Local quality assurance of serum or plasma (HIL) indices

Design, validation and performance of aspartate aminotransferase- and lactate dehydrogenase-reporting algorithms for haemolysed specimens including correction within quality specifications

BACKGROUND

In vitro haemolysis is a major operational challenge for medical laboratories. A new experimental design was used to investigate under what conditions algorithms could be designed to report either quantitative or qualitative aspartate aminotransferase and lactate dehydrogenase results outside the manufacturer's haemolysis specifications. Quantitative corrections were required to meet prespecified quality specifications.

METHODS

Twenty-five patient samples were used to design reporting algorithms and another 41 patient samples were used to validate the algorithms. Aspartate aminotransferase, lactate dehydrogenase and haemolysis index were determined using a Cobas 6000 analyser (Roche diagnostics, Mannheim, Germany). Correction factors were determined, and the accuracy of the correction was investigated. Reporting algorithms were designed based on (i) the manufacturer's cut-off for the haemolysis index, (ii) corrections within the total allowable error specification and (iii) qualitative reporting based on obtained results. The impact of the reporting algorithms was retrospectively determined by recalculating six months of aspartate aminotransferase and lactate dehydrogenase results.

RESULTS

No correction for aspartate aminotransferase/lactate dehydrogenase was possible for results below the upper reference interval limit, while results equal to or greater than the upper reference interval limit could, up to mild haemolysis, be corrected within the total error criterion. All samples generated from the validated patient cohort fulfilled the set criteria. The algorithms allowed reporting 88.5% and 85.9% of otherwise unreported aspartate aminotransferase and lactate dehydrogenase results, respectively.

CONCLUSIONS

An approach is presented that allows to generate and validate reporting algorithms for aspartate aminotransferase and lactate dehydrogenase compatible with prespecified quality specifications. The designed algorithms resulted in a significant reduction of otherwise unreported aspartate aminotransferase and lactate dehydrogenase results.

The preanalytical phase – from an instrument-centred to a patient-centred laboratory medicine

In order to guarantee patient safety, medical laboratories around the world strive to provide highest quality in the shortest amount of time. A major leap in quality improvement was achieved by aiming to avoid preanalytical errors within the total testing process. Although these errors were first described in the 1970s, it took additional years/decades for large-scale efforts, aiming to improve preanalytical quality by standardisation and/or harmonisation. Initially these initiatives were mostly on the local or national level. Aiming to fill this void, in 2011 the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) working group “Preanalytical Phase” (WG-PRE) was founded. In the 11 years of its existence this group was able to provide several recommendations on various preanalytical topics. One major achievement of the WG-PRE was the development of an European consensus guideline on venous blood collection. In recent years the definition of the preanalytical phase has been extended, including laboratory test selection, thereby opening a huge field for improvement, by implementing strategies to overcome misuse of laboratory testing, ideally with the support of artificial intelligence models. In this narrative review, we discuss important aspects and milestones in the endeavour of preanalytical process improvement, which would not have been possible without the support of the Clinical Chemistry and Laboratory Medicine (CCLM) journal, which was one of the first scientific journals recognising the importance of the preanalytical phase and its impact on laboratory testing quality and ultimately patient safety.

A Deep Learning-Based System for Assessment of Serum Quality Using Sample Images

BACKGROUND

Serum quality is an important factor in the pre-analytical phase of laboratory analysis. Visual inspection of serum quality (including recognition of hemolysis, icterus, and lipemia) is widely used in clinical laboratories but is time-consuming, subjective, and prone to errors.

METHODS

Deep learning models were trained using a dataset of 16427 centrifuged blood images with known serum indices values (including hemolytic index, icteric index, and lipemic index) and their performance was evaluated by five-fold cross-validation. Models were developed for recognizing qualified, unqualified and image-interfered samples, predicting serum indices values, and finally composed into a deep learning-based system for the automatic assessment of serum quality.

RESULTS

The area under the receiver operating characteristic curve (AUC) of the developed model for recognizing qualified, unqualified and image-interfered samples was 0.987, 0.983, and 0.999 respectively. As for subclassification of hemolysis, icterus, and lipemia, the AUCs were 0.989, 0.996, and 0.993. For serum indices and total bilirubin predictions, the Pearson's correlation coefficients (PCCs) of the developed model were 0.840, 0.963, 0.854, and 0.953 respectively. Moreover, 30.8% of serum indices tests were deemed unnecessary due to the preliminary application of the deep learning-based system.

CONCLUSIONS

The deep learning-based system is suitable for the assessment of serum quality and holds the potential to be used as an accurate, efficient, and rarely interfered solution in clinical laboratories.

External quality assessment of serum indices: Spanish SEQC-ML program

OBJECTIVES

Serum indices included in clinical chemistry instruments are widely used by laboratories to assess the quality of samples. Instruments that report quantitative results allow an evaluation of their diagnostic performance in a similar way to other biochemical tests. The Spanish Society of Laboratory Medicine (SEQC-ML) launched a monthly External Quality program of serum indices in 2018 using three lyophilized materials of simultaneous annual distribution. We present the results of the first three years of the program.

METHODS

The use of four different quality control materials with different concentrations in three alternate months allows an annual evaluation of the participant's accuracy. Assigned values are established by consensus among homogeneous groups, considering necessary at least 10 participants for a comparison at instrument level. The average percentage difference results per instrument allow the assessment of bias among groups.

RESULTS

The imprecision of the three indices ranges between 3 and 9%, with no major differences among instruments. Significant differences were observed in all indices among instruments with more than 10 participants (Roche Cobas, Abbott Architect, Abbott Alinity and Siemens Advia). The 90th percentile of the distribution of percentage differences was used as the analytical performance specification (APS). An improvement in performance was observed in the first three years of the program, probably due to the learning curve effect. In 2020, APS of 7.8, 12.2 and 9.7% were proposed for hemolytic, icteric and lipemic indices, respectively.

CONCLUSIONS

Serum indices have a great impact on the quality and the reliability of laboratory test results. Participation in proficiency testing programs for serum indices is helpful to encourage harmonization among providers and laboratories.

Using the hemolysis index of Abbott's Alinity c for the measurement of plasma free hemoglobin in ECMO patients

OBJECTIVES

Quantitative measurement of plasma free hemoglobin (fHb) concentrations is essential for monitoring pediatric ECMO patients, since hemolysis has a great impact on the patient's clinical outcome. The aim of this study was to validate the hemolysis index (HI) assay on Abbott's Alinity c system as a quantitative method to measure fHb.

METHODS

The performance of the HI assay, based on an automated spectrophotometric method recording the absorption at four different wavelength pairs, was evaluated using the 20 × 2 × 2 design according to the CLSI-EP05-A3 guidelines. LLOQ and LLOD were calculated according to CLSI-EP17 guidelines with CVs set to 10% and 20%, respectively. Furthermore, the method was tested for interferences with bilirubin and Intralipid®.

RESULTS

Linearity was ensured over an analytical measurement range of 30-7250 mg/L and the calculated LLOQ and LLOD were 80 mg/L and 50 mg/L, respectively. Intra-run and total imprecisions ranged from 0.9-3.4% and 1.0-3.4%, respectively. The HI assay correlated well with the Harboe method (HI (mg/L) = 0.998 * fHb (mg/L) + 28 mg/L, R = 0.998, n = 50) and interference testing showed no impact of bilirubin and Intralipid® up to 709 mg/L and 5580 mg/L, respectively.

CONCLUSIONS

The HI assay on Abbott's Alinity c system allows a precise and accurate determination of fHb concentrations with no significant interferences in a simple, rapid and cost-effective way.

Frequency of icteric interference in clinical chemistry laboratory tests and causes of severe icterus

Objectives

The aims of this study were to identify the causes of severe icterus in an academic medical center patient population and to assess the impact of icterus on clinical chemistry testing using assay package insert thresholds.

Design

and Methods: In this retrospective study at an academic medical center core clinical laboratory, icteric, hemolysis, and lipemia indices were available for all serum and plasma chemistry specimens analyzed on Roche Diagnostics cobas 8000 analyzers over a 12-month period, encompassing 414,502 specimens from 94,081 unique patients (51,851 females; 42,230 males) including children, inpatient, outpatient, and emergency department patients. Extensive chart review was done for all 57 patients (4 pediatric, 53 adult; 534 total specimens) who had one or more samples with an icteric index of 40 or higher (defined as severe icterus).

Results

Specimen icteric index exceeded package insert icteric index thresholds in 0.14% of clinical chemistry assays, with the highest number of instances for creatinine (1358 samples, 0.6% of total tests), total protein (1194 samples, 2.2%), and ammonia (161 samples, 3.9%). The 57 patients with an icteric index of 40 or higher accounted for 49.7% of all instances where the icteric index exceeded the specific assay package insert limit. The most common etiologies of this group of 57 patients were alcohol-related liver disease (34 patients), biliary tract disease (7 patients), and neoplasms (6 patients).

Conclusions

Approximately half of all instances where specimen icteric index exceeded assay package insert thresholds occurred in a small cohort of patients with severe liver/biliary tract disease.

•

Specimen icteric indices exceeded package insert icteric index thresholds for 0.14% of clinical chemistry tests ordered at an academic medical center.

•

Icteric interference exceeding package insert thresholds had the most overall occurrences for enzymatic creatinine, total protein, and ammonia.

•

Only 57 of 94,081 unique patients (0.06%) with severe icterus accounted for nearly half of instances where the icteric index exceeded the package insert limit for a specific assay.

•

The small cohort of patients with severe icterus had high mortality generally associated with cirrhosis, biliary disease, or aggressive metastatic cancer.

•

Icteric indices exceeding 20 were mostly comprised of patients with predominantly conjugated hyperbilirubinemia.

How to meet ISO15189:2012 pre-analytical requirements in clinical laboratories? A consensus document by the EFLM WG-PRE

The International Organization for Standardization (ISO) 15189:2012 standard aims to improve quality in medical laboratories through standardization of all key elements in the total testing process, including the pre-analytical phase. It is hence essential that accreditation bodies, assessing laboratories against ISO15189:2012, pay sufficient attention to auditing pre-analytical activities. However, there are significant differences in how technical auditors interpret the pre-analytical requirements described in ISO15189:2012. In this consensus document, the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) Working Group for Pre-analytical Phase (WG-PRE) sets out to review pre-analytical requirements contained in ISO15189:2012 and provide guidance for laboratories on how to meet these requirements. The target audience for this consensus document is laboratory professionals who wish to improve the quality of the pre-analytical phase in their laboratory. For each of the ISO requirements described in ISO15189:2012, members of EFLM WG-PRE agreed by consensus on minimal recommendations and best-in-class solutions. The minimal consensus recommendation was defined as the minimal specification which laboratories should implement in their quality management system to adequately address the pre-analytical requirement described in ISO15189:2012. The best-in-class solution describes the current state-of-the-art in fulfilling a particular pre-analytical requirement in ISO15189:2012. We fully acknowledge that not every laboratory has the means to implement these best-in-class solutions, but we hope to challenge laboratories in critically evaluating and improving their current procedures by providing this expanded guidance.

Hemolytic specimens in complete blood cell count: Red cell parameters could be revised by plasma free hemoglobin

Introduction

Hemolysis is the main cause of unqualified clinical samples. In this study, we established a method for detecting and evaluating hemolysis in whole blood test. We used a mathematical formula for correcting the influence of hemolysis on complete blood cell count (CBC) so as to avoid re‐venipuncture and obtain more accurate parameters of red blood cell detection, reduce the burden of patients, and improve the efficiency of diagnosis and treatment.

Methods

Hemolytic samples were selected and then corrected using the new formula. Plasma free hemoglobin (fHB) was used as the criterion to determine the degree of hemolysis; the uncertainty of measurement is acceptable as the limit value of deviation between the measured value and the revised value. Hemolysis simulation analysis in vitro and continuous monitoring of clinical patients were used to verify the correction effect.

Results

A total of 83 clinical samples with hemolysis were collected and analyzed; fHB 1.4 g/L was selected as the unacceptable value for clinical hemolysis detection. In hemolytic samples, the red blood cell parameters corrected by formula are significantly different from those uncorrected and had a good consistency with those before hemolysis.

Conclusion

The results show that the hemolysis phenomenon of CBC has a significant impact on routine blood testing. By using the new formula, the influence of hemolysis on erythrocyte and related parameters can be quickly and easily corrected, thus avoiding venipuncture again for re‐examination, reducing diagnostic errors, and saving medical resources.

Managing hemolyzed samples in clinical laboratories

Hemolysis is conventionally defined as membrane disruption of red blood cells and other blood cells that is accompanied by subsequent release of intracellular components into the serum or plasma. It accounts for over 60% of blood sample rejections in the laboratory and is the most common preanalytical error in laboratory medicine. Hemolysis can occur both in vivo and in vitro. Intravascular hemolysis (in vivo) is always associated with an underlying pathological condition or disease, and thus careful steps should always be taken by the laboratory to exclude in vivo hemolysis with confidence. In vitro hemolysis, on the other hand, is highly preventable. It may occur at all stages of the preanalytical phase (i.e. sample collection, transport, handling and storage), and may lead to clinically relevant, yet spurious, changes in patient results by interfering with laboratory measurements. Hemolysis interference is exerted through several mechanisms: (1) spectrophotometric interference, (2) release of intracellular components, (3) sample dilution and (4) chemical interference. The degree of interference observed depends on the level of hemolysis and also on the assay methodology. Recent evidence shows that preanalytical practices related to detection and management of hemolyzed samples are highly heterogeneous and need to be standardized. The Working Group for Preanalytical Phase (WG-PRE) of the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) has published many recommendations for facilitating standardization and improvement of this important preanalytical issue. Some key EFLM WG-PRE publications related to hemolysis involve: (i) a call for more transparency and some practical recommendations for improving the harmonization of the automatic assessment of serum indices and their clinical usefulness, specifically the hemolysis index (H-index), (ii) recommendations on how to manage local quality assurance of serum or plasma hemolysis/icterus/lipemia-indices (HIL-indices) and (iii) recommendations on how to detect and manage hemolyzed samples in clinical chemistry testing. In this review we provide a comprehensive overview of hemolysis, including its causes and effects on clinical laboratory assays. Furthermore, we list and discuss the most recent recommendations aimed at managing hemolyzed samples in everyday practice. Given the high prevalence of hemolyzed blood samples, the associated costs, the great heterogeneity in how hemolysis is handled across healthcare settings, countries and continents, and increasing patient cross-border mobility, standardization and quality improvement processes aimed at combatting this important preanalytical problem are clearly warranted.

Influence of hypertriglyceridemia, hyperbilirubinemia and hemolysis on thrombin generation in human plasma

Background

Although accumulating evidence suggests that the hemostatic balance is impaired in patients with hypertriglyceridemia, hyperbilirubinemia or hemolytic anemias, little is known on the underlying biological mechanisms. This experimental study was aimed at exploring whether increasing values of triglycerides, bilirubin or cell-free hemoglobin promote thrombin generation in plasma.

Methods

Three different pools were prepared from three different sets of 20 normal routine plasma citrate samples. The native pools were spiked with increasing amounts of exogenous triglycerides (up to 8.8 mmol/L), bilirubin (up to 350 μmol/L) or autologous hemolyzed blood (up to 3.5 g/L cell-free hemoglobin). Using the fully-automated thrombin generation analyzer ST Genesia, we measured the following parameters: lag time (LT), time to peak (TP), peak height (PH) and endogenous thrombin potential (ETP).

Results

A sustained increase of PH and ETP was found in parallel with increasing triglyceride concentrations, peaking in the aliquot with 8.8 mmol/L. Conversely, LT and TP displayed an opposite trend, reaching a maximum decrease in the 8.8 mmol/L aliquot. Increasing bilirubin concentrations promoted remarkable increases of PH and ETP and decreases of TP and LT, up to 211 μmol/L. After this threshold, all parameters tended to return towards baseline values. A constant increase of PH and ETP was also noted in hemolyzed samples, peaking in the 3.5 g/L cell-free hemoglobin aliquot, whereas the TP and LT remained unchanged in all hemolyzed aliquots.

Conclusions

Our findings suggest that hypertriglyceridemia, hyperbilirubinemia and hemolysis may promote a hypercoagulable state in human plasma.

Preanalytical challenges – time for solutions

The European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) Working Group for the Preanalytical Phase (WG-PRE) was originally established in 2013, with the main aims of (i) promoting the importance of quality in the preanalytical phase of the testing process, (ii) establishing best practices and providing guidance for critical activities in the preanalytical phase, (iii) developing and disseminating European surveys for exploring practices concerning preanalytical issues, (iv) organizing meetings, workshops, webinars or specific training courses on preanalytical issues. As education is a core activity of the WG-PRE, a series of European conferences have been organized every second year across Europe. This collective article summarizes the leading concepts expressed during the lectures of the fifth EFLM Preanalytical Conference “Preanalytical Challenges – Time for solutions”, held in Zagreb, 22–23 March, 2019. The topics covered include sample stability, preanalytical challenges in hematology testing, feces analysis, bio-banking, liquid profiling, mass spectrometry, next generation sequencing, laboratory automation, the importance of knowing and measuring the exact sampling time, technology aids in managing inappropriate utilization of laboratory resources, management of hemolyzed samples and preanalytical quality indicators.

Internal quality assurance of HIL indices on Roche Cobas c702

Automatic assessment of hemoglobin (H), lipaemia (L) and icterus (I) in serum or plasma (HIL indices) is the mainstay for evaluating sample quality. We planned this study to verify whether in-house prepared internal quality control (IQC) materials may be suitable for quality assurance of HIL indices. Pools containing different values of each of the three HIL indices were prepared from routine plasma samples, divided in aliquots and frozen at -20°C. Stability of frozen materials was assessed by thawing one aliquot of each pool after different days of freezing (1, 4, 8, 15, 22 and 29), and by measuring HIL indices on baseline fresh samples and frozen-thawed aliquots with Roche Cobas c702. Five fresh liquid IQCs materials were also measured at the same time points. Intra-assay and inter-assay imprecision of HIL indices calculated with commercial IQC materials ranged between 1.1-2.0% and 1.6-3.3%, respectively. When target values of HIL indices were calculated using frozen-thawed aliquots, the inter-assay imprecision of in-house prepared materials was optimal, even better than that of commercial liquid IQCs (H-index, 0.8% versus 1.6%; L-index, 2.2% versus 2.5%; I-index, 0.8% versus 3.3%). In conclusion, in-house prepared IQC materials are cost-effective alternatives to commercial liquid IQCs for HIL quality assurance.

Haemolysis index for the screening of intravascular haemolysis: a novel diagnostic opportunity?

The diagnostic approach to patients with intravascular haemolysis remains challenging, since no first-line laboratory test seems to be entirely suitable for the screening of this condition. Recent evidence shows that an enhanced cell-free haemoglobin (fHb) concentration in serum or plasma is a reliable marker of red blood cell injury, and may also predict clinical outcomes in patients with different forms of haemolytic anaemias. However, the routine use of the haemiglobincyanide assay, the current reference method for measuring fHb, seems unsuitable for a timely diagnosis of intravascular haemolysis, for many safety and practical reasons. The spectrophotometric assessment of fHb by means of the so-called haemolysis-index (H-index) has now become available in most clinical chemistry analysers. This measure allows an accurate, rapid and inexpensive assessment of fHb in a large number of serum or plasma samples, and its use has already proven to be useful for identifying some forms of haemolytic anaemias. Therefore, the aim of this article is to provide an update and a personal opinion about the potential clinical use of the H-index for screening patients with suspected intravascular haemolysis.