Background: Current guidelines recommend potent platelet inhibition with ticagrelor or prasugrel in patients after an acute coronary syndrome. However, data about optimal platelet inhibition in older patients are scarce. We aimed to investigate the safety and efficacy of clopidogrel compared with ticagrelor or prasugrel in older patients with non-ST-elevation acute coronary syndrome (NSTE-ACS). Methods: We did the open-label, randomised controlled POPular AGE trial in 12 sites (ten hospitals and two university hospitals) in the Netherlands. Patients aged 70 years or older with NSTE-ACS were enrolled and randomly assigned in a 1:1 ratio using an internet-based randomisation procedure with block sizes of six to receive a loading dose of clopidogrel 300 mg or 600 mg, or ticagrelor 180 mg or prasugrel 60 mg, and then a maintenance dose for the duration of 12 months (clopidogrel 75 mg once daily, ticagrelor 90 mg twice daily, or prasugrel 10 mg once daily) on top of standard care. Patient and treating physicians were aware of the allocated treatment strategy, but the outcome assessors were masked to treatment allocation. Primary bleeding outcome consisted of PLATelet inhibition and patient Outcomes (PLATO; major or minor bleeding [superiority hypothesis]). Co-primary net clinical benefit outcome consisted of all-cause death, myocardial infarction, stroke, PLATO major and minor bleeding (non-inferiority hypothesis, margin of 2%). Follow-up duration was 12 months. Analyses were done on intention-to-treat basis. This trial is registered with the Netherlands Trial Register (NL3804), ClinicalTrials.gov (NCT02317198), and EudraCT (2013–001403–37). Findings: Between June 10, 2013, and Oct 17, 2018, 1002 patients were randomly assigned to clopidogrel (n=500) or ticagrelor or prasugrel (n=502). Because 475 (95%) patients received ticagrelor in the ticagrelor or prasugrel group, we will refer to this group as the ticagrelor group. Premature discontinuation of the study drug occurred in 238 (47%) of 502 ticagrelor group patients randomly assigned to ticagrelor, and in 112 (22%) of 500 patients randomly assigned to clopidogrel. Primary bleeding outcome was significantly lower in the clopidogrel group (88 [18%] of 500 patients) than in the ticagrelor group (118 [24%] of 502; hazard ratio 0·71, 95% CI 0·54 to 0·94; p=0·02 for superiority). Co-primary net clinical benefit outcome was non-inferior for the use of clopidogrel (139 [28%]) versus ticagrelor (161 [32%]; absolute risk difference −4%, 95% CI −10·0 to 1·4; p=0·03 for non-inferiority). The most important reasons for discontinuation were occurrence of bleeding (n=38), dyspnoea (n=40), and the need for treatment with oral anticoagulation (n=35). Interpretation: In patients aged 70 years or older presenting with NSTE-ACS, clopidogrel is a favourable alternative to ticagrelor, because it leads to fewer bleeding events without an increase in the combined endpoint of all-cause death, myocardial infarction, stroke, and bleeding. Clopidogrel could be an alternative P2Y12 inhibitor especially for elderly patients with a higher bleeding risk. Funding: ZonMw.