S1194 Risk of Drug-Induced Liver Injury in Patients With Chronic Hepatitis B and Active Tuberculosis Co-Infection: A Systematic Review and Meta-Analysis

Introduction: Tuberculosis (TB) infection is associated with significant morbidity and mortality worldwide. While active TB is curable, anti-TB therapies can lead to hepatotoxicity and drug induced liver injury (DILI). TB patients co-infected with chronic hepatitis B virus (HBV) infection may have a higher risk of DILI when treated with TB therapies compared to TB patients without HBV co-infection. We aim to evaluate the prevalence of DILI among HBV-TB co-infected patients undergoing anti-TB therapy and to determine whether HBV-TB co-infected patients have a higher risk of DILI compared to TB patients without HBV co-infection. Methods: We conducted a systematic review and meta-analysis using MEDLINE/PubMed from inception to September 1, 2019, and identified studies that provided data regarding chronic HBV, TB, and DILI. We included studies evaluating patients with non-cirrhotic chronic HBV infection, defined by positive surface antigen without human immunodeficiency virus. Patients with active TB were defined as having positive sputum cultures. DILI was defined as elevations in liver enzymes at least 1.5 times upper limit of normal. Pooled risk ratios of DILI in HBV-TB vs TB without HBV were evaluated with meta-analysis using random-effects models. Results: Our search criteria identified 46 articles, among which 9 studies met criteria to be included in the final study cohort. 463 patients had HBV-TB co-infection, and 2659 patients had active TB without HBV. The prevalence of DILI was 22.8% (n = 106) in HBV-TB co-infected patients and 13.0% (n = 345) in TB patients without HBV. On meta-analysis, HBV-TB co-infected patients had significantly higher risk of DILI when treated with TB therapies compared to TB patients without HBV (pooled risk ratio 1.85, 95% CI 1.28-2.68), but significant heterogeneity was observed (I2 = 69%) (Figure 1). A sub-analysis of only prospective cohort studies showed similar findings (pooled risk ratio 2.28, 95% CI 1.79-2.90, heterogeneity, I2 =61%). However, a sub-analysis of prospective cohort studies conducted after year 2000 demonstrated a pooled risk ratio of 2.75 (95% CI 2.10-3.59) with no significant heterogeneity (I2 = 0%). Conclusion: Our comprehensive meta-analysis observed that HBV co-infection in TB patients undergoing TB therapy is associated with significantly greater risks of DILI. Routine HBV screening prior to initiation of TB therapy is critical for early identification of HBV-TB co-infection, so that clinicians can modify TB and HBV management to reduce risks of DILI.Figure 1.: Random effects meta-analysis model evaluating risk of DILI in HBV-TB co-infected patients vs. TB patients without HBV co-infection undergoing anti-TB therapies.