Efavirenz versus Protease Inhibitors in Patients with Human Immunodeficiency Virus – A Systematic Review and Meta-Analysis

Abstract

Background: Efavirenz- and protease inhibitor-based regimens remain viable options across the globe. We conducted a meta-analysis to compare the effectiveness of efavirenz-based regimens relative to protease inhibitor-based regimens (PIs). Methods: Embase, PubMed, Cochrane, and clinicaltrials.gov were searched for randomised controlled trials (RCTs) conducted between 1987 and 2018 comparing efavirenz- with PI-based regimens. This was followed by title, abstract, and full-text screens. The quality of selected studies was assessed using the Cochrane risk of bias tool. Meta-analysis of the odds of virological suppression was conducted using the robust variance estimation approach. Findings: Fifteen studies met the inclusion criteria and totaled 6,712 patients (efavirenz arm=3,339; PI arm=3,373), of which 1,610 (24.0%) were females. Follow-up ranged from 24 to 144 weeks. Mean/median age ranged from 33 to 44 years. Mean/median baseline CD4 count ranged from 32 to 557 cells/mL while mean/median baseline viral load ranged from log10 4·5 to log10 5·5 copies /mL. Meta-analysis showed that patients receiving efavirenz-based regimens had 37% higher odds of virological suppression compared to PI-based regimens (OR=1·37, 95% CI=1·06-1·77, p=0.02). The Egger test suggested the presence of publication bias (B=0·927, t=2·214, p=0·033). The main threat to the quality of evidence was attrition bias. Interpretation: Regarding virological suppression, efavirenz-based regimens were more effective than PI-based regimens and, therefore, might be ideal for the management of treatment naive patients with HIV in settings where NNRTIs and PIs are used. Funding Statement: The authors stated that there were no funding sources for the study. Declaration of Interests: The authors have no conflicts of interests to declare. Ethics Approval Statement: The study was approved by the University of Texas at Austin Institutional Review Board and determined to be non-human subjects research as it involved obtaining information from publicly available data. Hence, informed consent was not required. The study was registered with PROSPERO, an international database supported by PRISMA (ID: CRD42018100296).