European Journal of Cardiovascular Medicine

In clinical chemistry laboratories, interference can be defined as a cause of clinically significant bias in measured analyte concentration due to the effect of another component or property of the sample ^[1]. Interference is generally concentration dependent and can be quantified as a function of the concentration of the interfering compound ^[2]. Nonspecificity of the detection system, alteration of principle chemical reaction, alteration of analyte enzyme by potential interfering substance or some other specimen dependent bias are some of the reasons for interference in clinical laboratories ^[1]. Inaccuracy (total analytical error) is major contributed by three factors: imprecision, method-specific bias and sample-specific bias ^[3,4]. Generally, measurement evaluation procedures measures imprecision and method-specific bias but sample-specific bias (interference) is not measured by misinterpreting it as a sample specific problem ^[5,6]. However, such so called sample-specific bias is actually dependent on concentration of interfering substance which can be readily determined. Such sample-specific bias must be addressed properly to improve clinical outcome.

Serum total cholesterol can provide valuable information to assess lipid metabolism, risk of cardiovascular diseases, and effectiveness of dietary or pharmacological intervention. Estimation of total cholesterol in serum is considered as a ‘measurement system’, in which reagent, calibrator and instrument are desired from a single manufacturer. However, if reagent-calibrator-instrument system is not from single manufacturer, the onus of analytical performance surety shifts from manufacturer to user laboratory. So, in such case validation and verification of all the analytical performance criteria including interference needs to be done by user laboratory ^[7]. Bilirubin interference is one of the major problems as >5 mg/dL of bilirubin can significantly affect total cholesterol analysis ^[7].

There are several ways by which interference can occur like by altering the chemical reaction, overlapping properties of analyte and interferent, by altering physical properties of matrix, by altering the activity of an enzyme (analyte or reagent), Nonspecificity and cross-reactivity ^[1]. Understanding interference pattern enables laboratories to implement appropriate quality control measures and provide reliable diagnostic information for patient care management. Based on present knowledge, our study is primarily aimed to find out whether high concentration of bilirubin interferes with the detection of Serum Total Cholesterol in our laboratory settings. Secondly if bilirubin interference found significant then we also intend to evaluate effect of different concentrations of bilirubin on estimation of Serum Total Cholesterol concentration.

This study was carried out in the Department of Biochemistry at Surat Municipal Institute of Medical Education and Research (SMIMER), Surat, Gujarat, India, during October-2025 to December-2025. This study was done on left over serum samples of patients after routine testing and sample anonymity was maintained so patient consent was waived off. This project was approved by institutional ethical committee of SMIMER. This study was experimental work carried out as per Clinical and Laboratory Standard Institute (CLSI) approved guideline EP7-A2 [1]. Instruments and Reagents: Instruments: Fully automated clinical chemistry analyzer Erba-XL-1000 (Transasia Bio-Medicals Ltd. Mumbai, India) was used for bilirubin and serum total cholesterol estimation. Electronic balance ABJ 320-4 (Kern and Sohn, Germany) was used for weighing reagents. Both instruments were calibrated as per standard protocol. Reagents: Serum total cholesterol activity was measured by cholesterol oxidase-peroxidase based method (Pathozyme Diagnostics, Kolhapur, India) [9]. Concentration of bilirubin was measured by kit based on diazotized sulphanilic acid method (Lab-Care Diagnostics, Valsad, India) [10]. Both kits were calibrated using Randox calibrator and quality check was made by Randox quality control sera. Commercially available analytical grade anhydrous bilirubin powder was used for making bilirubin stock solution. Methods: Preparation of bilirubin stock solution: Anhydrous analytical grade bilirubin powder (60 mg) was accurately weighed and dissolved in 4 mL of 0.4% (w/v) NaOH solution. The solution was swirled continuously, and 0.4% NaOH was added stepwise (100 μL per addition) until all the bilirubin powder dissolved completely. To minimise the problems of photolysis and degradation of bilirubin, bilirubin stock solution was prepared freshly as and when required. High concentration of bilirubin in stock solution is required because when we spike the sample with this bilirubin stock solution, minimum possible volume of stock solution should be used so that sample matrix modification is kept minimum [1]. The concentration of total bilirubin in each newly prepared stock solution was estimated in 20 replicates. The mean and SD of these measurements was calculated and compared with calculated concentration. Stock solutions were considered acceptable for experimental use only upon agreement between the measured and calculated concentrations. Calculated Bilirubin Conc.(mg/dL) = (60 mg)/(Total volume of solution (in mL))×100 Preparation of Base Pool sample: In accordance with CLSI EP07-A2 approved guideline, analyte concentrations for interference testing should be selected in such a way so that it includes medical decision levels as far as possible. To fulfil this requirement in present study, we prepared three distinct pooled serum of low, medium and high serum total cholesterol samples aligned with the medical decision levels. Samples having a visible sign of icterus, hemolysis and lipemia were not included in the preparation of the base pool samples. Further as bilirubin is endogenously produced metabolite from heme catabolism, it was essential to keep bilirubin concentration at a minimal level in base pool serum. To make sure that bilirubin remains at the lowest possible level in base pool serum, samples with total bilirubin level less than 1 mg/dL were included for base pool serum. Finally, samples having total cholesterol levels below 150 mg/dL were utilised for preparing low base pool, between 150-250 mg/dL were utilised for medium base pool and more than 250 mg/dL were utilised for high base pool sera. Target volume for each base pool sera was kept minimum 10 mL so that it remains sufficient for the current experiment. Each final base pool sera were measure 20 times for total cholesterol and total bilirubin and their mean and SD were estimated to establish the baseline characteristics of pooled sera. Preparation of control and test pool sera: According to CLSI approved guidelines, the volume of bilirubin stock solution added to the base pool serum should not exceed 1/20th (5%) of the total base pool volume to minimise alteration in the sample matrix [1]. Control pools were prepared to match the test pools in all aspects, with the exception that the test interferent was replaced with an equal volume of the solvent (i.e. 0.4% NaOH) used for bilirubin stock solution preparation. Given that bilirubin is endogenously present in the base pool, the final bilirubin concentration in each pool was measured. Accordingly control pools were prepared by adding 125 µL of 0.4% sodium hydroxide (NaOH) to 4875 µl of each base pool serum, So three separate control pools were prepared corresponding to the low, medium, and high base pools. Test pools were similarly prepared by adding 125 µL of bilirubin stock solution to 4875 µL of each base pool serum. Three separate test pools were prepared corresponding to the low, medium, and high base pools. In total, six solutions were prepared – three control pools and three test pools – reflecting the three distinct total cholesterol concentration levels (low, medium and high) of the base pools. Number of replicates: The number of replicates required to detect interference effect with 95% confidence level (α=0.05) and 95% power (β=0.05) were calculated based on maximum allowable error derivative (dmax) and within-run repeatability (SD) of the analytic method. In present study dmax was using the maximum allowable error derived from the biological variation of serum total cholesterol. Ricos et al. (1999) estimated the total error (TE) of serum total cholesterol based on biological variation to be 9% [11] and this value was adopted for current study. The dmax of each pool was calculated using the following formula. d_max of each pool=Mean of base pool × (Total allowable error (%))/100 Within-run repeatability of total cholesterol was determined by analyzing each base pool (low, medium and high) 20 times and their standard deviation was calculated. These calculated dmax and SD values were used in the following formula to find number of required replicates (n) [12]. In case of non-integer value, rounding up to the next integer value was taken. n=2[(z_(1-α/2)+ z_(1-β) )x SD/d_max ]^2 Where, z values were used from standard statistical table for percentile for confidence level and power and accordingly Z0.975 and Z0.950 corresponds to 1.960 and 1.645 respectively. Sample analysis: The test (T) and control (C) samples were analyzed in alternate sequence (e.g. C1T1C2T2C3T3….CnTn) in fully autoanalyzer. To mitigate potential carryover effect from test samples to control samples, additional control serum samples (Cx) were intermittently introduced into the analytical sequence (e.g. C1T1CxCxC2T2CxCxC3T3…CxCxCnTn). Subsequently, the results obtained from these additional control samples were excluded from the final analysis. Measurement of both serum bilirubin and total cholesterol were carried out simultaneously for all samples. Data analysis: The analytical data obtained from control and test pools were used to calculate the point estimate of the observed interference effect (dobs) which is the difference between the mean values of the test and control samples. The cut-off value for two-sided test, dc, was employed to determine whether the null hypothesis should be accepted or rejected. In the present study, the null hypothesis states that bilirubin do not interfere with the estimation of total cholesterol (dnull = 0). The cut-off value (dc) was calculated using the following formula: d_c=(d_null+SD∙z_(1-α⁄2))/√n Where, dnull represents the value specified in the null hypothesis (0 in this study), n denotes the number of replicates, SD is the standard deviation of repeatability of the measurement procedure which was determined earlier in the base pool of low, medium and high total cholesterol base pool sera and z values were obtained from standard statistical considering α=0.05. If the point estimate (dobs) is less than or equal to the cut-off value (dc) than null hypothesis was accepted otherwise alternate hypothesis i.e. bilirubin interferes with the estimation of total cholesterol, was accepted. In case where interference effect was detected, a dose-response series of the interferent was conducted to characterise the degree of interference as a function of interferent concentration. Preparation of five dose-response samples A series of test samples were prepared by systematically varying the interferent (total bilirubin) concentration while keeping the analyte (total cholesterol) concentration constant. This was achieved by proportionate mixing of control and test pool. This procedure was done separately for low, medium and high total cholesterol pool. The control pool contained the lowest concentration of the interferent to be tested whereas the test pool contained the highest concentration. In this phase of the experiment, the previously prepared test pool sample was designated as high bilirubin concentration sample and control pool sample served as low bilirubin concentration sample. The proportionate mixture of high and low bilirubin concentration serum for all three total cholesterol level pools are shown in table-1. Table 1: Preparation method of dose-response samples along with measured concentration of bilirubin in each pool Series No. Proportionate mixture Measured concentration of bilirubin Low bilirubin concentration sample High bilirubin concentration sample Low total cholesterol pool Medium total cholesterol pool High total cholesterol pool 1 400 µl 0 µl 0.6 mg/dL 0.6 mg/dL 0.6 mg/dL 2 300 µl 100 µl 8 mg/dL 7.9 mg/dL 8 mg/dL 3 200 µl 200 µl 14.9 mg/dL 15.5 mg/dL 15.1 mg/dL 4 100 µl 300 µl 22.4 mg/dL 22.4 mg/dL 22.6 mg/dL 5 0 µl 400 µl 29.6 mg/dL 29.5 mg/dL 30.1 mg/dL In present study, a total five concentrations were analyzed in quadruplicate within a single analytical run. To minimize potential drift effects, all samples and replicates were analysed in a randomized sequence which was generated from table of random numbers. Data analysis of dose-response sample: The observed interference effect at each bilirubin concentration was calculated by determining the mean total cholesterol value of low bilirubin concentration pool and deducting it from each individual value of total cholesterol. These results were plotted with the observed effect on the y-axis and the interferent concentrations on x-axis to evaluate the dose-response relationship. If data appeared randomly distributed about a straight line, linear least square regression analysis was applied to determine slope, intercept and residual error from individual observations. When a linear relationship was established between interferent concentration and analyte measurement, the regression slope represents the bias per unit of interferent and the y-intercept represents the correction factor for the endogenous interferent concentration.

The baseline characteristics of the three base pool samples and the calculated parameters for interference testing are shown in Table-2.

Table 2: Baseline Characteristics of Base Pool Samples and Calculated Parameters for Interference Testing.

The mean total cholesterol concentration for the low, medium, and high base pools were 105.2 mg/dL, 198.65 mg/dL, and 293.05 mg/dL, respectively, with corresponding standard deviation of 3.53 mg/dL, 6.56 mg/dL, and 9.32 mg/dL. The mean endogenous total bilirubin concentration in the low, medium and high pools were 0.59 mg/dL, 0.65 mg/dL and 0.65 mg/dL, respectively, confirming minimum baseline bilirubin levels across all pools. Based on biological variation data, a total allowable error (TEa) of 9% was applied uniformly across all three pools, yielding maximum allowable error derivative (dmax) value of 9.47 mg/dL, 17.88 mg/dL, and 26.37 mg/dL for low, medium, and high pools, respectively. Using these parameters in conjunction with the within-run repeatability data, the calculated number of required replicates for interference testing was determined to be four (n=4) for each base pool to achieve 95% confidence and 95% power in detecting interference effect. The corresponding cut-off values (dc) for the 2-sided test were 3.46 mg/dL, 6.43 mg/dL, and 9.13 mg/dL for low, medium, and high pools, respectively.

Table 3: Results of the paired difference experiment

The results of the paired difference experiment are presented in table-3. The mean total cholesterol concentrations in the control pools were 100.25 mg/dL, 192 mg/dL, and 286 mg/dL for the low, medium, and high pools, respectively, with corresponding mean total bilirubin concentrations of 0.55 mg/dL, 0.58 mg/dL, and 0.58 mg/dL. In the test pools supplemented with bilirubin, the mean total cholesterol concentrations were 87.75 mg/dL, 173.25 mg/dL, and 240.75 mg/dL, respectively. With mean total bilirubin concentrations of 29.675 mg/dL, 29.65 mg/dL, and 29.8 mg/dL, respectively. The observed difference (dobs) between the control and test pools were -12.5 mg/dL, -18.75 mg/dL, and -45.25 mg/dL for low, medium, and high pools, respectively, indicating a negative bias in cholesterol measurement. In all 3 pools, the observed difference was more than the cut-off value, leading to rejection of null hypothesis and confirming statistically significant interference of elevated bilirubin concentration on total cholesterol measurement across all total cholesterol levels.

Dose response experiment of bilirubin interference on total cholesterol estimation:

The effect of different bilirubin concentrations on total cholesterol estimation was evaluated by linear least square regression analysis across three cholesterol levels (low, medium, and high).

Table 4: Summary of results from a five-level dose-response series for low total cholesterol concentration.

Figure 1: Regression analysis plot of results from dose response experiment described in Table-4

Table 4 and Figure 1 demonstrate the dose-response relationship between bilirubin concentration and its interference effect on low total cholesterol measurement. Linear regression analysis shown the equation Y = -0.42X + 0.23 (R² = 0.69, p < 0.05), indicating a statistically significant negative correlation. The regression slope of -0.42 mg/dL per mg/dL of bilirubin represents the systematic negative bias introduced per unit increase in bilirubin concentration.

Table 5: Summary of results from a five-level dose-response series for medium total cholesterol concentration.

Figure-2: Regression analysis plot of results from dose response experiment described in Table-5

Table 5 and Figure 2 illustrate the dose-response relationship for medium cholesterol concentrations. Linear regression analysis produced the equation Y = -0.64X + 0.27 (R² = 0.72, p < 0.05), demonstrating a stronger negative interference effect compared to the low cholesterol pool. The regression slope of -0.64 mg/dL per mg/dL of bilirubin indicates increased sensitivity to bilirubin interference at medium cholesterol concentrations.

Table 6: Summary of results from a five-level dose-response series for high total cholesterol concentration.

Figure-3: Regression analysis plot of results from dose response experiment described in Table-6

Table 6 and Figure 3 present the dose-response relationship for high cholesterol concentrations. Linear regression analysis revealed the equation Y = -1.52X + 1.34 (R² = 0.80, p < 0.05), demonstrating the most pronounced interference effect among all three cholesterol levels. The regression slope of -1.52 mg/dL per mg/dL of bilirubin, along with the highest R² value of 0.80, indicates both the strongest negative bias and the best model fit, confirming a robust dose-dependent relationship at elevated cholesterol concentrations.

It is the primary responsibility of the reagent and instrument manufacturer to characterise interference profiles during method development, but still clinical laboratories should verify the manufacturer's claim. In addition, as per CLSI guidelines, Laboratories should independently confirm method performance under their unique operational conditions.^[1,13]

Current study was done in two phase experimental design according to CLSI EP07-A2 guidelines. The first phase used an interference screening approach to determine whether high bilirubin concentration produce statistically and clinically significant effects on measurement of total cholesterol. This was done by paired difference testing. After determining the existence of significant interference in this initial screening phase, we moved on to the second phase, a thorough dose-response analysis intended to measure the connection between interferent concentration and analytical bias.

Our findings revealed that in all tested total cholesterol concentration ranges, elevated bilirubin concentrations have negative interference in total cholesterol measurement. As shown in Table 3, Bilirubin caused a negative bias of -12.5%, -9.8%, and -15.8% in low, medium, and high cholesterol pools, respectively. Our findings align with previous studies ^[14,15]. Such negative bias observed in present study as well as other study suggests that bilirubin may interfere by spectral overlap or by affecting the enzymatic reactions involved in cholesterol measurement, however the exact mechanism requires further research ^[8].

The data of dose response experiment is shown in Figure 1-3 and Table 4-6. The linear regression analysis of this data revealed slop of -0.42, -0.64 and -1.52 for low, medium and high total cholesterol pool respectively. This progressively increased slop suggest that the interference effect is not constant, rather it increases as concentration of total cholesterol increases. This finding has important clinical ramifications because patients who have both hypercholesterolemia and hyperbilirubinemia may have the highest analytical inaccuracy

Our analysis revealed an interesting pattern: both baseline cholesterol levels and absolute bilirubin concentration seemed to influence the magnitude of interference. This implies that compared to bilirubin concentration alone, the bilirubin-to-cholesterol ratio may be a better indicator of interference. This ratio may allow labs to create more advanced flagging algorithms that automatically detect specimens at high risk for interference, leading to either different measurement techniques or clinical interpretation guidelines. Pre-analytical quality management would significantly advance with the use of such predictive tools.

Clinical chemistry has a wide range of intricate interference mechanisms, including chemical artifacts like spectral overlap or redox reactions, detection artifacts like fluorescence or light scattering, physical artifacts like turbidity or viscosity changes, enzyme inhibition, cross-reactivity with antibodies or substrates, non-specific binding, and water displacement effects in ion-selective measurements^[16,17]. Since bilirubin have strong absorbance in the visible spectrum which can overlap with chromogenic detection systems frequently used in enzymatic cholesterol methods, spectral interference is probably the main mechanism in the case of bilirubin interference with cholesterol assays. Furthermore, hydrogen peroxide-based detection systems used in coupled enzymatic reactions may be hampered by bilirubin's antioxidant qualities.

Practically, a number of mitigating techniques should be considered when laboratories come across samples with increased bilirubin. To determine the possibility and degree of interference, total bilirubin should first be measured. Second, adjusted cholesterol values can be estimated using the regression equations given in this paper; however, this method may not take method-specific variables into account and assumes linearity holds across all bilirubin concentrations. Third, depending on laboratory capabilities and clinical urgency, alternate assessment methods such as direct cholesterol methods less vulnerable to bilirubin interference, sample dilution procedures, or physical removal of bilirubin through adsorbents may be used. Above all, since interference characteristics can range significantly between reagent formulations and equipment designs, laboratories should create their own interference profiles for their particular analytical platforms.

This study has few limitations that should be discussed. For sample spiking, we first used commercially available bilirubin, which is mostly found in the unconjugated form and may differ biochemically and photochemically from the combination of conjugated and unconjugated bilirubin seen in patient specimens. Second, although dilution used in this experiment was well within the range provided by CLSI guideline (<5%), but still it can unavoidably modify the sample matrix that could affect interference patterns. Third, because interference profiles are known to be method-dependent—that is, findings acquired on different analysers or with different reagent chemistries may generate distinct interference patterns—our study only looked at one analytical platform and reagent system.

Our study demonstrated clinically meaningful bilirubin interference in total cholesterol measurement. Based on our findings, we suggest each laboratory can create their own interference management procedures using the regression equations and interference thresholds determined in this work. The bilirubin-to-cholesterol ratio as a predictive tool, interference patterns across various analytical platforms, possible interactions between multiple interferents, and the clinical impact of uncorrected interference on patient management decisions and outcomes should be investigated in future research. 

This study carried out bilirubin interference on total cholesterol measurement as per CLSI guideline EP07-A2 at a tertiary care hospital laboratory. Our study findings showed that bilirubin produces significant dose-dependent negative interference on total cholesterol analysis, with bias ranging from -9.76% to -15.82% across different total cholesterol concentrations. In patients having high level of bilirubin, total cholesterol results should be interpreted with caution and should not be used as a standalone diagnostic marker for lipid assessment. The linear regression equations derived from this study can be applied by clinical laboratories to quantitatively estimate interference and predict corrected cholesterol values. As interference patterns may differ across instruments, methods and reagent systems, laboratories should validate the findings of these in their own setup. Our study emphasizes the importance of interference evaluation as part of routine quality assurance programs to ensure accurate and reliable patient results.

1. Clinical and Laboratory Standards Institute (CLSI). Interference Testing in Clinical Chemistry; Approved Guideline—Second Edition. CLSI document EP7-A2 (ISBN 1-56238-584-4). Clinical and Laboratory Standards Institute, 940 West Valley Road, Suite 1400, Wayne, Pennsylvania.
2. Glick MR, Ryder KW. Analytical system ranked by freedom from interferences. Clin Chem 1987;33:1453-8.
3. Kringle RO, Bogavich M. Statistical procedures. In: Burtis CA, Ashwood ER, eds. Tietz Textbook of Clinical Chemistry. 3rd ed. Philadelphia: W.B. Saunders Co.; 1999:265-309.
4. Krouwer JS. Estimating total analytical error and its sources: Techniques to improve method evaluation. Arch Pathol Lab Med. 1992;116:726-731.
5. Lawton WH, Sylvestre EA, Young-Ferraro BJ. Statistical comparison of multiple analytic procedures: Application to clinical chemistry. Technometrics. 1979;21:397-409.
6. Krouwer JS. How to improve total error modeling by accounting for error sources beyond imprecision and bias. Clin Chem. 2001;47(7):1329-1331.
7. Meeusen JW, Ueda M, Nordestgaard BG, Remaley AT. Lipids and lipoproteins. In: Rifai N, Chiu RWK, Young I, Burnham CAD, Wittwer CT, editors. Tietz textbook of laboratory medicine. 7th ed. Philadelphia (PA): Elsevier; 2023. p. 396-399.
8. AM Simundic, N Nikolac, WG Guder. Preanalytical Variation and Preexamination Processes. In: N Rifai, A Horwath, C Wittwer, editors. Tietz textbook of Clinical Chemistry and Molecular Diagnostics, 1st SEA Edn, India: RLX (an imprint of Elsevier); 2018. Pp. 92-93.
9. Liqvipath Cholesterol estimation kit. For photometric measurement of total cholesterol in serum based on CHOD-PAP method. Kit insert. Pathozyme diagnostics, Kolhapur, India (www.pathozyme.com)
10. Accucare Bilirubin kit. For photometric estimation of total and direct bilirubin concentration based on diazotized sulphanilic acid method. Kit Insert. Lab-care diagnostics (india) Pvt Ltd, Valsad, Gujarat, India (www.labcarediagnostics.com)
11. Ricos C, Alvarez V, Cava F, Garcia-Lario JV, Hernandez A, Jimenez CV, et al. Current databases on biological variation: pros, cons and progress. Scand J Clin Lab Invest. 1999;59(7):491–500.
12. Desu MM, Raghavarao D. Sample Size Methodology. San Diego: Academic Press, Inc.; 1990:7,9,30.
13. Saibaba KS, Bhaskar MV, Rao PV, Ramana GV, Dakshinamurty KV. Interferences in clinical chemistry analysis. Indian J Clin Biochem 1998;13:55–62.
14. Cano-Corres R, Sole-Enrech G, Aparicio-Calvente MI. Definition of icteric interference index for six biochemical analytes. Biochem Med (Zagreb). 2023;33(2):020702. doi:10.11613/BM.2023.020702.
15. Ho CK, Chen C, Setoh JW, Yap WW, Hawkins RC. Optimization of hemolysis, icterus and lipemia interference thresholds for 35 clinical chemistry assays. Practical Laboratory Medicine. 2021;25:e00232. doi:10.1016/j.plabm.2021.e00232.
16. Apple FS, Koch DD, Graves S, Ladenson JH. Relationship between direct-potentiometric and flame- photometric measurement of sodium in blood. Clin Chem. 1982;28:1931-1935.
17. Durst RA, Siggard-Andersen O. Electrochemistry. In: Burtis CA, Ashwood ER, eds. Tietz Textbook of Clinical Chemistry. 3rd ed. Philadelphia: W.B. Saunders Co.; 1999.