OBJECTIVES: There is no head-to-head trial demonstrating the risk of myocardial infarction (MI) across angiotensin receptor blockers (ARBs). This review assessed the MI incidences associated with the use of ARBs through network meta-analyses of randomised controlled trials (RCTs). METHODS: Embase® and MEDLINE® were searched until June 2012 for RCTs assessing the safety of approved ARBs versus active control/placebo. Studies were included based on a pre-specified protocol. Data were extracted by two reviewers, with any discrepancy being reconciled by a third, independent reviewer. A network meta-analysis was conducted and probability-based rank (P of being best) was generated using WinBUGS®. Subgroup analyses by disease cohorts (hypertension, heart failure, and diabetes) were also performed. RESULTS: Of the 3099 studies, 33 RCTs enrolling 143,205 patients were included. Most studies reported data for up-titrated doses with continued background therapy. No significant differences were observed between ARBs and controls in meta-analysis (relative risk [95% confidence interval]: 1.04 [0.99-1.10] vs. active-control; 0.94 [0.83-1.07] vs. placebo). Network meta-analyses indicated no significant difference among the ARBs. Subgroup analyses by disease cohorts showed no significant difference between ARBs and active/placebo as well as within ARBs. The probability of least MI incidence was highest with olmesartan 40 mg (P=41%) followed by candesartan 16 mg (P=19%), losartan 200 mg (P=10%), valsartan 80 mg (P=9%), and candesartan 4 mg-8 mg (P=4%-5%). Conversely, the probability of least MI incidence was <1% with telmisartan 80 mg, irbesartan 300 mg, losartan 100 mg, valsartan 160 mg. A dose-risk assessment within ARBs remained inconclusive due to limited data. CONCLUSIONS: ARBs did not differ significantly from active control/placebo or from each other for the risk of MI. However, per probability-based ranking, olmesartan 40 mg may present the least risk of MI. Limited data warranted careful interpretation of these results and further research in this area using monotherapy data.