Background/Purpose: Hyperuricemia and gout are associated with an increased risk of type 2 diabetes (T2D). Xanthine oxidase inhibitors (XOI), allopurinol and febuxostat, are the main therapy to treat gout patients with hyperuricemia. Little is known whether treating hyperuricemia with a XOI has any effect on future risk of T2D. We examined the risk of T2D in gout patients initiating a XOI versus untreated patients with hyperuricemia.    

Methods: We conducted a cohort study using a U.S. commercial insurance claims database. Patients aged ≥40 years with gout and hyperuricemia (≥ 6.8mg/dl) who had an enrollment period for ≥ 365 days were eligible. Propensity score (PS) matching was used to simultaneously control for baseline demographic factors, comorbidities, medications, health care utilization, and time trend. From January 2004 to December 2012, XOI initiators and non-initiators matched on a PS were identified with a 1:2 ratio in each calendar month (a total of 108 calendar months). The first day of each month was the index date for both groups. We excluded patients with diabetes, use of XOI or anti-diabetic drugs, end-stage renal disease and renal transplantation prior to the index date. Follow-up continued until the outcome occurrence, discontinuation or initiation of XOI, disenrollment, or administrative censoring. We calculated incidence rates (IR) of T2D based on a new diagnosis of T2D and a receipt of anti-diabetic medication. Due to violation of the proportional hazards assumption, Cox proportional hazards models stratified by treatment duration compared the risk of T2D in XOI initiators versus non-initiators.

Results: There were 4,045 XOI initiators and 8,090 non-initiators. Baseline characteristics were well-balanced between the matched groups. Mean age was 54 years and 89% male in both groups. Common comorbidities include hypertension (64%), hyperlipidemia (61%), CVD (10%), obesity (10%) and chronic kidney disease (8%). Use of systemic steroids at baseline was common (33%). The mean serum uric acid level at baseline was 8.9 mg/dl in XOI initiators and 8.3 mg/dl in non-initiators.  The mean HgbA1c level at baseline was 5.9% in XOI initiators and 5.8% in non-initiators. The IR of T2D per 100 person-years was 1.88 (95% CI 1.41-2.51) in XOI initiators and 1.57 (95% CI 1.36-1.81) in non-initiators.  XOI treatment for 0-90 days was associated with an increased risk of T2D versus non-initiators, whereas the use of XOI for longer than 360 days may be associated with a decreased risk of T2D (Table).

Conclusion: Nearly 2% of gout patients were newly diagnosed with T2D during follow-up. Short-term use of XOI was associated with a greater risk of T2D in gout patients compared to non-initiators, but a potential long-term beneficial effect of XOI on T2D cannot be excluded. Future research such as a randomized clinical trial ensuring treatment adherence may be needed to examine the long-term effect of XOI on T2D.