiver Operating Curve (ROC) with Analyse-it (Excel, USA).

Results: Twenty LPR patients and twelve healthy controls were enrolled. Significantly higher salivary elastase, higher salivary pH and lower salivary cholesterol were found in LPR patients compared to healthy controls (p<0.05). At a cut-off of >31.5 µg/mL, elastase reported sensitivity (sens) and specificity (spec) of 84% and 79%. The cut-off for pH was >7.3 (sens=51%, spec=96%) and cholesterol <3.3 mg/dL (sens, spec=100%). Total bilirubin remained undetected.

Discussion: Basic research has been done to roll out a non-invasive diagnostic method in saliva of reflux patients. Decisional criteria have been found for the clinical support of reflux when salivary elastase is above 31.5 µg/mL, pH above 7.3 and total cholesterol beneath 3.3 mg/dL. Total bilirubin failed as a salivary marker in this patient population.

Conclusion: This study shows successful use of saliva in reflux patients for the measurement of elastase, pH and cholesterol. Expertise with saliva specimen is growing in our salivary clinic. In the coming year, further research will be done with saliva in other medical settings to enlarge the scope of salivary biomarkers.

(international patent pending EP23188695.3)

EVALUATION OF SEMI-QUANTITATIVE URINARY DRUG SCREENING TESTS BY LIQUID CHROMATOGRAPHY COUPLED WITH TANDEM MASS SPECTROMETRY (LC-MS/MS)

K. Benboujida ¹, F. Ponthieux¹, F. Cotton¹

¹LHUB-ULB, Université Libre de Bruxelles, Bruxelles, Belgium

Objective: Urinary drug assays are designed to quickly detect drug use in suspected intoxicated patients. Our aim was to evaluate the analytical performances of the kinetic interaction of microparticles in solution (KIMS) screening assays in urine (Roche Diagnostics) in comparison with a reference method (LC-MS/MS).

Material and methods: Between January and March 2023, 1758 urine samples received at the laboratory for routine drug screening were analyzed with the KIMS tests for one or several of the following families: amphetamines, benzodiazepines, cannabis, cocaine, methadone and opiates. Positive urines were then confirmed by LC-MS/MS using the MassTox® Drugs of Abuse testing in Urine kit (Chromsystems Diagnostics) on a UPLC-1290 and MS 6490 system (Agilent Technologies). The analytical performance of the KIMS assays was evaluated in comparison with LC-MS/MS.

Results: Out of the 1758 samples analyzed, 821 were positive for at least one screening test. The confirmation by LC-MS/MS allowed to calculate the results summarized in table 1.

Low benzoylecgonine levels, observed in case of past cocaine intake, gave frequently negative screening tests but were well detected by LC-MS/MS. The amphetamines screening test presented a high false positive rate of 72% (n=77) whereas 96 samples containing opiates had a negative screening result.

Conclusion: According to those results, KIMS screening tests are suitable to detect most of drugs of abuse. However, they suffer from the cross-reactivity of many interfering molecules with the amphetamines test and do not detect some members of the opiate family. This underlines the need for more performant methods like LC-MS/MS, even in first line.

AN APPARENT DISCREPANCY IN BONE TURNOVER MARKERS LEADING TO THE DIAGNOSTIC OF A RARE GENETIC DISEASE

A. Ladang ¹, C. Wirth², M. Bosse¹, E. Cavalier¹

¹Department of Clinical Chemistry, CHU de Liège, Liège, Belgium, ²Rhumatology department, CHL, Luxembourg, Luxembourg

Case description: A 69 years old woman was followed for several years for unexplained fractures and, in particular, for atypical fractures occurring at the start of a bisphosphonate treatment.

Laboratory investigations: A full evaluation of bone turn-over markers was done in our lab, revealing normal levels of bone alkaline phosphatase (14.1 µg/L) and a slight increase of P1NP (98.6 µg/L). On the other hand, C-telopeptide collagen type 1 (CTXS) was slightly decreased (291.7 ng/L), but Trap-5b was above all levels ever measured in the lab (> 56 U/L). Interference with hemolysis or rheumatoid factor were excluded, as was cross-reaction with oestradiol.

Discussion: MD was suspecting type 2 osteopetrosis without biochemical prove but Trap-5b level combined with low CTXS suggests loss of osteoclastic function, thus confirming this hypothesis. Indeed, while a high Trap-5b reflects a high number of osteoclasts, a low CTXS shows there is no proper bone resorption. Type 2 osteopetrosis is an autosomal dominant disorder due to mutations on the chloride channel 7 gene. This mutation enables acidification of bone matrix needed to perform bone resorption. As bone resorption is ineffective, bone density increases, leading to increased osteoclasts. The low bone turnover rate sensitizes patient to fractures as microdamage cannot be repaired anymore, leading to a more fragile bones. As bisphosphonate slow bone resorption by acting on osteoclasts, the few remaining efficient osteoclasts were blocked, emphasizing the phenotype.

Conclusion: Osteopetrosis has several phenotypes, mainly characterized by atypical fractures, but only the type 2 has a late onset. In addition to increased bone density and pathological fractures, patients with osteopetrosis are at risk of cranial nerve compression and moderate hematological failure.

VERIFICATION OF A HIGH-SPEED CENTRIFUGATION PROCESS TO ELIMINATE LIPEMIA INTERFERENCE

J. Cabo ¹, J. Favresse¹

¹Clinique Saint-Luc Bouge, Bouge, Belgium

Objective: Lipemia constitutes an unavoidable analytical interference, contingent upon the fasting state and metabolic condition of the patient. The objective of this study was to evaluate the impact of lipemia on multiple measurands and to determine whether high-speed centrifugation process could effectively alleviate the interference.

Materials and Methods: A reference pool was generated using leftover serum samples (n=25). Intravenous lipid emulsion (20%) was spiked to generate derivative pools, each representing varying degrees of lipemia, with lipemia indexes of 500, 1000, 2000, and 4000, respectively. Each lipemia pool was subsequently divided into two distinct sets: one tested in its original state and the other tested after high-speed centrifugation (miniSpin, 13.500 rpm), followed by the retrieval of the infranatant. A total of 70 different measurands were evaluated on cobas pro analyzers. The bias between the reference pool and each tested pool was calculated and compared against allowable specifications for bias gathered from the EFLM Biological Variation Database.

Results: In uncentrifuged pools, 8 (11.4%), 12 (17.1%), 23 (32.9%) and 31 (44.3%) measurands were significantly impacted by lipemia interference for L-indexes of 500, 1000, 2000, and 4000, respectively while in high-speed centrifuged pools, 3 (4.3%), 5 (7.1%), 11 (15.7%) and 19 (27.1%) measurands were impacted. The L-index showed a strong correlation with intravenous lipid emulsion concentrations (R² = 0.9994) (Figure 1). High-speed centrifugation reduced the L-index by an average of 94.9%.

Conclusion: High-speed centrifugation effectively reduces the L-index and efficiently eliminates lipemia interference for numerous measurands. In some situations, the centrifuged sample did not meet the allowable bias while the uncentrifuged sample did. The potential impact of the centrifugation process on the measurand itself should therefore be carefully evaluated.

Figure 1: Impact of high-speed centrifugation on L-indexes.

NMR-BASED METABOLOMICS: A NEW TOOL FOR PATIENTS’ FITNESS MONITORING AND INJURY PREVENTION

M. Schoumacher¹, P. Massonnet², P. De Tullio³, E. Cavalier², C. Le Goff ²

¹Clinical Metabolomics group (CliMe), Center for Interdisciplinary research on Medicines (CIRM), University of Liège, Liège, Belgium, ²Department of Clinical Chemistry, University of Liège (Liège), CHU Sart Tilman, Liège, Belgium, ³Clinical Metabolomics group (CliMe), Center for Interdisciplinary research on Medicines (CIRM), University of Liège, Liège, Belgium

Objectives: By providing a holistic view of the metabolic status of an individual, metabolomics is an essential tool for biomarker discovery and personalized health monitoring. In this work, we aim to develop an NMR-based metabolomic workflow to spot specific metabolic signature of patients’ fitness and recovery.

Materials and methods: To achieve our goals, we build up a longitudinal study design in which participants will exercise at the maximum of their capacity. While sedentary individual (n= 15) will run for 1 hour on a treadmill, UT patients (n=20) will participate to the Ultratour of Liège (Belgium), a 67 km long ultra-trail running with 1500 D+. Blood samples was collected at three time points: before (TO), immediately after (T1), and three hours after the start of their effort (T3). NMR-based metabolomic analysis was employed to analyze intragroup and intergroup metabolic variations using dedicated tool for spectra processing and multivariate statistical analysis.

Results: Our strategy highlighted valuable information about individuals’ metabolic status. Indeed, intragroup analysis spotted metabolic profile changes from T0 to T3 in both groups. Moreover, our data shown that three hours of recovery were insufficient for athletes to return to their baseline metabolic status while sedentary individual return closer to baseline. Significant differences between athletes and sedentary individuals were also highlighted. In this field, blood-derived biomarkers appear to give the best representation of an athlete’s condition and helps clinicians/coach/team to individualize workload and prevent injuries.

Conclusions: This proof-of-concept study spotted the importance of metabolomics approach in the field of athletes’ follow-up, performance monitoring and injury prevention. Indeed, by following the metabolic changes over the time, we can assess the fitness of the studied individuals and gather information about their recovery and avoid situation of high-risk of muscle injury.

ISO 15189:2022 PAVES THE WAY TO GLUCOSE METER ACCREDITATION

E. Catry ¹, L. Moreno Y Banuls¹, M. Closset¹, F. Mullier¹

¹CHU UCL Namur, Yvoir, Belgium

Introduction and objective(s): Since December 6^th 2022, ISO 22870, which provided specific requirements applicable to point-of-care testing (POCT), has been withdrawn and revised by ISO 15189:2022. POCT must now adhere to all the requirements of ISO 15189:2022, in addition to some specific items. Currently, BELAC accreditation for laboratory medicine remains non-mandatory. However, our two-site CHU UCL Namur laboratory has been BELAC-accredited since April 2011, we are now exploring the accreditation process for POCT, with an initial focus on glucose meters.

Material and Methods: First, we conducted an inventory of “Dos and Don’ts” in accordance with ISO 15189:2022 and glucose meter accreditation. As in situ-analyses, traceability is of significant importance in ISO processes, including but not limited to competence requirements and frequency of assessment, equipment maintenance and repair, and metrology. Moreover, documentation, instructions for use, standards, manuals, and reference data must be readily accessible to all operators. Participation in external quality control (EQA) programs is mandatory for POCT accreditation.

Results: A major challenge for glucose meter accreditation is ensuring the traceability of operator training and certification, as well as the frequency of assessments. Typically, the use of a glucose meter is learned during nursing training, and evaluating the operator’s competences by the laboratory (or by a delegate such as the POCT coordinator) is often met with resistance during this period of pressure on healthcare professionals. Nevertheless, we decided to initiate a certification process on April 3^rd, 2023, based on the implementation of “Observer Test Sequence (OTS)”. Operators were given a three-months window to complete their OTS certification; otherwise, their user ID would be deactivated by the system on July 2, 2023.

To date, a total of 501 operators have been certified with a 2-year certification, representing less than 20% of the total operators registered in the management software, COBAS IT 1000. Among them, over 40% of operators achieved certification within one month after the OTS implementation. Numerous uncertified operators are associated with short-term staff or interim personnel. Their user ID will be deactivated by September 2023.

Conclusion: Certification through the OTS was just the initial step in lengthy accreditation process. It is crucial to foster the interest of operators in ISO 15189:2022 requirements and to collaborate in enhancing quality and patient safety. Several tools are already in place, including a multidisciplinary POCT committee dedicated to promoting and developing POCT in ISO respect, as well as POCT coordinators, who serve as key interlocutors with operators, in clinical chemistry or hemostasis.

ORTHO CLINICAL DIAGNOSTICS VITROS XT7600 HBA1C TEST: PERFORMANCE ASSESSEMENT OF GLYCATED HEMOGLOBIN MEASUREMENT BY ENZYMATIC METHOD

M. Serrari ¹, C. Renaux², L. Boitquin², L. De Mol², J. Brauner², A. S. Adams³, S. Benyaich³, P. Nguyen²

¹Epicura, Brussels, Belgium, ²Epicura, Hornu, Belgium, ³LHUB, Brussels, Belgium

Objective: The aim of this study was to evaluate an enzymatic measurement method of glycated hemoglobin (HbA1c) using Vitros XT7600 (Ortho Clinical Diagnostics, New Jersey, US).

Material and Methods: A comparison with the high-performance liquid chromatography technique (HPLC) from G8 analyzer (Tosoh Bioscience, Tokyo, Japan) was carried-out, regarding the following parameters: repeatability, reproducibility, accuracy, stability and hemoglobin variants interference testing. Performance characteristics were determined on internal quality control (IQC) and patient samples. French Society of Clinical Biology (SFBC) acceptance criteria were used. Comparison was assessed with Passing-Bablok and Bland-Altman methods. 55 different patient samples whole blood EDTA material were used. 13 hemoglobin variants samples were performed to determine analytical assays interference.

Results: Repeatability, for normal (5%) and high (11%) HBA1C levels (n=15), was 5.1 and 9.8% with a coefficient of variation (CV) of 1.2 and 1.3%, respectively. Reproducibility was, for normal (6.4%) and high (11%) IQC levels (n=15), 6.3 and 10.6% with a CV of 2.2 and 0.8%, respectively. Accuracy, for normal (6.4%) and high (11%) HbA1c levels (n=15), was 6.3 and 10.6%, a relative bias of -1.9 and -3.4%, respectively. The patient samples were stable for 3 days at 2-8°C. Comparison with Tosoh G8 demonstrated a Bland–Altman plot: 0.128 -0.338, with a mean bias value of 0.233. The curve of Passing–Bablok regression was y = 0.959x + 0.0286. Hemoglobin variants performed (Hb S, Hb C, Hb D-Punjab and Hb Lepore) appear to give consistent results for both methods. HbA1c measurements lower than 4% on Vitros XT7600 could suspect homozygous or compound heterozygosity hemoglobin variants as no HbA1c was detected on Tosoh G8. Moreover, there is no hemoglobin variant detection alert on Vitros XT7600.

Conclusion: Vitros HbA1c solution reaches the predefined acceptance criteria for analytical performance. Results are comparable with Tosoh Bioscience G8. Uncommon hemoglobin variants may interfere.

Acknowledgments: Epicura laboratory, LHUB-ULB laboratory.

QUANTITATIVE ABNORMALITIES IN THE Β-REGION OF THE ELECTROPHORETIC PROFILE OF SERUM PROTEINS AS PREDICTIVE MARKERS OF MONOCLONALITY

S. C. Cherkaoui ¹, S. P. Penickova¹, F. Cotton¹

¹LHUB-ULB, Bruxelles, Belgium

Introduction: Monoclonal gammopathies are a group of hematological disorders characterized by the abnormal proliferation of a single clone of plasma cells, resulting in the production of monoclonal immunoglobulins known as M-proteins. Identifying and quantifying M-proteins that migrate within the β-region is complicated by the presence of native serum proteins such as transferrin or C-reactive protein. Although reflex testing guidelines for γ-region anomalies are well established, the same does not hold true for β-region irregularities.

Objectif of the study: To develop a predictive model that allows us to detect M-protein when the β-globulin fraction is abnormal.

Materials and method: We included patients who underwent routine serum protein electrophoresis and immunofixation (IF) in our laboratory. The patients with electropherogram showing elevated β1 and/or β2 fraction levels were included. Following strict exclusion criteria such as acute inflammation, iron deficiency anemia or supposed liver disease, 309 patients were selected, aged from 30 to 97 years with a male to female ratio of 54/46.

To construct a predictive model for the need to perform IF, the following data were used: age, sexe, β1-globulin, β2-globulin, hypogammaglobulinemia, total serum protein, IgA, IgG and IgM. The dataset underwent a random split, dividing it into a foundational training set (80%, 247 samples) and a foundational test set (20%, 62 samples). The training set was subjected to five different algorithms: logistic regression, decision tree, random forest, gradient boosting, and support vector.

Results: Among 309 selected patients, 149 exhibited a negative IF (no M-protein demonstrated) and 160 a positive IF (M-protein detected).

The statistical analysis showed a high correlation between the age, β2-globulin, total serum protein, IgA and a positive IF.

The evaluation of the five tested models demonstrated very good performance, the chosen model was random forest for its high sensitivity (85%) and AUROC (91%). The figure 1 shows the ability of the model to discriminate between true positive (positive IF = 1, in red) and true negative (negative IF = 2, in blue) samples.

Figure 1: Scatter plot showing threshold at 0.5 classifying instances into the “positive IF” or “negative IF” categories.

Conclusion: This study used several analyzed parameters to determine those that had a real link with a positive IF. The machine learning study enabled us to choose a model (Random Forest) that demonstrated the ability to discriminate between positive and negative cases effectively.

The main limitation of this study is the small cohort, considering that machine learning models can be self-taught with additional data, the aforementioned results are subject to improvement but promising.

The project was carried out with the support of LHUB-ULB 2023 grant.

EXAMINING THE EFFICACY OF CLINELL SINK DRAIN DISINFECTANT IN THE INTENSIVE CARE UNIT OF THE UZ BRUSSEL

R. Vanstokstraeten ¹, B. Gordts², L. Blommaert¹, N. Verbraeken¹, G. van der Kelen¹, K. Emmerechts¹, I. Wybo¹

¹Department of Microbiology and Infection Control, Vrije Universiteit Brussel (VUB), Universitair Ziekenhuis Brussel (UZ Brussel), 1090 Brussels, Belgium, Brussel, Belgium, ²Dialex Biomedica, Bilzen, Belgium

Aim: Sink drains in hospitals often harbor multidrug-resistant microorganisms, acting as a reservoir for nosocomial outbreaks. Considering that disinfectants based on peracetic acid are both non-corrosive and effective in eliminating biofilm and planktonic microorganisms, they present a potential solution for decontaminating sink drains. In this study, we examined the effectiveness of Clinell Drain Disinfectant®, a peracetic acid-based disinfectant, in the intensive care unit (ICU) of UZ Brussel, a department that has been confronted with nosocomial infections linked to sinks for several years.

Methods: All 10 sinks in one of the ICU subunits proven to be heavily contaminated were treated with Clinell Drain Disinfectant® for one month. Throughout the treatment period, bacterial growth in the P-traps was systematically monitored both quantitatively (using eSwab®) and qualitatively (employing a sterile catheter and syringe) on various selective agar plates, processed, and incubated in the WASPLab® system. MALDI-TOF MS was used to identify all morphologically distinct colonies. Additionally, the utilization of O.K.N.V-RESIST5® allowed the identification of some clinically relevant acquired resistance mechanisms.

Results: At baseline, most of the sink drains were colonized by high concentrations of multidrug-resistant microorganisms, primarily VIM-producing P. aeruginosa. Our study demonstrated the disinfectant’s efficacy in decontaminating the P-trap liquid, as the cultures taken immediately after decontamination yielded negative results, with only a very few exceptions. Nonetheless, it was observed that the biofilm in the upper section of the drain system remained unaffected, and within two days, it was capable of recolonizing the liquid in the P-traps at a concentration exceeding 100 000 CFU/mL.

Conclusion: Clinell Drain Disinfectant® has demonstrated its effectiveness in completely eliminating bacteria from the liquid in P-traps in the sink drainage system. However, in our specific setting, the disinfectant does not affect significant parts of the biofilm present in the sink drains. As a result, the P-trap water experiences rapid and massive recolonization after disinfection. Our study demonstrates that the decontamination of the draining system and the eradication of biofilm could be more complex than previously thought.

LEGIONELLA PNEUMOPHILA ENDOCARDITIS: A CASE REPORT

E. De Hert ¹, M. Vanden Driessche¹, H. Jansens¹, V. Matheeussen¹, S. Vandamme¹

¹Department of Microbiology, University Hospital Antwerp, Edegem, Belgium

Objective: We describe a case of Legionella pneumophila endocarditis and its diagnostic challenges.

Case description: A 75-year-old man with a history of aortic valve replacement with a bio-prosthesis was hospitalized after a 3-month trip to Thailand, due to fever chills, night sweats, stomachache, headache and weight loss since one month. Endocarditis of the bio-prosthesis was diagnosed via transesophageal echocardiography, flucloxacillin and vancomycin were started. Blood cultures remained negative, so serological and molecular tests were deployed. Legionella urinary antigen was negative, but IgG and IgM of L. pneumophila serotype 1-7 and PCR for L. pneumophila on whole blood came back positive. Respiratory samples were not investigated. Blood cultures remained negative despite subculturing on Legionella specific agars. Antibiotics were switched to levofloxacin and the patient received a new aortic valve bio-prosthesis. Culture of the infected prosthesis showed growth on BCYE and GVPC agar after 6 days, which was identified as L. pneumophila via MALDI-TOF. Serotyping showed presence of serotype 2-15. Two weeks later the patient was discharged in good condition, levofloxacin was continued for 5 months.

Conclusion: L. pneumophila is mostly associated with pneumonia, however extrapulmonary infections, including endocarditis, occur and are important to recognize. The diagnosis remains difficult, as standard blood and tissue culture media do not support growth of Legionella and not all patients show respiratory symptoms, as demonstrated in this case. Therefore, Legionella should always be considered in case of culture negative endocarditis. Here, we describe a case of biological valve endocarditis caused by L. pneumophila. Diagnosis was made by serology and PCR, confirmed by a positive culture of the infected bio-prosthesis. Literature corroborates that the diagnosis of Legionella endocarditis is mostly made by positive serology and/or PCR on whole blood or prosthesis.

Acknowledgments: We thank our colleagues of the Cardiology and Medical Microbiology department of the Admiraal De Ruyter Ziekenhuis Goes for the initial diagnosis. We thank our colleagues of the Cardiac Surgery Department of the Antwerp University Hospital for the good care of this patient.

ESTABLISHING A THRESHOLD FOR AUTOMATED TREPONEMAL SCREENING TESTS: DISCRIMINATING NON-SPECIFIC REACTIONS FROM A TRUE DIAGNOSIS OF SYPHILIS USING THE ELECSYS SYPHILIS ASSAY (ROCHE DIAGNOSTICS, MANNHEIM, GERMANY)

S. C. Cherkaoui ¹, S. V. van den Wijngaert¹, H. Dahma¹, D. M. Martiny¹

¹LHUB-ULB, Bruxelles, Belgium

Introduction: Syphilis is a globally significant STI caused by Treponema pallidum. Its early detection is crucial for effective disease management and control [1]. Serology is widely used as the primary diagnosis tool, employing a combination of treponemal tests (TT) and non-treponemal tests (NTT) [2].The frequent need for a second confirmatory manual TT in case of positive results in automated TT and negative results in NTT has proven to be both costly and time-consuming [3].

Aim of the study: To determine a reliable S/CO ratio in automated TT to differentiate non-specific reactions from true syphilis infections and therefore minimize the need for unnecessary confirmation tests.

Material and method: From January 1, 2020, to December 31, 2022, a cohort of 1397 patients was analyzed as part of the routine operations at LHUB-ULB. Our work focused on samples that were ECLIA-positive using the Elecsys syphilis assay (Roche Diagnostics, Mannheim, Germany). All positive syphilis screenings underwent subsequent confirmatory testing, starting with a NTT, the Rapid Plasma Reagin (RPR) (Lorne laboratories, RPR CARBON KIT, Berkshire, UK). If the RPR is non-reactive, a manual TT, T. pallidum particle agglutination assay (TPPA) (Fujirebio, Europe N.V) is performed, regardless of the S/CO ratio, to differentiate between a non-specific positive automated TT result and a true syphilis result.

A receiver operating characteristic (ROC) curve was used to determine the S/CO threshold that effectively discriminates between true positive samples (TPPA positive) and false positive samples (TPPA negative).

Results: An optimal S/CO value of 30.00 was selected to achieve a balance between sensitivity and specificity (Figure 1).

Figure 1. Receiver operating characteristics (ROC) curve comparing S/CO screening results with TPPA results (considered as the reference value).

This threshold value effectively differentiated between syphilis-positive and syphilis-negative individuals with a high level of accuracy. Importantly, 98% of patients with an S/CO screening value ≥ 30 and a negative RPR test were TPPA positive.

Conclusion: The proposed cutoff effectively distinguishes between non-specific reactions and true syphilis diagnoses, resulting in a 40% reduction of unnecessary confirmations.

References: 1. Yi J, Choi W, Shin S, Choi J, Kim H, Chung HJ, Moon HW, Hur M, Yun YM. Strategy for performing treponemal tests in reverse-sequence algorithms of syphilis diagnosis. Clin Biochem. 2019 Jan;63:121-125

2. French P, Gomberg M, Janier M, et al; IUST. IUSTI: 2008 European guidelines on the management of syphilis. Int J STD AIDS 2009; 20:300–9

3. Berry GJ, Loeffelholz MJ. Use of Treponemal Screening Assay Strength of Signal to Avoid Unnecessary Confirmatory Testing. Sex Transm Dis. 2016 Dec;43(12):737-740

EVALUATION OF 3 METHODS FOR ANTIMICROBIAL SYNERGY TESTING BETWEEN AZTREONAM AND CEFTAZIDIME-AVIBACTAM

E. Cottone ¹, A. De Bel¹, L. Coorevits¹, M. Boudewijns¹

¹AZ Groeninge Kortrijk, Kortrijk, Belgium

Objectives: Antibiotic resistance is globally increasing, causing a higher need for tailored antibiotic combinational therapies, e.g. for NDM-metallo-beta-lactamase. Aztreonam (ATM) is active upon metallo-beta-lactamase, but is normally inactivated by concomitant ESBL- or AmpC-type beta-lactamase. Combination with ceftazidime-avibactam (CZA) can restore its activity due to inhibition of ESBL/AmpC with proven clinical success (1).

Material and method: We compared three methods during three different days on ten strains of Enterobacterales with different resistance patterns from clinical samples:

• Double-disk synergy test (DDST): ATM and CZA disks were placed 2 cm apart on Mueller-Hinton (MH) agar; the presence of an inhibition zone between the two disks was evaluated (1).

• Cross method: ATM and CZA E-tests were placed on MH-agar, intersecting at their individual MIC-value for the organism; after overnight incubation. ∑FIC was calculated (see ref. 2); ∑FIC ≤ 0,5 is indicative for synergy (2).

• Superposition method: E-test of CZA was placed for 10 minutes resp. 1 hour on MH-agar, before removing and applying E-test of ATM on the exact same position. After overnight incubation, synergy was evaluated based upon 2 criteria: 1) MIC ATM ≤ 1 mg/L and 2) a 4-time decrease of MIC ATM in superposition compared to the MIC-value for ATM alone (2,3).

Results: DDST method, superposition method after 1 hour with both interpretation criteria and superposition method after 10 minutes with the ATM MIC criterium showed fully concordant results. The cross method was the least reproducible method with concordant results for only 6 out of 10 isolates.

Conclusion: The most feasible method for the microbiological laboratory is the superposition method, as is the DDST method.

References: (1) Falcone et al. Efficacy of Ceftazidime-avibactam Plus Aztreonam in Patients With Bloodstream Infections Caused by Metallo-beta-lactamase-Producing Enterobacterales.Clinical Infectious Diseases. 72 (2021) 1871-1878.

(2) Pankey et al. Comparison of 3 Etest® methods and time-kill assay for determination of antimicrobial synergy against carbapenemase-producing Klebsiella species. Diagnostic Microbiology and Infectious Disease 77 (2013) 220–226.

(3) Rawson et al. A practical laboratory method to determine ceftazidime-avibactam-aztreonam synergy in patients with New Delhi metallo-beta-lactamase (NDM)–producing Enterobacterales infectionJournal of Global Antimicrobial Resistance 29 (2022) 558–562

EVALUATION OF THE PERFORMANCE OF LUMIRADX SARS-COV-2 AG ULTRA AND LUMIRADX SARS-COV-2 & FLUA/B

P. Mbouche ¹, M. Tré-Hardy¹, L. Blairon¹

¹Iris Hospitals South, Brussels, Belgium

Objective: The endemic circulation of SARS-CoV-2 requires the continuation of a reliable screening strategy. Rapid antigen diagnostic tests (Ag-RDT) are widely used, enabling rapid decision-making without the need for heavy equipment. Our aim was to verify the analytical performance of the two rapid sandwich-type microfluidic immunofluorescence assays SARS-CoV-2 Ag-Ultra and SARS-CoV-2 & Flu A/B on LumiraDx device (LumiraDx, Alloa, UK).

Material and Methods: Frozen nasopharyngeal samples issued from the routine, collected on VST transport medium and previously analyzed with a RT-PCR technique were retested with the two LumiraDx kits. The 77 positive samples with known cycle threshold (Ct) values were classified into 3 categories: group 1 (N=27), very high viral load (VL) (17 ≤ Ct < 25); group 2 (N=35) high VL (25 ≤ Ct < 30); group 3 (N=15), moderate VL (30 ≤ Ct < 36). 100 fresh negative samples were also tested. A 100-µl aliquot of the samples was absorbed using a swab, transferred into the extraction buffer, and then analyzed with the two LumiraDx assays in parallel. The results were compared to the reference RT-PCR method.

Results: The table 1 summarizes the main results.

Conclusion: The LumiraDx device is an easy-to-use polyvalent point-of-care instrument capable of running chemistry and coagulation tests, cardiac markers, as well as viral antigens. The SARS-CoV-2 Ag-Ultra test meets WHO recommendations for the use of SARS-CoV-2 Ag-RDT (cumulative sensitivity >90% at a Ct of 30), and demonstrates a substantial level of agreement with RT-PCR. Conversely, the SARS-CoV-2 & Flu A/B does not comply with these recommendations although it can provide valuable additional information. As expected, the sensitivities of these Ag-RDTs are highly influenced by the VL.

VERIFICATION OF THE RESPIRATORY PATHOGEN PANEL (RPP12) PERFORMED ON SANITY 2.0

K. De Letter ¹, M. Berth¹

¹AZ Alma, Eeklo, Belgium

OBJECTIVE. This study aimed to verify the performance of a ready-to-use respiratory multiplex PCR performed on Sanity 2.0 (Xiamen Zeesan Biotech).

MATERIAL AND METHODS. Sanity 2.0 offers a respiratory pathogen panel (RPP12) for the qualitative detection of nine respiratory viruses, incl. influenza A & B and respiratory syncytial virus, as well as three bacteria: M. pneumoniae, C. pneumoniae and B. pertussis. The analytical precision of the assay was evaluated by repeated measurement of Amies medium pools with different viruses, in one run. To evaluate the analytical accuracy, a commercial respiratory pathogen panel (Vircell) was analysed. A method comparison of 50 nasopharyngeal swabs, collected in Amies medium (eSwab), was performed between the RPP12 on Sanity 2.0 and the respiratory panel 2.1 plus on BioFire FilmArray (bioMérieux). The respiratory panel 2.1 plus detects the same pathogens as RPP12 but allows additional detection of MERS-CoV, SARS-CoV-2 and B. parapertussis.

RESULTS. The precision and accuracy experiments showed good repeatability and no bias was observed. The method comparison with BioFire FilmArray displayed a qualitative concordance of 70% on virus/bacterium specific level.

CONCLUSION. The instrument portrayed a good analytical performance fit for routine use. The method comparison showed a moderate correlation between Sanity 2.0 and BioFire FilmArray. It should be noted that there is no reference method available for the detection of respiratory pathogens through PCR. The main assets of the RPP12 are the extent of the panel (incl. bacterial pathogens) and the automated interpretation of the PCR by the software, as well as the possibility for manual review of the melting curves. The assay has a good turnaround time (150 minutes) and a reasonable price. The drawbacks of the system include the fact that this is a batch system (one run requires four samples), the incapability of the detection of SARS-CoV-2 and the absence of clinical performance data.

PRE-ANALYTICAL PHASE OF PARAMETERS WITHIN HUMORAL IMMUNOLOGY: LI-HEPARIN TUBES VS. ALIQUOTS AND STABILITY ASSESSMENT

S. De Koninck ¹, J. Hemelings¹, E. Linskens¹, S. Demeester¹

¹Free University of Brussels (VUB), University Hospital Brussels, Department of Laboratory Haematology, Brussels, Belgium

Objective: The aim of this project is the comparison of the parameters IgG, IgG₂, IgG₃, IgG₄, IgA, IgM, C3c, C4, C1-inhibitor (C1-inh), λ_Free, and κ_Free over a 7-day period after sampling, juxtaposing two methods (Li-heparin tubes with (LH-GS) and without (LH) gel separator) against the aliquot, our reference method, to determine whether excluding the aliquoting step poses any risk to the stability of these parameters.

Materials and methods: Venous blood of 27 volunteers was collected in two different types of tubes, the S-Monovette^® LH tube and S-Monovette^® LH-GS tube (’Sarstedt AG&Co’). An aliquot was taken from one S-Monovette^® LH tube and served as reference for further analysis, as routinely done in the laboratory. Analysis of the different parameters was performed on the Optilite^® (‘The Binding Site Group Ltd’). The stability of the parameters was evaluated after 3 and 7 days storage at 2-8°C by calculating the mean % difference (±95% CI) vs the aliquot and results were compared respecting the total allowable error (TEa, EFLM) or total coefficient of variation (CV) (‘The Binding Site Group Ltd’).

Results: Only for C3c, the mean % difference [95%CI] exceeded the TEa (7.80%) in both tubes at day 7 (LH tube 10.39% [8.61%-12.16]; LH-GS tube 7.28% [5.46%-9.10%]), but not at day 3. For the other parameters, TEa or total CV was not exceeded up to 7 days.

Conclusion: 10 of the 11 investigated parameters (IgG, IgG₂, IgG₃, IgG₄, IgA, IgM, C4, C1-inh, λ_Free, and κ_Free) are stable up to 7 days after sampling in both the LH and LH-GS tube, when compared to the reference method, the aliquot. Notably, only C3c exhibits stability up to only 3 days in both examined tubes.

Disclosure: No author has financial relationships with any commercial organization that might appear to represent a conflict of interest with the material presented in this report.

METHOD VALIDATION OF A NEW STAGO APTT REAGENT (PTT-A) ON STA COMPACT/SATELLITE MAX

T. Vanhove ¹, I. Torrez¹, L. De Keersmaecker¹

¹Department of Laboratory Medicine, Jan Ypermanziekenhuis, Ieper, Belgium

Objectives: The activated partial thromboplastin time (APTT), used as a general screening of the intrinsic coagulation cascade, is a coagulation test that is routinely requested in clinical laboratories. Due to a discontinuation of our current reagent (Cephascreen), a transition to an alternative assay (PTT-A) was imminent. The objective of this study was therefore to verify the new approach.

Material and methods: The APTT was evaluated using both Cephascreen and PTT-A assays on the STA Compact Max 2 coagulation automate as well as on the back-up STA Satellite Max (Diagnostica Stago). Correlation study analyses were performed on routine patient samples, solely on the Compact Max. We anticipated that the correlation would show a positive bias, likely exceeding the total error allowable (TEa = 15%). Precision and bias data was gathered through quality control (STA Routine QC) repetitions. The mean normal patient time (MNPT), as well as the reference range, was updated.

Results: The reagent switch presents a profound change. Our validation study shows a 15% increase on mean clotting time. This was previously communicated by the producer and is confirmed in literature [1-3]. The APTT upper reference value (syn. upper limit of normal; ULN) changes from <33.6 sec to <40.7 sec. This can presumably be explained by the type of activator: silica versus ellagic acid [2,3]. The APTT ratio, calculated by dividing the patient APTT value by the MNPT (28.8 sec previously, now 33.4 sec), shows no discrepancy and thus can simply continue to be used for monitoring patients under unfractionated heparin (UFH) therapy [4]. Furthermore, the PTT-A shows acceptable analytical performance in regard to precision and bias. Verification after continuous data clean-up (p97,5%) resulted in an ULN of 39.3 sec.

An interaction between CRP and APTT has been described: PTT-A shows prolonged clotting times from ± 100.0 mg/L [5]. No major differences have been described in terms of sensitivity to factor deficiencies, unfractionated heparin (UFH) and lupus anticoagulant [3,6-7]. The impact of in vitro interference by hemolysed samples corresponds to shorter clotting times (>10% reduction, in a range of <0.5 - 2.1 g/L free plasma hemoglobin) [8]. The relationship between APTT and anti-Xa, in context of UFH monitoring, has been extensively studied in literature [9-14]. We attempted a correlation between anti-Xa and APTT in kidney dialysis patients, in order to pass along a representative reagent-specific APTT-ratio range that correlates with a therapeutic anti-Xa range. As for now we apply a theoretical range of 1.4 - 3.4 [3].

Conclusion: The Stago PTT-A aPTT method fulfilled all verification criteria and has been successfully implemented.

References

[1] Stago BNL. APTT reagens studie: Maxima Medisch centrum en ZNA Middelheim. Diagnostica Stago, Asnières, France.

[2] Stroobants AK et al. Gevoeligheid van verschillende APTT-reagentia. Ned Tijdschr Klin Chem Labgeneesk. 2008.

[3] Rozen L et al. Evaluation of three APTT reagents in a routine laboratory: toward a compromise. Clin Lab. 2013.

[4] CLSI. H47-A3 One-stage prothrombin time (PT) test and activated partial thromboplastin time (APTT) test; approved guideline – 3rd edition.

[5] Devreese KMJ et al. Interference of C-reactive protein with clotting times. Clin Chem Lab Med. 2015.

[6] Lawrie AS et al. Determination of APTT factor sensitivity – the misguiding guideline. International journal of laboratory hematology. 2013.

[7] Dumoulin EN et al. Investigation of sensitivity for coagulation factor deficiency in APTT and PT: how to perform it? Clin Chem Lab Med. 2016.

[8] Woolley A et al. Effects of haemolysis, icterus and lipaemia on coagulation tests as performed on Stago STA-compact-max analyser. International journal of laboratory hematology. 2016.

[9] Gouin-Thibaut et al. Monitoring unfractioned heparin with APTT: A French collaborative study comparing sensitivity to heparing of 15 APTT reagents. Thombosis Research. 2012.

[10] Sta-Liquid anti-Xa Brochure. Diagnostica Stago France.

[11] Whitman-Purves E et al. Performance of anti-factor Xa versus activated partial thromboplastin time for heparin monitoring using multiple nomograms. Clinical and Applied Thrombosis/Hemostasis. 2018.

[12] Trucco M et al. Retrospective cohort study comparing activated partial thromboplastin time versus anti-factor Xa activity nomograms for therapeutic unfractionated heparin monitoring in pediatrics. Journal of Thrombosis and Haemostasis. 2015.

[13] Coene KLM et al. Protocolled redefinition of the therapeutic range for unfractionated heparin: lost in translation. Clinical and Applied Thrombosis/Hemostasis. 2018.

[14] Mulder MMG et al. ECMO and anticoagulation: a comprehensive review. Netherlands Journal of Critical Care. 2017.

CLASSIFICATION OF MDS PATIENTS BASED ON THE 5TH EDITION OF THE WORLD HEALTH ORGANIZATION

M. Vanrenterghem ¹, J. Dom ¹, M. Briers ¹, J. De Bie², L. Michaux², C. Van Laer³, M. Tajdar¹, N. Boeckx⁴

¹Department of Laboratory Medicine, University Hospitals Leuven, Leuven, Belgium, ²Center of Human Genetics, University Hospitals Leuven, Leuven, Belgium, ³Department of Laboratory Medicine, University Hospitals Leuven & Department of Cardiovascular Sciences, Centre for Molecular and Vascular Biology, University of Leuven, Leuven, Belgium, ⁴Department of Laboratory Medicine, University Hospitals Leuven & Department of Oncology, Laboratory of Experimental Hematology, University of Leuven, Leuven, Belgium

Objectives: Recently, the 5^th edition of the World Health Organization classification for myeloid and histiocytic/dendritic tumours was published. The impact of this classification, specific for myelodysplastic syndromes (MDS), was studied on previously classified MDS-patients. This entity is renamed as myelodysplastic neoplasms, and is defined by two major groups: 1) MDS with defining genetic abnormalities and 2) MDS, morphologically defined.

Materials and methods: A retrospective study was performed at University Hospitals Leuven between January 2020 - December 2021. Patient’s medical records were reviewed including morphologic findings on bone marrow aspirate, biopsy, cytogenetic- and molecular analysis.

Results: In total, 45 adults and 3 children with known MDS (n=48) were reclassified. Morphological examination of bone marrow aspirate, biopsy, and cytogenetics were available for all patients. Next generation sequencing (NGS) is routinely not performed in patients ≥ 70 years old, unless decided otherwise after multidisciplinary deliberation. Consequently, NGS data were only available for 24 patients (50%), including 5 patients ≥ 70 year.

Eleven patients were reclassified as ‘MDS with defining genetic abnormalities’, including 4 MDS-5q, 1 MDS-5q with ring sideroblasts, 3 MDS-SF3B1 and 3 MDS-biTP3. Five patients were classified as having a secondary neoplasm, of which 3 had an MDS, post cytotoxic therapy and 2 children with germline predisposition (germline GATA2 variant). There was 1 childhood MDS with low blasts - hypocellular, formerly classified as refractory cytopenia in childhood. All other patients (n=31) were reclassified as ‘MDS, morphologically defined’, as no genetic abnormality was found and/or extensively investigated.

Conclusion: The reclassification of MDS patients posed no significant problems. However, for patients ≥70 years, NGS was routinely not performed, given that the morphological classification of MDS will remain of great importance.

INTERFERENCE OF CHOLESTEROL CRYSTALS IN AUTOMATIC LEUKOCYTE COUNT IN PLEURAL FLUID: A CASE

R. Deley ¹, K. Truijens¹, N. Boeckx¹, M. Tajdar¹, C. Van Laer¹

¹Department of Laboratory Medicine, University Hospitals Leuven, Leuven, Belgium

Objective: For years, automatic leukocyte count and differential in body fluids are performed on hematology analyzers(1). We describe a case of cholesterol crystals interfering with automatic leukocyte count.

Case: A 77-year-old man with history of rheumatoid arthritis presented with recurring exertional dyspnoea. Imaging showed idiopathic pulmonary fibrosis and pleural effusion. Thoracentesis was performed and pleural fluid (PF) was send to the clinical laboratory.

Methods and results: Macroscopic cloudy, yellow coloured PF was analysed with the body-fluid module of Sysmex XN-9100 analyzer. Based on the interpretation of the WDF scattergram (figure 1A), both automatic WBC count and differential were unreliable. WBC were manually counted using a counting chamber and showed a lower leukocytes count compared to automatic count (table 1). WBC differential was performed on a cytospin preparate (table 1). Microscopic evaluation showed multiple ‘broken glass-like’ crystals, indicating cholesterol crystals (figure 1B&C). Chemical analysis showed high cholesterol (284.3 mg/L) and lower triglycerides (29.5 mg/L).

Conclusion: Cholesterol crystals are a typical feature of pseudochylothorax, a cholesterol-rich pleural effusion occurring in patients with underlying chronic inflammation(2,3). Therapy includes treatment of the underlying disease and thoracentesis. The presence of cholesterol crystals interfere with automatic cell counts in PF. It is important to correctly interpret the raw data of automated cell counters, and perform manual cell count to ensure correct results. In addition, recognising cholesterol crystals in PF is important for clinical diagnosis of pseudochylothorax and appropriate patient management.

Table 1: Automatic and manual cell counts.

Figure 1. A. WDF scattergram of XN analyzer showing unreliable WBC count and differentiation. B. Cholesterol crystals in counting chamber (Fuchs Rosental) (×50). C. Cholesterol crystals in cytospin preparate (May-Grünwald Giemsa) (×50).

References

1. Martín MJA, Adv Lab Med. 2021

2. Lama A, J Thorac Dis. 2016

3. AbdullGaffar B, Cytopathol. 2022

PROOF-OF-CONCEPT OF RAPID AND EASY-TO-PERFORM OPTICAL BIOSENSING ASSAYS FOR VWF ANTIGEN AND FUNCTIONALITY

B. Calcoen ¹, I. Pareyn¹, M. Jacquemin², N. I. Baikan³, J. S. Bérangère³, A. Veyradier⁴, C. Dekimpe¹, C. Tersteeg¹, K. Vanhoorelbeke⁵, S. F. De Meyer¹

¹Laboratory for Thrombosis Research, Kortrijk, Belgium, ²Department of Vascular Medicine and Hemostasis. University Hospitals Leuven and National Coordinating Hemophilia Center, University Hospitals Leuven, Leuven, Belgium, ³Service d’Hématologie biologique, Hôpital Lariboisière, AP-HP.Nord, Université Paris Cité, 75010 Paris, France, Paris, France, ⁴EA-3518 Recherche Clinique en hématologie, immunologie et transplantation, Institut de Recherche Saint Louis, Université Paris Cité, 75010 Paris, France, Paris, France, ⁵Laboratory for Thrombosis Research, Kortrijk and PharmAbs, the KU Leuven Antibody Center, KU Leuven, Leuven, Kortrijk, Belgium

Objectives: Von Willebrand factor (VWF) is an essential hemostatic protein. Mutations leading to either a quantitative or qualitative VWF deficit are the cause of von Willebrand disease (VWD). A laboratory-proven abnormality of VWF is needed for VWD diagnosis. However, laboratory determination of VWF parameters is limited to expert centra due to the need for multiple platforms and experienced personnel. Fiber optic – surface plasmon resonance (FO-SPR) could serve as a simple and easy-to-use alternative as its measurement method is based on a simple plasma dip. Our goal was to develop FO-SPR-assays for all crucial VWF parameters.

Materials and Methods: First, proof-of-concept experiments were performed for each FO-SPR assay, including VWF antigen (VWF:Ag), VWF binding to collagen (VWF:CB) and VWF binding to normal (VWF:GPIbR) and mutant (VWF:GPIbM) recombinant fragments of the GPIbα receptor (rGPIbα). Next, dose-response was investigated using a dilution series of normal human plasma. A type 3 VWD plasma sample was used to determine lower analytical sensitivity. Lastly, 90 relevant plasma samples were used to perform method comparison with in-house ELISAs, except for the VWF:GPIbR and VWF:GPIbM assays.

Results: Proof-of-concept was obtained for both FO-SPR VWF:Ag and VWF:CB assays. The FO-SPR VWF:GPIbR assay yielded an unstable signal, which was shown to be attributed to the co-factor function ristocetin, which did not result in a reliable signal in FO-SPR settings. We then produced mutant rGPIbα fragments that spontaneously bind to VWF. Proof-of-concept of this FO-SPR VWF:GPIbM assay was successful. For all three FO-SPR assays, excellent dose-responsive behavior was noticed with good lower range analytical sensitivity. Method comparison revealed good Spearman r coefficients, a significant relative bias between both VWF:Ag assays and no bias between both VWF:CB assays

Conclusions: We presented easy-to-perform assays to measure multiple VWF parameters. Because the FO-SPR platform allows connecting four fibers, this offers the possibility to determine a patient’s complete VWF status within a single run.

ESTIMATION OF PERIPHERAL LYMPHOCYTE SUBSETS REFERENCE INTERVALS FROM ROUTINE DATA USING THE REFINER ALGORITHM

B. Vanmechelen ¹, G. Frans¹, X. Bossuyt¹

¹UZ Leuven, Leuven, Belgium

Objective: To estimate local reference intervals (RIs) of peripheral blood lymphocyte subsets through retrospective analysis.

Material and Methods: Peripheral blood lymphocyte subsets extracted between January 2010 to November 2022 from the laboratory information system of the clinical laboratory of UZ Leuven. Patients with repeat measurements were excluded. The 2,5th and 97,5th percentiles RIs were calculated with the RefineR package in the R programming language [1]. Age partitioning was determined by visually inspecting the distribution and scatter plots for overall trends. The results were compared with RIs from published studies using the direct approach if available.

Results: A total of 15 355 samples were included to calculate the RIs. After visual inspection of the boxplots and the number of inclusions, it was decided to choose the following age categories for adults: 18-39y, 40-59y and 60-89y. The obtained indirect RIs were in the same order of magnitude as the direct RIs obtained from different representative studies. For the pediatric population the following categories were chosen: 0-3m, 3-12m, 1-2y, 3-5y, 6-11y and 12-17y. To date, no representative RIs studies using the direct approach are available for the pediatric population that comply with CLSI guidelines (CLSI EP28-A3c). The pediatric indirect RIs were compared with the current used direct RIs [2]. To enable comparison, the 5th and 95th percentiles were calculated. The greater the number of inclusions in the subgroups of the direct RI study, the better the obtained indirect RIs correlated.

Conclusion: The indirect approach to determine RIs serves as a practical complement to the direct method and can offer a feasible way for laboratories to estimate their laboratory specific RIs, especially for the pediatric population. However, the direct approach remains the golden standard because current indirect algorithms cannot fully deal with pathological population interference.

References

[1] T. Ammer, A. Schützenmeister, H. U. Prokosch, M. Rauh, C. M. Rank, and J. Zierk, “refineR: A Novel Algorithm for Reference Interval Estimation from Real-World Data,” Sci. Rep., vol. 11, no. 1, Dec. 2021.

[2] W. M. Comans-Bitter et al., “Immunophenotyping of blood lymphocytes in childhood: Reference values for lymphocyte subpopulations,” J. Pediatr., vol. 130, no. 3, pp. 388–393, 1997.

IMPLEMENTATION OF A LABORATORY WORKFLOW FOR OPTIMIZING SYNOVIAL LEUKOCYTE COUNT BASED ON SCATTERGRAM INTERPRETATION

M. Vanrenterghem ¹, J. Dom¹, N. Boeckx^1,2, M. Depypere^1,3, G. Frans¹, G. Vles^4,5, C. van Laer^1,6

¹Department of Laboratory Medicine, University Hospitals Leuven, Leuven, Belgium, ²Department of Oncology, Laboratory of Experimental Hematology, KU Leuven, Leuven, Belgium, ³Department of Microbiology, Immunology and Transplantation, Laboratory of Clinical Bacteriology and Mycology, KU Leuven, Leuven, Belgium, ⁴Department of Orthopaedic Surgery, University Hospitals Leuven, Leuven, Belgium, ⁵Institute for Orthopaedic Research and Training (IORT), Catholic University Leuven, Leuven, Belgium, ⁶Department of Cardiovascular Sciences, Centre for Molecular and Vascular Biology, University of Leuven, Leuven, Belgium

Objective: Falsely elevated synovial white blood counts (WBC) have been reported, particularly in the setting of joint arthroplasty.[1] We evaluated the implementation of a laboratory workflow based on Sysmex XN-1000 automated cell counting and scattergram interpretation for optimizing synovial cell counts.

Methods: Synovial fluid samples from 76 patients were collected, of which 29 from native joints and 47 from prosthesis patients. All scattergrams were evaluated by lab technicians for possible incorrect WBC and/or differential count: results ok (OK), only WBC count ok (WBC OK, DIFF NOT OK) or no results reliable (NOT OK). A specific finding were ‘banana-shaped’ scattergrams, which indicate possible interferences. Manual and automated cell counts were correlated using Passing-Bablok regression. The European Bone & Joint infection society (EBJIS) guidelines were applied to categorize prosthesis patients to identify possible prosthetic joint infections (PJI).[2]

Results: Lower manual WBC counts were seen in samples categorized as NOT OK in comparison with automated counts (Fig 1). The correlation between automated and manual WBC counts, calculated for samples with WBC count below 50 000/µL, was higher for native joints (r=0,938) in comparison with patients known with arthroplasty material (r=0,906). Scattergrams classified as OK showed overall a higher correlation compared to scattergrams which were interpreted as NOT OK. Banana-shaped scattergrams (n=19) showed falsely elevated WBC count and 17/19 (89%) patterns were seen in prosthesis patients. Six of 47 (13%) prosthesis patients were reclassified from ‘confirmed’ to ’unlikely’ PJI according to the EBJIS criteria.

Conclusion: Our workflow based on scattergram interpretation results in reporting correct WBC counts. ‘Banana-shaped’ scattergrams must be recognized because of WBC overestimation. Automated synovial WBC count can be used for patients with arthroplasty if the scattergram is visually acceptable.

References [1] Deirmengian et al., J Arthroplasty 2020 [2] McNally et al., Bone Joint J 2021

Figure 1: Correlation between manual and automated WBC count (µL)

THE VIRTUOUS CIRCLE OF APPROPRIATE LABORATORY TESTING: SUSTAINABILITY, PATIENT SAFETY, AND ECONOMIC IMPLICATIONS

L. Devis ¹, M. Closset¹, J. Degosserie¹, S. Lessire¹, P. Modrie¹, D. Gruson², E. J. Favaloro³, G. Lippi⁴, F. Mullier¹, E. Catry¹

¹CHU UCL Namur, Yvoir, Belgium, ²Cliniques Universitaires Saint-Luc, Brussels, Belgium, ³Institute of Clinical Pathology and Medical Research, Westmead Hospital, Sidney, Australia, ⁴University Hospital of Verona, Verona, Italy

Objective. To review the literature on the complex interplay between safety, economic and environmental factors, associated with inappropriate use of laboratory resources.

Material and Methods. We searched PubMed, Embase, Scopus, and Google scholar between April and September 2023 for articles from inception to present.

Results. We outline the impact of inappropriate laboratory testing on patient safety and economic outcomes. We examine the available literature on the environmental consequences of laboratory activity. We present a number of practical resolutions for reducing the environmental effects of clinical laboratory. We emphasize that decreasing unnecessary laboratory testing results in cost savings and environmental benefits, as evidenced by interventional studies, without compromising patient safety.

Conclusions. The implementation of sustainable practices in laboratories can create a virtuous circle in which reduced testing enhances cost-efficiency, reduces greenhouse gas emissions, and ensures patient safety, thereby benefiting the triple bottom line or ’3Ps’ (people, profit, and the planet). This review highlights the critical need for appropriate laboratory resource utilization in achieving sustainability in healthcare.

The authors have nothing to disclose.

ANALYTICAL PERFORMANCE OF LUMIRADX POINT OF CARE TESTING FOR TWO PARAMETERS OF CARDIOPULMONARY EMERGENCIES

N. Riahi ¹, R. Cupaiolo¹, L. Blairon¹

¹Iris Hospitals South, Brussels, Belgium

Background: Acute dyspnoea (AD) is one of the most frequent symptoms leading to the emergency department. The differential diagnosis of AD is complex, and requires relevant and rapid analysis of clinical, radiological and biological parameters. In this study we evaluate the performance of LumiraDX as a Point of Care Testing (POCT) for two parameters often used for the differential diagnostic of AD.

Materials and methods: Repeatability was evaluated by measuring two levels of internal quality control (IQC) for D-dimer and three levels for NT-proBNP.

Inter-assay precision was evaluated by analysing the same IQC twice a day, for fifteen days.

A total of 90 whole blood samples were collected (50 samples in a citrated tube for D-dimer measurement and 40 samples in a heparinised tube for NT-proBNP). For D-dimer, samples were analysed on both LumiraDX and Siemens CS-5100, while for NT-proBNP, results obtained on LumiraDX were compared with samples analysed on the cobas 8000 platform. Correlations were performed using a Bland Altman plot (BAP) and/or Passing Bablok linear regression (PBR).

Results:

Conclusion: Our results indicate that, for D-dimer and NT-proBNP, LumiraDx has good precision and demonstrates a good correlation with the routine methods around cut-offs, which can be interesting for ruling out in the case of AD.