© The Author 2017. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. Objective. The aim was to examine whether allopurinol use is independently associated with a reduction in the risk of incident peripheral arterial disease (PAD) in the US elderly. Methods. We used the 5% random Medicare sample from 2006 to 2012 to examine the association of allopurinol use and duration of use with the risk or hazard of incident PAD in a retrospective cohort study using a new user design. Multivariable Cox regression models adjusted for demographics, co-morbidity, cardiac medications and cardiac conditions. Hazard ratios (HRs) and 95% CIs were calculated. Results. We identified 26 985 episodes of incident allopurinol use in 25 282 beneficiaries; 3167 allopurinol use episodes (12%) ended in incident PAD. In multivariable-adjusted analyses, allopurinol use was associated with an HR of 0.88 (95% CI: 0.81, 0.95) for incident PAD, as was female gender, HR 0.84 (95% CI: 0.78, 0.90). In a separate multivariable-adjusted model, compared with no allopurinol use, longer durations of allopurinol use were associated with lower HR of PAD: 181 days to 2 years, 0.88 (95% CI: 0.79, 0.97); and >2 years, 0.75 (95% CI: 0.63, 0.89). Other factors significantly associated with a higher HR of PAD were age 75 to<85 and ≥ 85 years, higher Charlson index score and black race. Sensitivity analyses that adjusted for cardiac conditions and medications confirmed these findings, with minimal to no attenuation of HRs. Conclusion. New allopurinol use was independently associated with a lower risk of PAD in the elderly. Longer allopurinol use durations seemed more protective. Mechanisms of the protective effect need to be investigated in future studies.