|Year : 2023 | Volume
| Issue : 1 | Page : 55-60
Study of analytical error in lipase assay
Sweta Kumari, Ravi Shekhar, Rekha Kumari, Pritam Prakash
Department of Biochemistry, Indira Gandhi Institute of Medical Sciences, Patna, Bihar, India
|Date of Submission||30-Jun-2021|
|Date of Decision||18-Mar-2022|
|Date of Acceptance||26-Apr-2022|
|Date of Web Publication||24-Jan-2023|
Department of Biochemistry, Indira Gandhi Institute of Medical Sciences, Patna - 800 014, Bihar
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Background: Lipase and amylase are the most frequently used biomarker for the diagnosis of pancreatic diseases, especially acute pancreatitis. Lipase has better diagnostic accuracy in comparison to amylase for the analysis of acute pancreatitis. However, lipase assay in random access analyzer is sometimes difficult to perform as it is exposed to different types of samples or reagents which may act as interference. Materials and Methods: In our laboratory, we found the raised values (>500 IU/L) of lipase with normal amylase in some samples. However, the immediate rerun of these samples for lipase only showed normal (<80 IU/L) lipase level. To root out this fallacy, we performed reagent and sample carryover studies. Results: The cause of the falsely raised value of lipase was revealed by reagent carryover studies. All samples which assayed triglyceride (TGL) followed by lipase immediately after it showed elevated (>500 IU/L) lipase value. This is due to the interference of microbial lipase used in TGL reagents. This was corrected by separating the analysis of lipase and TGL into two different instruments. Conclusion: If interference testing is not done, the laboratories are prone to have an analytical error in reporting and hence lead to diagnostic error. Hence, after analyzer installation, interference testing should be included in the validation protocol.
| Abstract in French|| |
Contexte: La lipase et l'amylase sont les biomarqueurs les plus fréquemment utilisés pour le diagnostic des maladies pancréatiques, en particulier la pancréatite aiguë. La lipase a une meilleure précision diagnostique par rapport à l'amylase pour l'analyse de la pancréatite aiguë. Cependant, le dosage de la lipase dans un analyseur à accès aléatoire est parfois difficile à réaliser car il est exposé à différents types d'échantillons ou de réactifs qui peuvent agir comme des interférences. Matériels et méthodes: Dans notre laboratoire, nous avons trouvé des valeurs élevées (> 500 UI/L) de lipase avec une amylase normale dans certains échantillons. Cependant, la réexécution immédiate de ces échantillons pour la lipase n'a montré qu'un niveau de lipase normal (<80 UI/L). Pour éliminer cette erreur, nous avons effectué des études de transfert de réactifs et d'échantillons. Résultats: La cause de la valeur faussement élevée de la lipase a été révélée par des études de transfert de réactif. Tous les échantillons qui dosaient les triglycérides (TGL) suivis de la lipase immédiatement après présentaient une valeur de lipase élevée (> 500 UI/L). Cela est dû à l'interférence de la lipase microbienne utilisée dans les réactifs TGL. Cela a été corrigé en séparant l'analyse de la lipase et de la TGL dans deux instruments différents. Conclusion: Si les tests d'interférence ne sont pas effectués, les laboratoires sont susceptibles d'avoir une erreur analytique dans les rapports et donc de conduire à une erreur de diagnostic. Par conséquent, après l'installation de l'analyseur, les tests d'interférence doivent être inclus dans le protocole de validation.
Mots-clés: Pancréatite aiguë, amylase, erreur diagnostique, laboratoires, lipase
Keywords: Acute pancreatitis, amylase, diagnostic error, laboratories, lipase
|How to cite this article:|
Kumari S, Shekhar R, Kumari R, Prakash P. Study of analytical error in lipase assay. Ann Afr Med 2023;22:55-60
| Introduction|| |
Acute pancreatitis is the inflammation of the pancreas that develops quickly and could be fatal if not promptly treated. Several research groups recommended the presence of two of the three criteria for the diagnosis and management of acute pancreatitis. They are characteristic abdominal pain, typical computed tomography finding, and serum amylase and/or lipase level ≥3 times the upper limit of normal.,, In this study, sensitivity and specificity of lipase for the diagnosis of acute pancreatitis were 96.6% and 99.4%, respectively. In contrast, the sensitivity and specificity of amylase for the diagnosis of acute pancreatitis were 78.6% and 99.1%, respectively., Lipase (EC 126.96.36.199; triacylglycerol acylhydrolase: LIP) offers a wider diagnostic window in comparison to amylase as lipase activity increases within 3–6 h of acute attack of pancreatitis, peaks in 24 h, and remain elevated for 8–14 days. Whereas amylase level elevates within 5–8 h of symptom onset, attains maximal level in 12–72 h, and returns to normal by the 4–7 day., Lipase has increased sensitivity in the diagnosis of acute pancreatitis due to alcohol consumption and also in patients presenting later in their clinical course.
The measurement of lipase activity has been the most sensitive and specific marker for the diagnosis of pancreatic injury. The interpretation of this test sometimes can be difficult since several nonpancreatic conditions can present with abnormal serum lipase levels. There could be falsely elevated values of lipase that may result in misdiagnosis and wrong treatment.
With the increasing dependency on diagnostic tests for clinical diagnosis, the medical laboratories are confronted with problems such as analyzing a larger number of analytes at a time, maintaining faster Turn around time (TAT), and reducing the number of trained laboratory personnel. Hence, laboratories rely on an automatic analyzer to meet the above needs. Analyzers normally use 1 or 2 probes that are exposed to different varieties of reagents/samples in quick succession. Similarly, cuvettes also come in contact with different reagents/samples. This may result in carryover interference. It usually occurs when the analyte from a high-concentration sample/reagent is incompletely removed by the analytical system washing process, either probe mixer or cuvette washing.
In our hospital laboratory, the newly installed biochemistry analyzer was giving erroneously high values of lipase with normal serum amylase in otherwise normal patients. On repeating the sample for lipase only, the values came to normal. This episode occurred intermittently with few samples. High lipase values were found irrespective of the triglyceride (TGL) level; however, most of the TGL values were within the normal range.
The study is aimed to find out the reason for falsely elevated serum lipase and to take corrective action for it. The study also intended to exclude the chances of carryover from different reagents or specimens in serum lipase assay.
| Materials and Methods|| |
It is a laboratory-based study in which 75 samples of elevated serum lipase with normal serum amylase were rerun for the serum lipase only. Samples showing a change in lipase values on rerun were assessed for any interference or carryover. All the tests were performed on the biochemistry automated analyzers Beckman Coulter DxC 700 AU and AU5800, using reagents from the same company. Two levels of Bio-Rad internal quality controls are run daily. The statistical analysis was done using software of IBM SPSS Statistics for Windows, version 20 (IBM Corp., Armonk, N.Y., USA).
To perform a sample carryover study, a sample pool of normal lipase values was run five times. Then, a sample pool of very high lipase values was run five times. Then, the previous normal pool sample was run again five times. The mean values of normal lipase sample pool before and after the run of high lipase sample pool would be compared for significant interference. Interference of TGL reagent in lipase assay was mentioned in its pack insert provided by the manufacturer. In the study, all samples which showed an appreciable change in lipase values on rerun were also assayed for TGL in their first run. The reason for this interference could be the carryover of TGL reagent which contains a large amount of lipase. A carryover study of TGL reagent on lipase assay was performed. The protocol for this was taken from Boneno et al. work on reagent carryover studies. The scheme was as [Table 1]:
This same scheme was used after the corrective action to check its effectiveness and measure any residual carryover.
Here, Bio-Rad controls were run as samples. Saline was used as both sample and reagent in the case of saline.
Ethical permission for this study was taken from the Institutional Research Ethics Committee.
| Results|| |
All the physical conditions of the laboratory were within acceptable limits and did not indicate a problem which may affect the results. The temperature of the laboratory was maintained at 22.0° ± 1°C. The lamp energy of the analyzers was monitored daily. The incubation bath temperature was sustained at 37.0°C and that of coolant at 7.8°C. The conductivity of water was 0.023 ± 0.011 μS/cm.
Seventy-five samples of elevated serum lipase with normal serum amylase were rerun for lipase alone on two different instruments DxC 700 AU and AU5800, randomly. All the values were used to draw box and whisker plot – the first box for raised values of lipase, the second for serum lipase values repeated on DxC 700 AU, and the third for lipase values repeated on AU5800.
[Figure 1] shows lipase values in the first run and then repeated in different instruments randomly. The median value of lipase was 497 IU/L for all the samples of raised lipase with normal amylase at the first run. While samples that were repeated on AU700 have the median lipase of 52 IU/L, a difference of 89% and those repeated on AU5800 have a median of 44 IU/L, a difference of 91% from the median of the first value. To carry out a sample carryover study, a sample pool of normal lipase was run five times. The mean value was 36 ± 1.48 IU/L and CV = 4.4%. Then, a sample pool of very high lipase value was run five times. The mean value was 1200 ± 28.3 IU/L and CV was 2.6%. The previous pool sample of normal lipase level was run again five times. The mean value was 38 IU/L. The percentage difference between the two values of lipase was 5.56% which is <10% and is acceptable.
|Figure 1: Comparison of median values of lipase on the first run and repeated run|
Click here to view
In this reagent carryover study, a sample pool was run for lipase in combination with each one of the parameters being run in the analyzer in alternating fashion, for example, lipase, serum glutamic pyruvic transaminase, lipase, serum glutamic-oxaloacetic transaminase, lipase, albumin, and so on to observe a difference in the value. There was a significant increase observed in the value of lipase after TGL. Once the donor reagent was found, a specific reagent carryover study was done as per the protocol mentioned earlier. The results were tabulated in [Table 2] and [Figure 2] as the following:
At first, the Quality Control (QC) samples were run for lipase five times as a sample. The mean value was 93.16 ± 2.1 IU/L and CV was 2.25%. Then, the same QC sample was run for TGL five times. The reason for running TGL is to contaminate the probes or cuvettes with the probable interfering agent, which in this case is the TGL reagent having microbial lipase. The mean value of TGL was 126 mg/dl. Then, the same QC was run again for lipase five times. The first lipase value after the TGL run was 139.6 IU/L that was significantly increased from the target mean value; there was a difference of 49.8% from the mean value of lipase. The rest of the lipase value was close to the mean having a difference of <10%. This clearly shows the interference of TGL reagent on lipase analysis.
To prevent the interference, corrective action was taken and the carryover study was repeated as per the protocol proposed by Joseph Boneno et al. The results are tabulated in [Table 3].
|Table 3: Result of reagent carry-over study on lipase after decontaminating with cleaning solution ('Contamination prevention parameter' setting)|
Click here to view
In [Table 3] mean value of lipase before contamination was 94.16 ± 2.3 IU/L. After that, TGL was run, and later on, three cycles of cleaning solution were also run. Lipase values run after cleaning solution were close to the initial mean value having a difference of <5%. This shows that the effect of TGL reagent was overcome by the use of cleaning solution.
| Discussion|| |
Haeckel described carryover as a process by which materials are carried into a reaction mixture to which they do not belong. These materials can be either part of a specimen or reagents including the diluent or wash solution.” Carryover can be classified by material (e.g., specimen, diluent, reagent, reaction mixture, and wash solution) or by the site of carryover (e.g., specimen cup, sample probe, reagent probe, and wash station). The following types of carryovers have been observed in practice: (i) specimen carryover in a sample probe, (ii) carryover from diluent to specimen in a specimen cup, (iii) carryover from the reaction mixture to the reaction mixture, and (iv) carryover from reagent to the reaction mixture from a reagent probe. Haeckel proposed various formulae for the measurement of carryover and describes various specific scenarios that should be considered. He concludes by noting that “new analytical systems should be designed to avoid any type of carryover as far as possible.”,
The analysis of TGL is based on the series of coupled enzymatic reactions. In the first reaction, TGL is hydrolyzed by microbial lipase into glycerol and three fatty acids. The glycerol is phosphorylated by adenosine triphosphate in the presence of glycerol kinase to produce glycerol-3-phosphate. In the concluding reaction, hydrogen peroxide (H2O2) reacts with 4-aminophenazone and N, N-Bis (4-sulfobutyl)-3,5-dimethylaniline, disodium salt in the presence of peroxidase to produce a chromophore, the absorbance of which at 660/800 nm is proportional to the TGL content of the sample.
The lipase used in above-mentioned TGL reagent, if not washed out completely from the probes and cuvettes, poses serious interference in lipase analysis. This interference occurs if lipase analysis is done immediately after TGL testing. The frequency of error in lipase values was found irrespective of the TGL level. There was no significant difference in the frequency of this episode among the high, low, or normal level of TGL. [Table 2] and [Figure 2] show that TGL reagent was interfering with the results of lipase and the difference was 48.9% which was statistically as well as clinically significant. Hence, the manufacturer's package inserts also recommended to exclude running TGL assays just before a lipase assay and utilizing suggested measures for contamination parameters.
Initially, the measures to reduce the interference in lipase assay were not being implemented in our biochemistry laboratory during installation and validation, as we have two biochemistry analyzers out of which DxC 700 AU was used for plasma glucose, HbA1c, and serum lipase. The rest of the parameters were being run on AU5800 instrument. Thus, lipase and TGL along with all lipid profile parameters were running on two different instruments. Later on, as the institute sample load increased, all the test parameters were started on both the instruments which led to running TGL and lipase assay and in AU5800,assays further separated by assigning reactions of these two in different cuvette rings. [Table 3] shows that the interference of TGL reagent was nullified by decontamination using explicit cleaning solution. These alterations not just helped us in releasing accurate lipase reports but also prevented the delay and reagent wastage due to repeat testing.
Beckman Coulter provides a list of “contamination prevention parameters” specific for the different sets of instruments. For carryover of TGL reagent to lipase assay in DxC 700 AU, the company recommends cleaning solution ODR 20067, which should be replaced weekly or as required. Three wash cycles should be programmed between the analysis of TGL and lipase test. In AU5800 instrument, the TGL and lipase assay could be separated by designating them in two different cuvette rings. These rings have different sets of reagents and sample probes and thus minimizing the interference. However, this measure could be taken on AU5800 only as it has two different rings for the placement of reagents, and both rings have their own reagent and sample probe. Beckman Coulter advises the programming of these donor-recipient tests in such fashion that recipient is the first test to be run, followed by the donor. The TGL assay running just before a lipase assay is excluded from programming.
A similar type of reagent carryover incidence had been mentioned by Patrick B Kyle in 2009. He observed a significant decrease (≤100 mg/dl) in serum cholesterol values run after creatine kinase (CK). During the cholesterol test procedure, H2O2 is produced which oxidatively converts 4-aminophenazone and phenol to a red quinone-imine dye, which is directly proportional to the cholesterol concentration. The CK reagent contains N-acetylcysteine which acts as a scavenger for reactive oxygen and inactivates hydroxyl radicals, superoxide, and and H2O2. Carryover of CK reagent having N-acetylcysteine into assay for cholesterol would decompose the H2O2 which was released during the reaction, and thus decreased the level of cholesterol. This problem which seemed to occur intermittently was eliminated by the replacement of the reagent probe.
A carryover study was done as early as 1988 was reported by Krouwer JS, Stewart WN, Schlain B when he proposed a 12-sample multifactor protocol as a screening method to evaluate random access analyzers for imprecision, nonlinearity, linear drift, and reagent carryover. These authors recognized that random access instruments differed from batch analyzers in that they are susceptible to reagent cross-contamination in addition to sample carryover. They found evidence of reagent carryover in aspartate transaminase (AST)/ Lactate Dehydrogenase (LD) pair in which the donor was AST assay and the recipient LD assay. This is because the AST reagent contained LD in high concentrations.
Bernhard Frank and Klaus Gottlie in 1992 explained several conditions with raised lipase and normal amylase, for example, renal insufficiency, nonpancreatic sources of lipolytic enzymes due to malignant tumor, acute cholecystitis or esophagitis, delayed blood withdrawal, hypertriglyceridemia, or subclinical pancreatitis in patients without abdominal pain. Isolated raised lipase with normal amylase should not be considered evidence of acute pancreatitis; hence both amylase and lipase should be tested simultaneously in the presence of characteristic abdominal pain.
David A. Armbruster in 2006 performed sample carryover studies on the Abbott ci8200 integrated clinical chemistry/immunoassay systems to assess whether the clinical chemistry module causes sample carryover in immunoassay analysis. He observed sample carryover of HBsAg, alpha feto protein (AFP), β-Human chorionic gonadotropin (hCG), and other analytes were <0.1 ppm or below the assay limit of detection.
Joseph Boneno in his reagent carryover studies evaluated reagent cross-contamination on the Abbott c8000 analyzer. He tested 32 Abbott clinical chemistry reagents and 15 specific protein and other assays from EQual Diagnostics. He found >10% difference in the target value of the recipient analytes for donor/recipient pair such as ferritin/alkaline phosphatase, C-reactive protein/immunoglobulin (IgM), and IgA/CO2. Discrete analyzers are classified as “closed” or “open.” In closed systems, only tests for which the manufacturer offers reagents can be performed. Open systems are more flexible and reagents from different manufacturers may be used. In closed systems, manufacturers that provide both analyzers and reagents perform reagent carryover studies to ensure that test results are not affected, so laboratories working on open systems using third-party reagents should always perform carryover study before starting the test.
P. M. G. Broughton in 1984 mentioned the methods for the measurement of carryover in continuous flow and discrete analyzers and also discussed their effects on analytical precision. He put forward the different formulas for quantification of sample and reagent carryover, respectively.
The purpose of this article was to make the new user aware of the error experienced by our laboratory. Carryover testing should be performed during validation before starting a new parameter. By employing the methods cited in this article as well as respective manufacturers' guidelines, the laboratories will be able to report accurate lipase results. This will avoid multiple repeat testing of lipase leading to less reagent consumption and improved TAT.
| Conclusion|| |
Laboratories, especially those using the open system with third-party reagents, should remain vigilant for possible interference on results due to carryover. Once the evidence of carryover is found, as in this study, TGL reagent carryover found in lipase assay, the acceptable limit is set so that it might be analytically significant but must be clinically insignificant. Suitable measures such as decontamination with cleaning solution, physically separating the two reagents in two different rings or instruments, and programming for performing the recipient test first followed by donor test should be taken to check the carryover to its acceptable limit. Manufacturers' recommendation for contamination prevention should be implemented before starting a new test. Laboratories cannot completely eliminate the carryover but judicious application of laboratory practices and clinical insight, amount of it can be well-tolerated.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Hameed AM, Lam VW, Pleass HC. Significant elevations of serum lipase not caused by pancreatitis: A systematic review. HPB (Oxford) 2015;17:99-112.
Banks PA, Freeman ML. Practice parameters committee of the american college of gastroenterology. Practice guidelines in acute pancreatitis. Am J Gastroenterol 2006;101:2379-400.
Pezzilli R, Zerbi A, Di Carlo V, Bassi C, Delle Fave GF. Working group of the Italian association for the study of the pancreas on acute pancreatitis. Practical guidelines for acute pancreatitis. Pancreatology 2010;10:523-35.
Ismail OZ, Bhayana V. Lipase or amylase for the diagnosis of acute pancreatitis? Clin Biochem 2017;50:1275-80.
Gomez D, Addison A, De Rosa A, Brooks A, Cameron IC. Retrospective study of patients with acute pancreatitis: Is serum amylase still required? BMJ Open 2012;2:e001471.
Batra HS, Kumar A, Saha TK, Misra P, Ambade V. Comparative study of serum amylase and lipase in acute pancreatitis patients. Indian J Clin Biochem 2015;30:230-3.
Lippi G, Valentino M, Cervellin G. Laboratory diagnosis of acute pancreatitis: In search of the Holy Grail. Crit Rev Clin Lab Sci 2012;49:18-31.
Gumaste V, Dave P, Sereny G. Serum lipase: A better test to diagnose acute alcoholic pancreatitis. Am J Med 1992;92:239-42.
Boneno J, Fokakis M, Armbruster D. Reagent carryover studies: Preventing analytical error with open clinical chemistry systems. Lab Med 2005;36:705-10.
Haeckel R. Proposals for the description and measurement of carry-over effects in clinical chemistry. Pure Appl Chem 1991;63:302-6.
Haeckel R. Recommendations for definition and determination of carry-over effects. J Automat Chem 1988;10:181-3.
Kyle PB. Beware of carryover in modern chemistry analyzers. Clin Chem Lab Med 2010;48:519-21.
Krouwer JS, Stewart WN, Schlain B. A multi-factor experimental design for evaluating random-access analyzers. Clin Chem 1988;34:1894-6.
Frank B, Gottlieb K. Amylase normal, lipase elevated: Is it pancreatitis? A case series and review of the literature. Am J Gastroenterol 1999;94:463-9.
Armbruster DA, Alexander DB. Sample to sample carryover: A source of analytical laboratory error and its relevance to integrated clinical chemistry/immunoassay systems. Clin Chim Acta 2006;373:37-43.
Broughton PM. Carry-over in automatic analysers. J Automat Chem 1984;6:94-5.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3]