OBJECTIVES: To estimate the relative efficacy of clopidogrel versus low-dose aspirin plus extended-release dipyridamole (ASA+ERDP) in preventing serious vascular events among stroke patients. Additionally, to test whether a network meta-analysis (NMA) can give reliable estimates of treatments’ relative efficacy in the absence of direct evidence. METHODS: A systematic literature review was conducted in EMBASE and MEDLINE to identify randomized controlled trial (RCT) evidence on the endpoint “serious vascular events” (including myocardial infarction, stroke and vascular death). A NMA with fixed effects was fitted to the data using Winbugs. The NMA was first run on indirect evidence only (NMA_indirect), and these results were compared with results from a NMA including direct and indirect evidence (NMA_all). RESULTS: Three RCTs were identified (ESPS2, ESPRIT, CAPRIE) comparing clopidogrel or ASA+ERDP with aspirin; one RCT (PRoFESS) provided direct evidence, two RCTs (CHARISMA and MATCH) provided additional data on clopidogrel through the link with aspirin+clopidogrel. Furthermore, one meta-analysis (ATC) compared the efficacy of different aspirin dosages; this was added in the network to link the aspirin arms which differed in dosage. The odds ratio (OR) of ASA+ERDP versus clopidogrel for NMA_indirect was 1.15 (0.95 : 1.37); PRoFESS reported 0.99 (0.92 : 1.07); and NMA_all resulted in 1.02 (0.95 : 1.10). All analyses have OR close to “1” and confidence intervals overlapping “1”. The point estimate of PRoFESS is within the confidence bounds of NMA_indirect, but the OR from these analyses are in opposite directions. The confidence interval of NMA_all tightens when adding PRoFESS into the network. CONCLUSIONS: The results of the analysis indicate that clopidogrel and ASA+ERDP are of equivalent efficacy in preventing secondary serious vascular events. Furthermore, in the absence of direct evidence, statistical techniques such as NMA can provide a reasonable estimate of relative efficacy.