Diagnostic accuracy, sensitivity, and specificity of serum AST, α-GST, ALT, ALP, and gender in predicting Hepatitis B infection

Authors

Abstract

Aim Hepatitis B virus (HBV) infection remains a major public health concern, particularly in resource-limited settings. Accurate early detection is essential for effective intervention and disease control. This study evaluates the diagnostic performance of our previously developed predictive model based on gender and serum biomarkers. It assesses the diagnostic accuracy, sensitivity, and specificity of the predictive model for HBV status using gender, alpha-glutathione S-transferase (α-GST), aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALP).
Materials and Methods A total of 304 participants (53.9% male, 46.1% female) were evenly stratified into HBV-positive and HBV-negative groups based on standard serological assays. Our previously published model was applied to this new dataset. Performance metrics, including sensitivity, specificity, and overall accuracy, were calculated.
Results The model achieved a sensitivity of 78.1%, specificity of 77%, and overall accuracy of 77.6%, indicating strong diagnostic capability. These findings reinforce the earlier published results identifying AST as a significant predictor of HBV status, while α-GST, ALT, and ALP demonstrated lower individual predictive value.
Discussion The validated model offers a practical and efficient tool for HBV screening, especially in settings with limited diagnostic resources. Broader application across diverse populations is recommended to confirm generalizability and enhance public health outcomes.

Keywords

Introduction

Hepatitis B virus (HBV) infection poses a persistent global health challenge, affecting over 296 million individuals and causing an estimated 820,000 deaths annually due to complications such as liver cirrhosis and hepatocellular carcinoma [1, 2]. Despite the availability of effective vaccines and antiviral treatments, underdiagnosis and delayed detection, especially in low- resource settings, remain major barriers to early intervention and disease management [3, 4].
Standard diagnostic approaches, serological assays, and molecular techniques are reliable but often inaccessible in primary care due to cost and infrastructure requirements [5– 7]. This has prompted growing interest in predictive models leveraging clinical and biochemical variables to offer rapid, cost-effective alternatives for early screening [8]. Gender and serum biomarkers such as alpha glutathione S-transferase (α-GST), aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALP) have emerged as potential indicators of liver injury, inflammation, and viral hepatitis progression [9–11].
Beyond viral hepatitis, these enzymes are sensitive markers of hepatic dysfunction under various physiological conditions. For instance, a previous investigation into the hepatomodulatory properties of a di-herbal aqueous extract containing Ocimum gratissimum and Gongronema latifolium revealed significant modulation of liver enzymes in hyperglycemic rats, affirming the responsiveness of these biomarkers to hepatic stress [12]. In an earlier study, we developed and published a logistic regression-based predictive model incorporating gender and the aforementioned biomarkers to classify HBV status [13]. The findings highlighted AST as a statistically significant predictor, while α-GST, ALT, ALP, and gender demonstrated lesser individual contributions. However, the need to validate this model in a broader, balanced population remained.
Also, updated clinical guidelines underscore the importance of early diagnosis and risk stratification in HBV management to prevent disease progression [14], while insights into immune pathogenesis have further emphasized the role of host factors in HBV persistence and reactivation [15]. Thus, there is a growing need for accessible predictive tools to aid in early detection.
This study, therefore builds upon our earlier work by evaluating the diagnostic performance, specifically the accuracy, sensitivity, and specificity of the predictive model using data from a stratified cohort of HBV-positive and HBV-negative individuals. The aim is to determine the clinical utility of this model for HBV risk assessment and early detection, particularly in resource-constrained environments where conventional diagnostics may be impractical.

Materials and Methods

Study Design and Population
A cross-sectional analytical study was conducted to evaluate the diagnostic performance of a predictive model for Hepatitis B virus (HBV) status based on gender and serum biomarkers. A total of 304 participants (53.9% male, 46.1% female) were recruited from Enugu State University Teaching Hospital (ESUTH), Parklane, Enugu, Nigeria. Participants were evenly divided into HBV-positive (n = 152) and HBV-negative (n = 152) groups based on confirmed HBsAg serological status.
Inclusion and Exclusion Criteria
Adults aged 18–65 years with confirmed HBV status were included. Individuals co-infected with HIV or HCV, pregnant women, and those with a history of chronic liver disease or hepatotoxic drug use were excluded to minimize potential confounding variables.
Biochemical Assays
Venous blood (5 mL) was collected aseptically from each participant. Serum was separated by centrifugation and stored at –20°C until analysis. The following serum biomarkers were evaluated:
• Alpha Glutathione S-transferase (α-GST): Measured using
commercial ELISA kits (Elabscience®, USA).
• AST and ALT: Analyzed via the Reitman–Frankel colorimetric method.
• ALP: Assessed using the kinetic method based on p-nitrophenyl phosphate hydrolysis.
All procedures were carried out according to manufacturer instructions with appropriate quality controls.
Model application and performance evaluation
The predictive model developed [13] was applied to the dataset. Independent variables included gender, α-GST, AST, ALT, and ALP. Binary logistic regression analysis was used to predict HBV status.
Statistical Analysis
Data were analyzed using SPSS version 25. Descriptive statistics were presented as frequencies (%), where appropriate. The model’s diagnostic performance was evaluated using a classification table comparing predicted and actual HBV statuses, with outcomes categorized as either positive (1) or negative (0). Model performance was further assessed using sensitivity, specificity, and overall accuracy metrics. Receiver operating characteristic (ROC) curves were generated to determine the model’s discriminative capacity. A p-value of < 0.05 was considered statistically significant.
Ethical Approval
This study was approved by the Ethics Committee of Enugu State University Teaching Hospital (Date: 2023-07-10, No: ESUTHPK-MHC/RA/034/VOL2165).

Results

The predictive performance of the gender- and serum biomarker-based model for hepatitis B virus (HBV) status was evaluated using a classification table comparing predicted outcomes with actual serological results. Out of 304 participants, the model correctly identified 117 HBV-negative individuals (true negatives) and 118 HBV-positive individuals (true positives). However, it misclassified 35 HBV-negative individuals as positive (false positives) and 33 HBV-positive individuals as negative (false negatives). The model achieved a sensitivity of 78.1%, indicating its ability to correctly identify HBV-positive individuals. The specificity was 77.0%, reflecting the model’s capability to correctly identify HBV-negative cases. The overall predictive accuracy was 77.6%, demonstrating strong diagnostic potential (Figure 1; Table 1). This figure illustrates the key performance metrics—sensitivity (78.1%), specificity (77.0%), and overall accuracy (77.6%)—of the logistic regression-based predictive model used to classify HBV status based on gender and serum biomarkers (α-GST, AST, ALT, ALP). The cut-off probability threshold was set at 0.500. The classification threshold was set at 0.500 for binary logistic regression. These findings support the model’s clinical utility in HBV risk stratification, particularly for preliminary screening in low-resource settings.

Discussion

This study evaluated the diagnostic performance of a predictive model for hepatitis B virus (HBV) status using gender and serum biomarkers (α-GST, AST, ALT, and ALP), building upon our previously published work [13]. The model demonstrated a sensitivity of 78.1%, specificity of 77.0%, and an overall accuracy of 77.6%, confirming its robustness in accurately predicting HBV status among the study population.
These findings reinforce the clinical relevance of integrating liver enzyme markers, particularly AST, into predictive algorithms for HBV. AST was previously identified as a statistically significant predictor of HBV infection [13], a finding consistent with multiple studies that have reported elevated AST activity as a hallmark of hepatocellular injury associated with viral hepatitis [16–18]. Although ALT and ALP are conventionally associated with hepatic dysfunction, their individual predictive power in this context was limited, which supports earlier observations on their non-specificity in differentiating chronic HBV from other liver pathologies [16, 19, 20].
The balanced sensitivity and specificity observed in this model indicate its potential to minimize both false positives and false negatives, key attributes for screening tools in public health contexts. This is particularly relevant in low-resource settings where confirmatory molecular diagnostics may be unavailable [21, 22]. Several predictive models utilizing non-invasive biochemical markers have shown promise in similar settings, lending support to the feasibility of such tools for HBV screening [17, 22, 23].
Furthermore, the inclusion of gender in the model reflects an appreciation for biological variability, even though gender did not independently contribute to prediction in our earlier findings [13]. Sex-specific responses in hepatic enzyme expression and viral replication have been previously reported, suggesting a nuanced role of gender in HBV pathogenesis [24, 25].
Interestingly, the application of this model aligns with research that explores the modulation of liver enzymes under various physiological states. A study involving herbal extracts demonstrated significant alterations in hepatic biomarkers in hyperglycemic models, reinforcing the diagnostic value of AST, ALT, and ALP in detecting subclinical liver dysfunction [12, 15].

Limitations

Despite its strengths, this study is not without limitations. The sample was restricted to a single institution, which may affect external validity. Furthermore, the exclusion of virological parameters such as HBV DNA levels and HBeAg status may limit the model’s precision in differentiating active versus inactive infection phases. Future studies should incorporate longitudinal data and external validation cohorts to assess predictive consistency across diverse populations and settings. Future studies involving larger, multicenter cohorts and longitudinal validation are recommended to enhance generalizability and clinical applicability. Incorporating additional virological and immunological markers may further refine the model’s predictive accuracy for different phases of HBV infection.

Conclusion

This study demonstrates the utility of a predictive model incorporating gender and hepatic enzyme biomarkers (α-GST, AST, ALT, and ALP) for accurately identifying hepatitis B virus (HBV) status. The model showed strong diagnostic performance, with balanced sensitivity (78.1%), specificity (77.0%), and overall accuracy (77.6%), underscoring its potential as a reliable screening tool, particularly in low-resource settings where access to molecular diagnostics is limited. The findings support the integration of routine liver enzyme panels into predictive algorithms for early HBV detection and suggest that such models could serve as effective adjuncts to standard diagnostic protocols.

References

1. World Health Organization (WHO). Hepatitis B [Internet]. World Health Organization Fact Sheet. Published July 2023 [cited 2025 Jul 20]. Available from: https://www.who.int/news-room/fact-sheets/detail/hepatitis-b
   PubMed Google Scholar
2. Tang LSY, Covert E, Wilson E, Kottilil S. Chronic hepatitis B infection: a review. JAMA. 2018;319(17):1802-13.
   PubMed Google Scholar
3. Schweitzer A, Horn J, Mikolajczyk RT, Krause G, Ott JJ. Estimations of worldwide prevalence of chronic hepatitis B virus infection: a systematic review. Lancet. 2015;386(10003):1546-55.
   PubMed Google Scholar
4. MacLachlan JH, Cowie BC. Hepatitis B virus epidemiology. Cold Spring Harb Perspect Med. 2015;5(5):a021410.
   PubMed Google Scholar
5. Lok ASF, McMahon BJ. Chronic hepatitis B: update 2019. Hepatology. 2019;67(4):1560-99.
   PubMed Google Scholar
6. European Association for the Study of the Liver (EASL). Clinical practice guidelines on the management of hepatitis B virus infection. J Hepatol. 2017;67(2):370-98.
   PubMed Google Scholar
7. Viganò M, Lampertico P, Colombo M. New insights into the pathogenesis and clinical management of chronic hepatitis B virus infection. Hepatol Int. 2014;8(4):389-401.
   PubMed Google Scholar
8. Wang J, Xia J, Zhang R, et al. A novel index using routine clinical parameters for predicting significant liver inflammation in chronic hepatitis B. J Viral Hepat. 2018;25(5):530–539.
   PubMed Google Scholar
9. Kim WR, Flamm SL, Di Bisceglie AM, Bodenheimer HC. Serum activity of ALT as an indicator of liver disease. Hepatology. 2008;47(4):1363-70.
   PubMed Google Scholar
10. Njoku CH, Ejim PC, Ogoina D, Akinsulie AO, Okany CC, Onwubere BJ. Liver enzyme profile among hepatitis B surface antigen-positive and negative blood donors in Nigeria. Niger Med J. 2011;52(2):52-4.
    PubMed Google Scholar
11. Olayanju AO, Onigbinde OA, Akinrinmade AF. Gender and liver enzyme pattern in chronic hepatitis B infection. Niger J Clin Pract. 2019;22(1):89-93.
    PubMed Google Scholar
12. Okerulu AL, Onwuka OM, Ajuzie GC. Hepatic enzymatic activities of di- herbal aqueous extract of Ocimum gratissimum and Gongronema latifolium on liver dysfunction in experimental hyperglycemic rats. J Int Res Med Pharm Sci. 2019;18(2):13-22.
    PubMed Google Scholar
13. Ezeadichie SO, Ikekpeazu JE, Wariso M, Katchy AU, Ugwu CC, Onwuka OM. Development of a predictive model for hepatitis B virus (HBV) status using gender and serum biomarkers. Precis Nanomed. 2025;8(2):1511-7.
    PubMed Google Scholar
14. Terrault NA, Bzowej NH, Chang KM, Hwang JP, Jonas MM, Murad MH. AASLD guidelines for treatment of chronic hepatitis B. Hepatology. 2016;63(1):261-83.
    PubMed Google Scholar
15. Bertoletti A, Ferrari C. Adaptive immunity in HBV infection. J Hepatol. 2016;64(1 Suppl):S71-83.
    PubMed Google Scholar
16. Wu Y, Zhang X, Liu M, Chen H, Li Y. Noninvasive diagnosis model for predicting significant liver inflammation in chronic hepatitis B patients. Sci Rep. 2025;15(1):87756.
    PubMed Google Scholar
17. Zhou K, Xu L, Liu J, Chen Y, Zhang W. A roadmap for serum biomarkers for hepatitis B virus. Front Med. 2022;16(2):146-60.
    PubMed Google Scholar
18. Wang Y, Fan Y, Zhang T, Li X, Gao Z. Review of serum biomarkers and models derived from them in HBV-related liver diseases. J Clin Transl Hepatol. 2020;8(3):280-9.
    PubMed Google Scholar
19. Pandit A, Shah K. Liver function tests. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2023 [cited 2025 Jul 20]. Available from: https://www. ncbi.nlm.nih.gov/books/NBK482489/
    PubMed Google Scholar
20. Pratt DS, Kaplan MM. Evaluation of abnormal liver-enzyme results in asymptomatic patients. CMAJ. 2000;162(3):367-79.
    PubMed Google Scholar
21. Chen J, Liu Q, Wang X, Huang R, Zhang H. Liver volume-based prediction model for patients with hepatitis B virus-related acute-on-chronic liver failure. BMC Gastroenterol. 2023;23(1):45.
    PubMed Google Scholar
22. Xiao G, Yang J, Yan L. Noninvasive models for assessment of liver fibrosis in patients with chronic hepatitis B. World J Gastroenterol. 2016;22(31):7227-36.
    PubMed Google Scholar
23. Chen CH, Hung CH, Hu TH, Wang JH, Lu SN, Lee CM. Comparison of noninvasive models for liver fibrosis in chronic hepatitis B patients with normal ALT levels. World J Gastroenterol. 2015;21(36):10453-61.
    PubMed Google Scholar
24. Ruggieri A, Gagliardi MC, Anticoli S. Sex-dependent outcome of hepatitis B and C virus infections: synergy of sex hormones and immune responses? Front Immunol. 2018;9:2302.
    PubMed Google Scholar
25. Yin J, Zhang H, He Y, et al. Gender-specific effects of estrogen receptor alpha on HBV replication and liver fibrosis in transgenic mice. J Microbiol Biotechnol. 2024;34(1):1-10.
    PubMed Google Scholar

Declarations

Additional Information

Publisher’s Note
Bayrakol MP remains neutral with regard to jurisdictional and institutional claims.

Rights and Permissions

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0). To view a copy of the license, visit https://creativecommons.org/licenses/by-nc/4.0/

View full license details →

About This Article