Introduction: Previous studies have shown an association between gout and/or hyperuricemia and a subsequent increase in cardiovascular disease (CVD) outcomes. Allopurinol reduces vascular oxidative stress, ameliorates inflammatory state, improves endothelial function, and prevents atherosclerosis progression. Accordingly, we tested the hypothesis that a positive association between allopurinol therapy in gout patients and future cardiovascular outcomes is present using a population-based matched-cohort study design.
Methods: Patients aged ≥40 years with newly diagnosed gout having no pre-existing severe form of CVD were separated into allopurinol (n = 2483) and non-allopurinol (n = 2483) groups after matching for age, gender, index date, diabetes mellitus, hypertension, hyperlipidemia, and atrial fibrillation. The two groups were also balanced in terms of uric acid nephrolithiasis, acute kidney injury, hepatitis, and Charlson comorbidity index.
Results: With a median follow-up time of 5.25 years, the allopurinol group had a modest increase in cardiovascular risk [relative risk, 1.20; 95% confidence interval (CI), 1.08-1.34]. A Cox proportional hazard model adjusted for chronic kidney disease, uremia, and gastric ulcer gave a hazard ratio (HR) for cardiovascular outcomes of 1.25 (95% CI, 1.10-1.41) in gout patients receiving allopurinol compared with the non-allopurinol group. In further analysis of patients receiving urate-lowering therapy, the uricosuric agent group (n = 1713) had an adjusted HR of 0.83 (0.73-0.95) for cardiovascular events compared with the allopurinol group.
Conclusions: The current population-based matched-cohort study did not support the association between allopurinol therapy in gout patients with normal risk for cardiovascular sequels and beneficial future cardiovascular outcomes. Several important risk factors for cardiovascular disease, such as smoking, alcohol consumption, body mass index, blood pressure were not obtainable in the current retrospective cohort study, thus could potentially bias the effect estimate.