Background and purpose: Patients actively smoking at the time of primary hip or knee arthroplasty are at increased risk of direct perioperative complications. We investigated the association between smoking status and risk of revision and mortality within 2 years following hip or knee arthroplasty. Methods: We used prospectively collected data from the Dutch Arthroplasty Register. All primary total hip arthroplasties (THAs), total knee arthroplasties (TKAs), and unicondylar knee arthroplasties (UKAs) with > 2 years’ follow-up were included (THA: n = 140,336; TKA: n = 117,497; UKA: n = 14,807). We performed multivariable Cox regression analyses to calculate hazard risks for differences between smokers and non-smokers, while adjusting for confounders (aHR). Results: The smoking group had higher risk of revision (THA: aHR 1.3, 95% confidence interval [CI] 1.1–1.4 and TKA: aHR 1.4, CI 1.3–1.6) and risk of mortality (THA: aHR 1.4, CI 1.3–1.6 and TKA: aHR 1.4, CI 1.2–1.6). Following UKA, smokers had a higher risk of mortality (aHR 1.7, CI 1.0–2.8), but no differences in risk of revision were observed. The smoking group had a higher risk of revision for infection following TKA (aHR 1.3, CI 1.0–1.6), but not following THA (aHR 1.0, CI 0.8–1.2). Conclusion: This study showed that the risk of revision and mortality is higher for smokers than for non-smokers in the first 2 years following THA and TKA. Smoking could contribute to complications following primary hip or knee arthroplasty.