Introduction
Infection with hepatitis B virus is one of the major public health problems worldwide. About 30% of the world’s population has serological evidence of the current and past HBV infection,[ 1 ] and about 1 million people die every year from HBV-related chronic liver diseases, including severe complications such as cirrhosis of the liver and hepatocellular carcinoma.[ 2 ] Screening and clinical diagnosis of infection with hepatitis B virus is mainly based on the presence of hepatitis B surface antigen (HBsAg), which is considered to be the most important serum marker.[ 3 ]
For the detection of HBsAg, various screening tests are currently available, including automated tests such as electro-chemiluminescence assay (ECLIA) and enzyme-linked immunosorbent assay (ELISA). In addition to these tests, there are several rapid tests available for HBsAg identification. The decision to use a specific serological test in screening or diagnosis of symptomatic cases depends on the tools, epidemiological information, and genetics of the infectious agents which could affect the assay limit.[ 4 ]
In recent years, ECLIA is being considered, by the diagnostic laboratories in developed nations, a preferred assay in diagnosing HBsAg due to its better testing output and objective interpretation of results. However, developing nations and third world countries are still dependent on ELISA or any other rapid diagnostic test due to cost-related constraints.
Although some of the studies[ 4–6 ] have performed comparative analyses for the efficiency of these tests, however, much is still to be explored especially the systematic analyses of these three tests in a single cohort. Said that there are still many grey areas where clear guidelines are required to give the reports as reactive or nonreactive. In this cross-sectional systematic diagnostic study, we performed ECLIA for detection of HBsAg and compared the same with other methods such as ELISA and rapid immunochromatographic test (RICT) which are alternatively used in resource-limited laboratories. Here, we attempted to identify the potential cut-off index (COI) and identify the grey zone for HBsAg-ECLIA testing to determine when an additional supporting or confirmatory test is required.
Material and Methods
Specimens
The study was conducted during June-December 2018 in the Department of Microbiology, JN Medical College, AMU, Aligarh. Three thousand eight hundred and forty-six consecutive patients with hepatitis attending the Hepatitis clinic or admitted in the Medicine wards of Jawaharlal Nehru Medical College were screened for viral hepatitis. All patients underwent complete physical examination, and the detailed clinical history was elicited from them. The study was approved by the Institutional Ethical Committee of JN Medical College, AMU, Aligarh.
Exclusion criteria
Patients with autoimmune hepatitis, alcoholic hepatitis, drug-induced hepatitis, patients giving a history of recent infection, surgery, trauma within the preceding two months, renal insufficiency, or with other acute or chronic inflammatory diseases were excluded from this study. None of the participants had received any antiviral or immunosuppressive therapy before or during this study.
Description of three tests used for detecting hepatitis B surface antigen
ECLIA
The Roche Diagnostic Cobas e411 Immunoassay System, a fully automated, random access, software-controlled immunoassay analysis system based on ECLIA method was used to detect the level of HBsAg. Elecsys HBsAg II was the kits used. Samples with concentration values ≥1 COI were considered clearly reactive as per the manufacturer’s instructions, however samples between 0.9 and 1 were considered borderline reactive. For the sake of epidemiological calculations, we combined these two reactive groups and collectively considered them as positive by ECLIA.
Enzyme-linked immunosorbent assay
ELISA method (Qualisa/HBsAg, Microwell enzyme Immunoassay, Qualpro Diagnostics, India) was also used to detect HBsAg samples and the positive results were interpreted at optical density ≥NC (average absorbance value of negative control) +0.15. The tests were performed according to the manufacturer’s instructions.
Rapid immunochromatographic test
The third test used for comparison was Alere Truline rapid test kit for HBsAg (Alere Medical Pvt Ltd, Gurgaon, India) and the results were interpreted as positive when the color band appeared at test band and control band of the result window.
Selection of samples for hepatitis B surface antigen confirmatory tests through nucleic acid amplification test
Of 259 samples tested positive by ECLIA, five representative samples from different concentration bands were confirmed by nucleic acid amplification test (NAAT) [Table 1 ]. The NAAT testing was a courtesy by the Blood Bank of the Department of Pathology of JN Medical College.
Table 1: Results of the representative samples tested by nucleic acid amplification test
Statistical analysis
Cohen’s kappa coefficient (k) with 95% confidence interval was computed to see the extent of agreement between the values of different methods beyond which we would expect by chance alone.
Interpretation of Cohen’s kappa coefficient (k)
Kappa <0: No agreement
Kappa between 0.00 and 0.20: Slight agreement
Kappa between 0.21 and 0.40: Fair agreement
Kappa between 0.41 and 0.60: Moderate agreement
Kappa between 0.61 and 0.80: Substantial agreement
Kappa between 0.81 and 1.00: Almost perfect agreement.
Receiver operating characteristic (ROC) curve analysis (MedCalc Software, Mariakerke, Belgium) was performed on the ECLIA-COI results considering ELISA as a reference to compare its sensitivity and specificity. The area under the ROC curve (AUC), which can be used as a measure of the accuracy of the test, was also calculated.
Results
Demographic information of patients and seroprevalence of hepatitis B surface antigen based on ECLIA
Out of 3846 samples tested for HBsAg, 259 (6.73%) were positive by ECLIA and 3587 (93.26%) were negative by ECLIA, thus the seroprevalence for HBsAg was 6.73% in our cohort. Out of these 259 HBsAg-positive samples, 98 (37.8%) were male and 161 (62.2%) were female. The age distribution of patients positive for HBsAg is given in Figure 1 .
Figure 1: Distribution of hepatitis B surface antigen -positive patients according to age group
Concordance rates of the three tests used
Out of these 259 samples, 68 were positive by both ECLIA and ELISA and had cut off index (COI) of >5 in ECLIA. Whereas 191 were positive by ECLIA only (COI between 0.9 and 5) hence the concordance rates of the ECLIA and ELISA in detecting serum HBsAg was 26.25% while the same for ELISA and RIT was 31.57%. Figure 2 shows the detailed distribution of COI of ECLIA-positive samples. Representative samples tested by NAAT were positive by NAAT except for the sample which had a COI <5 and hence was also negative by ELISA [Table 1 ].
Figure 2: Distribution of hepatitis B surface antigen-cut off index in positive samples in ECLIA
Receiver operating characteristic of ECLIA-cut off index
The ROC curve obtained from the comparative analysis of ECLIA-COI considering ELISA as reference is shown in Figure 3 and Table 2 .
Figure 3: Receiver operating characteristic of ECLIA-cut off index, the true-positive rate (sensitivity [y axis]) is plotted against the false-positive rate (1 - specificity [x axis])
Table 2: Confidence interval SE and Standard error based on ROC Curve
Table 3 shows the accuracy metrics of ECLIA and Rapid tests for the detection of HBsAg considering ELISA as reference. Cohen’s k was 0.39 between ELISA and ECLIA which denotes fair agreement while Cohens’ K was 0.54 between ELISA and RDT which means moderate agreement.
Table 3: Accuracy metrics point estimates and 95% confidence intervals of electrochemiluminescence immunoassay and rapid test for the detection of hepatitis B surface antigen
Discussion
In this cross-sectional diagnostic study, we found the prevalence of HBsAg to be 6.73% which is in concordance with the prevalence rate in our region. In the majority of studies from India, the overall HBsAg positivity rate has been documented to vary between 2% and 8%.[ 7 , 8 ]
Various screening or diagnostic tests are being currently used for the detection of HBsAg and the choice of the test is based on many factors, financial constraints being the primary factor considered. Moreover, accurate screening of HBsAg is of utmost clinical importance.[ 9 , 10 ]
In this study, we compared three different methods, ECLIA, ELISA, and RICT for the detection of HBsAg. ECLIA (Elecsys) uses the amplification mechanism of streptavidin-biotin and triple pyridinium to continuously receive electrons from the electric field supplied by tripropylamine.[ 11 , 12 ] In our study, the concordance rates of the ECLIA and ELISA was 26.25% while the same for ELISA and RICT was 31.57% that suggested the occurrence of false-negative or false-positive results while detecting HBsAg from serum. It is to be emphasized that false-negative results can lead to potential health risks to receivers of blood or blood components.[ 13 ] Hence, it is crucial to choose a method with good sensitivity and specificity (or a combination) for HBsAg screening and detection.
ECLIA, being a high-sensitivity assay, is likely to detect trace amounts of HBsAg, allowing HBV infection to be diagnosed earlier in the evolution of the disease that may lead to more effective treatment. Currently, HBsAg detection is done in most of the laboratories (especially in developing or third-world nations) by either ELISA or rapid tests, which may fail to detect it when the antigenemia is low in serum, thus resulting in under diagnosis.[ 14 ] On the other hand, advanced laboratories especially in developed nations use ECLIA, however, such improvements in sensitivity often are associated with more false-positive results.
Furthermore, studies have laid emphasis on discrepancies in the results from different diagnostic modalities and pointed out the absence of detectable DNA in the presence of HBsAg or otherwise, i.e., presence of detectable DNA in the absence of HBsAg.[ 15 ] Hence, HBV DNA and HBsAg could be screened simultaneously as discrepancies between different serological assays are common.
According to the manufacturer, ECLIA-COI>1 is considered positive, but contrary to this, various studies have suggested raising the COI for considering ECLIA as positive and considering a gray zone which needs to be confirmed by some other tests.[ 16 , 17 ] Another study[ 18 ] suggested that COI >12.1 should be taken as true positive and gray zone area be considered between <2 and 12.1. In our study, we found that samples with COI >5 in ECLIA were also positive by ELISA, but those having COI between 1 and 5 were negative by ELISA and rapid tests. It is emphasized here that the majority (28%) of the patients in our cohort were having ECLIA-COI in the range between 1 and 5 [Figure 2 ], this means that these patients were reported negative by ELISA and RIT. In addition, four representative samples which were positive by ECLIA and ELISA (i.e., COI >5) were positive with NAAT, and one sample which was negative by ELISA and positive by ECLIA (i.e., COI <5) was negative by NAAT. Though ECLIA is a sensitive and good test method for the detection of HBsAg, our preliminary study suggests that ECLIA-results with COI <5 should be reported cautiously and need to be confirmed by any confirmatory test method to rule out false positives. Or else, ECLIA-COI for positivity to be raised above 5.
Our main limitation was a smaller number of representative samples tested through NAAT and this was because of resource limitation at our end. NAAT is an expensive test and the representative samples were tested by the Blood Bank of the Department of Pathology on a courtesy basis. To determine the exact gray zone (COI) for interpretation by ECLIA, larger population-based studies are required with large number of representative samples tested through ECLIA, ELISA and NAAT.
In nutshell, ECLIA is a highly sensitive test however positivity-COI should be raised above 5 and in doubtful cases (i.e., COI <5), a combination of ECLIA and NAAT should be used.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
Acknowledgments
Mohammad Shahid gratefully acknowledges Arabian Gulf University (AGU), Bahrain, and the College of Medicine and Medical Science of AGU for sanctioning the funds for presenting the part of this research as a poster presentation in Pathology Update 2020, which unfortunately got canceled due to COVID-19 pandemic. The abstract was published in the supplementary issue of the journal Pathology (Volume 52, Supplement 1, February 2020, Pages S126-S127; https://doi.org/10.1016/j.pathol. 2020.01.431
References
1. Szabó E, Páska C, Kaposi Novák P, Schaff Z, Kiss A. Similarities and differences in hepatitis B and C virus induced hepatocarcinogenesis. Pathol Oncol Res 2004;10:5–11.
2. Okuda K. Hepatocellular carcinoma. J Hepatol 2000;32:225–37.
3. Dienstag JL. Hepatitis B virus infection. N Engl J Med 2008;359:1486–500.
4. Bisseye C, Ntsame Ndong JM, Mouity Matoumba A, Bengone C, Mouelet Migolet G, Nagalo BM. Comparison of electrochemiluminescence and ELISA methods in the detection of blood borne pathogens in Gabon. Asian Pacific J Trop Biomed 2017;7:805–8.
5. Das D, Roy S, Mondal S. Evaluation of performance characteristics of enzyme chemiluminescence immunoassay (ECLIA) and rapid diagnostic test (RDT) for HBV, HIV and HCV infections. J Clin Diagn Res 2018;12:DC14–7.
6. Mathur A, Dontula S, Jagannathan L. A study of centralized individual donor nucleic acid testing for transfusion transmitted infections to improve blood safety in Karnataka, India. Glob J Transfus Med AATM 2017;2:24–8.
7. Abraham P. Viral hepatitis. India Clin Lab Med 2012;32:159–74.
8. World Health Organization. Regional Office for South-East Asia. Prevention of Hepatitis B in India: an overview. WHO Regional Office for South-East Asia 2002. Available from:
https://apps.who.int/iris/handle/10665/205655 .
9. Eko Mba JM, Bisseye C, Ntsame Ndong JM, Mombo LE, Bengone C, Mouelet Migolet G, et al. Prevalent hepatitis B surface antigen among first-time blood donors in Gabon. PLoS One 2018;13:e0194285.
10. Zhang K, Liu Y, Chen R, Li Q, Xu Z, Si L, et al. Antigenicity reduction contributes mostly to poor detectability of HBsAg by hepatitis B virus (HBV) S-gene mutants isolated from individuals with occult HBV infection. J Med Virol 2018;90:263–70.
11. Wang T, Li D, Yan K, Yuan Y, Yang T, Du X, et al. Performance evaluation of a new fourth-generation HIV Ag/Ab combination electrochemiluminescence immunoassay
– Evaluation of a new HIV assay. Int J STD AIDS 2014;25:267–72.
12. Prusa AR, Hayde M, Unterasinger L, Pollak A, Herkner KR, Kasper DC. Evaluation of the Roche Elecsys Toxo IgG and IgM electrochemiluminescence immunoassay for the detection of gestational Toxoplasma infection. Diagn Microbiol Infect Dis 2010;68:352–7.
13. Kiely P, Stewart Y, Castro L. Analysis of voluntary blood donors with biologic false reactivity on chemiluminescent immunoassays and implications for donor management. Transfusion 2003;43:584–90.
14. Safia B, Waquaruddin A, Syed Ejaz A. Comparison of rapid test with ELISA for the detection of hepatitis B surface antibodies. Pak J Med Res 2014;53:60–2.
15. Allain JP, Candotti D. Diagnostic algorithm for HBV safe transfusion. Blood Transfus 2009;7:174–82.
16. Ryhan Bashir MH, Vidhya S, Saxena V. Hepatitis B surface antigen detection by enhanced chemiluminescence –A pilot study to determine the gray zone results. Int J Curr Microbiol App Sci 2016;5:778–2.
17. Ghai S, Dutta P. Comparative evaluation of electrochemiluminescence (ECLIA) with immunochromatographic test (ICT) available for hepatitis b surface antigen. Galore Int J Health Sci Res 2018;3:43–6.
18. Ning H, Li D, Bi X, Wang T, Du X, Yan K, et al. Compare HBsAg cut off index values of borderline or initially weakly reactive samples with neutralization results or electrochemiluminescence immunoassays in China. African J Microbiol Res 2012;6:3024–8.
