Background The effectiveness of durable polymer drug-eluting stents comes at the expense of delayed arterial healing and subsequent late adverse events such as stent thrombosis (ST). We report the 4 year follow-up of an assessment of biodegradable polymer-based drug-eluting stents, which aim to improve safety by avoiding the persistent inflammatory stimulus of durable polymers. Methods We did a multicentre, assessor-masked, non-inferiority trial. Between Nov 27, 2006, and May 18, 2007, patients aged 18 years or older with coronary artery disease were randomly allocated with a computer-generated sequence to receive either biodegradable polymer biolimus-eluting stents (BES) or durable polymer sirolimus-eluting stents (SES; 1: 1 ratio). The primary endpoint was a composite of cardiac death, myocardial infarction, or clinically-indicated target vessel revascularisation (TVR); patients were followed-up for 4 years. Analysis was by intention to treat. This trial is registered with ClinicalTrials.gov, number NCT00389220. Findings 1707 patients with 2472 lesions were randomly allocated to receive either biodegradable polymer BES (857 patients, 1257 lesions) or durable polymer SES (850 patients, 1215 lesions). At 4 years, biodegradable polymer BES were non-inferior to durable polymer SES for the primary endpoint: 160 (18.7%) patients versus 192 (22.6%) patients (rate ratios [RR] 0.81, 95% CI 0.66-1.00, p for non-inferiority <0.0001, p for superiority=0.050). The RR of definite ST was 0.62 (0.35-1.08, p=0.09), which was largely attributable to a lower risk of very late definite ST between years 1 and 4 in the BES group than in the SES group (RR 0.20, 95% CI 0.06-0.67, p=0.004). Conversely, the RR of definite ST during the first year was 0.99 (0.51-1.95; p=0.98) and the test for interaction between RR of definite ST and time was positive (p(interaction)=0.017). We recorded an interaction with time for events associated with ST but not for other events. For primary endpoint events associated with ST, the RR was 0.86 (0.41-1.80) during the first year and 0.17 (0.04-0.78) during subsequent years (p(interaction)=0.049). Interpretation Biodegradable polymer BES are non-inferior to durable polymer SES and, by reducing the risk of cardiac events associated with very late ST, might improve long-term clinical outcomes for up to 4 years compared with durable polymer SES.