Osteoarthritis (OA) and the non-steroidal anti-inflammatory drugs (NSAIDs) used to relieve OA-associated pain have been linked independently to increased cardiovascular risk. We examined the risk of cardiovascular events associated with NSAID use in patients with OA. We employed linked nationwide administrative registers to examine NSAID use between 1996 and 2015 by Danish patients with OA aged ≥18 years. Using adjusted Cox proportional hazard analyses, we calculated the risk of the composite outcome of cardiovascular death, non-fatal myocardial infarction and non-fatal ischaemic stroke/TIA, and of each outcome separately, up to 5 years after OA diagnosis. Of 533 502 patients included, 64.3% received NSAIDs and 38 226 (7.2%) experienced a cardiovascular event during follow-up. Compared with non-use, all NSAIDs were associated with increased risk of the composite outcome: hazard ratio (HR) for rofecoxib, 1.90 (95% confidence interval, 1.74-2.08); celecoxib, 1.47 (1.34-1.62); diclofenac, 1.44 (1.36-1.54); ibuprofen, 1.20 (1.15-1.25); and naproxen, 1.20 (1.04-1.39). Similar results were seen for each outcome separately. When celecoxib was used as reference, ibuprofen (HRs: 0.81 [CI: 0.74-0.90]) and naproxen (HRs: 0.81 [0.68-0.97]) exhibited a lower cardiovascular risk, even when low doses were compared. Low-dose naproxen and ibuprofen were associated with the lowest risks of the composite outcome compared to no NSAID use: HRs: 1.12 (1.07-1.19) and 1.16 (0.92-1.42), respectively. In patients with OA, we found significant differences in cardiovascular risk among NSAIDs. Naproxen and ibuprofen appeared to be safer compared to celecoxib, also when we examined equivalent low doses. In terms of cardiovascular safety, naproxen and ibuprofen, at the lowest effective doses, may be the preferred first choices among patients with OA needing pain relief.