Exploring the utility of a spontaneous adverse drug reaction reporting system in identifying drug–drug interactions between antiretrovirals, antitubercular drugs, and cotrimoxazole: a case/non-case analysis

Abstract

Drug–drug interactions (DDIs) cause significant morbidity and mortality, especially in patients with HIV with opportunistic infections such as tuberculosis. However, the literature on quantitative signal detection analyses for DDIs within the national spontaneous reporting systems (SRSs) of countries with high HIV/tuberculosis burdens is lacking. Our objective was to explore the utility of using post-marketing SRSs in quantitative signal detection analyses of DDIs. A case/non-case analysis using the Zimbabwean adverse drug reaction (ADR) database obtained from VigiBase® was utilized for quantitative signal detection using 2 × 2 contingency table calculations. Cases were defined as individual case safety reports (ICSRs) with the ADR of interest, and non-cases included the rest of the ICSRs. The exposure of interest was the use of a drug of interest. Signals of disproportionate reporting (SDRs) were observed for hepatotoxicity for the combined use of highly active antiretroviral therapy (HAART) and antitubercular treatments (ATT) [ADR reporting odds ratio (ROR) 43.78; 95% CI 5.24–366.08], HAART and isoniazid (ROR 44.84; 95% CI 5.36–374.99), and isoniazid and nevirapine-based HAART (ROR 35.60; 95% CI 9.39–134.89). SDRs were also observed for the combined use of nevirapine-based HAART and co-trimoxazole for Stevens–Johnson syndrome (ROR 28.91; 95% CI 14.00–59.70), severe cutaneous ADRs (ROR 16.10; 95% CI 9.40–27.57), and rash (OR 2.18; 95% CI 1.69–2.81). It is feasible to conduct signal detection analyses for DDIs within relatively small SRS databases. However, the observed potential SDR for the respective DDI should be investigated further as the method is only a hypothesis-generation analysis.