Allopurinol and the risk of ventricular arrhythmias in the elderly: a study using US Medicare data

Abstract

Background: There are no published human studies investigating whether the use of allopurinol, the most commonly used medication for the treatment of hyperuricemia in gout, the most common type of inflammatory arthritis in adults, has any beneficial effects on ventricular electrophysiology. The objective of our study was to assess whether allopurinol use is associated with a reduction in the risk of ventricular arrhythmias (VA).

Methods: We used the 5% random sample of Medicare beneficiaries from 2006-2012 to examine new allopurinol use and the risk of incident VA. Multivariable Cox regression analyses were adjusted for demographics (age, race, sex), comorbidity, cardiac medications, and conditions associated with VA. We calculated hazard ratios (HR) and 95% confidence intervals (CI).

Results: Of the 28,755 episodes of new allopurinol use, 2538 were associated with incident VA (8.8%). Among patients with incident VA, 54% were male, 78% were White, 75% had gout as the underlying diagnosis, and the mean Charlson-Romano comorbidity score was 4.8. The crude incidence of VA per 1,000,000 person-days declined as the duration of allopurinol use increased: 1-180 days, 151; 181 days to 2 years, 105; and > 2 years, 85. In multivariable-adjusted analyses, compared to non-use, allopurinol use was associated with lower HR of VA of 0.82 (95% CI, 0.76-0.90). Compared to allopurinol non-use, longer allopurinol use durations were significantly associated with lower multivariable-adjusted HR for VA: 1-180 days, 0.96 (95% CI, 0.85-1.08); 181 days to 2 years, 0.76 (95% CI, 0.68-0.85); and > 2 years, 0.72 (95% CI, 0.60-0.87). Multiple sensitivity analyses adjusting for cardiac conditions, anti-arrhythmic drugs and alternate definitions confirmed our findings with minimal/no attenuation of estimates.

Conclusion: Allopurinol use and use duration of more than 6 months were independently associated with a lower risk of VA. Future studies need to assess the pathophysiology of this potential benefit.

Keywords: Allopurinol; Elderly; Medicare; Risk factor; Ventricular arrhythmias.

Fig. 1

Patient selection flow chart. The flow chart shows the selection of new allopurinol exposure episodes after applying all the eligibility criteria, including an absence of VA and the absence of any allopurinol filled prescription in the baseline period of 365 days (new user design). We found 28,755 new allopurinol exposure episodes in 26,905 patients. Of these, 2538 ended in incident VA and 26,217 ended without incident VA. * We followed each eligible patient with a new filled allopurinol prescription until the patient lost full Medicare coverage, had VA (the outcome of interest), died or reached the each of the study period on December 31, 2012, whichever came first. For some of these patients, the VA occurred on days covered by allopurinol exposure (n = 1525), yet other patients had periods of no allopurinol exposure after an initial qualifying allopurinol exposure during which VA occurred (n = 1013). Nb Number of beneficiaries, T _E treatment episodes, Np Number of allopurinol prescriptions, NE Number of qualified episodes of new allopurinol prescriptions, VA Ventricular arrhythmia

Fig. 2

Examining the effect of previous myocardial infarction (MI) on the associations of allopurinol use (2a) and allopurinol use duration (2b, 2c) with incident ventricular arrhythmias (VA). a Association of allopurinol use with VA by previous MI: Models 1 and 5. b Association of allopurinol use duration with VA by previous MI: Model 2. c Association of allopurinol use duration with VA by previous MI: Model 6. Each solid bar represents the hazard ratio estimate for allopurinol use (vs. non-use) for both Models 1 and 5 (panel a) or allopurinol use duration for Model 2 (panel b; multivariable model adjusted for demographics, Charlson–Romano score, beta blockers, diuretics, ACE inhibitors and statins) and model 6 (panel c; multivariable model adjusted for demographics, beta blockers, diuretics, ACE inhibitors, statins, VA risk factor conditions, aspirin, digoxin, calcium channel blockers, amiodarone, flecainide, and ranolazine), each panel given the presence or absence of previous MI. A hazard ratio of 1.0 represents the reference hazard with no exposure to allopurinol. Error bars represent the 95% confidence interval for each hazard ratio and inclusion of 1.0 in this range indicates that the hazard ratio is not significant