© Article author(s) (or their employer(s) unless otherwise stated in the text of the article) 2017. All rights reserved. Objective T o assess the comparative effectiveness of allopurinol versus febuxostat for preventing incident renal disease in elderly. Methods In a retrospective cohort study using 2006-2012 Medicare claims data, we included patients newly treated with allopurinol or febuxostat (baseline period of 183 days without either medication). We used 5:1 propensity-matched Cox regression analyses to compare the HR of incident renal disease with allopurinol use (and dose) versus febuxostat (reference). Sensitivity analyses included multivariable-adjusted regression models. Results T here were 31 465 new allopurinol or febuxostat treatment episodes in 26 443 patients; 8570 ended in incident renal disease. Crude rates of incident renal disease per 1000 person-years were 192 with allopurinol versus 338 with febuxostat. Crude rates of incident renal disease per 1000 person-years were lower with higher daily dose: allopurinol <200, 200-299 and ≥300 mg/day with 238, 176 and 155; and febuxostat 40 and 80 mg/day with 341 and 326, respectively. In propensity-matched analyses, compared with febuxostat, allopurinol use was associated with lower HR of incident renal disease, 0.61 (95% CI 0.49 to 0.77). Compared with febuxostat 40 mg/day, allopurinol doses <200, 200-299 and ≥300 mg/day were associated with lower HR of incident renal disease, 0.75 (95% CI 0.65 to 0.86), 0.61 (95% CI 0.52 to 0.73) and 0.48 (95% CI 0.41 to 0.55), respectively. Sensitivity analyses using multivariable-adjusted regression confirmed these findings. Conclusions Allopurinol was associated with a lower risk of incident renal disease in elderly patients than febuxostat. Future studies need to examine the mechanism of this potential renal benefit of allopurinol.