Febuxostat和Allopurinol之心血管安全性比較：以傾向分數配對之世代研究

Abstract

目的： Febuxostat和allopurinol為痛風及高尿酸血症的首選藥品治療，在藥理學分類上皆屬xanthine oxidase inhibitors (XOIs)。然而，美國Food and Drug Administration (FDA)不良事件通報系統的研究顯示febuxostat有潛在較高心血管栓塞不良事件的訊號，而在臨床試驗中也發現病人使用febuxostat相較於使用allopurinol後有顯著較高的心血管死亡和全死因死亡風險。本研究之目的係探討febuxostat使用者相較於allopurinol使用者是否增加心血管相關事件和死亡的風險，同時亦分析心血管不良事件是否和febuxostat的藥品使用劑量間存在劑量反應效果。 方法： 本研究為一回溯性世代研究，以臺灣全民健康保險資料庫為研究材料。研究納入於2012年4月1日至2015年12月31日間首次處方XOIs的病人，並分為febuxostat使用者和allopurinol使用者兩組，為平衡兩組病人之基礎特性，以傾向分數、指標日期 (index date)、有無腎功能不全病史和有無使用benzbromarone作為配對條件進行1：1配對。研究終點為主要心血管事件 (包含心肌梗塞、缺血性中風和心血管相關死亡)、靜脈栓塞、心血管疾病相關住院、心血管相關死亡和全死因死亡，使用Cox比例風險模式 (Cox proportional hazards models) 進行統計分析，估計兩組病人發生研究終點之風險比 (hazard ratios，HRs) 和95%信賴區間 (confidence intervals，CIs)。本研究亦將XOIs使用者之每日平均劑量以定義每日劑量 (defined daily dose，DDD) 表示，分為低劑量 (≤0.5 DDD)、中劑量 (>0.5 DDD, ≤1 DDD) 和高劑量 (>1 DDD) 3個類別，並以此作為模型中隨時間變化的變數，分析使用中、高劑量febuxostat相對於低劑量febuxostat發生研究終點事件的HRs和95% CIs。 結果： 在配對後兩組各包含44,111位病人，所有共變數在兩組間均平衡。在根據治療 (as-treated) 分析方法中，febuxostat使用者相較於allopurinol使用者有顯著較高的心衰竭住院 (HR=1.22；95% CI：1.13-1.33；兩組事件發生率分別為每1000人年60.1和51.0)、心房顫動住院 (HR=1.19；95% CI：1.05-1.36；兩組事件發生率分別為每1000人年23.7和20.7) 及心血管相關死亡 (HR=1.19；95% CI：1.03-1.36；兩組事件發生率分別為每1000人年19.2和17.2) 發生風險，而在主要心血管事件、靜脈栓塞、心肌梗塞住院、缺血性中風住院及全死因死亡無顯著相關。其中，多項敏感性分析結果均支持febuxostat的使用增加心衰竭住院風險這項研究結果。本研究之研究對象接受febuxostat或allopurinol的平均日劑量分別為53.1 mg (0.66 DDD) 和133.3 mg (0.33 DDD)，劑量反應分析結果發現使用中、高劑量的febuxostat相較於使用低劑量的febuxostat，病人發生心血管事件的風險呈階梯式的增高。 結論： Febuxostat使用者相較於allopurinol使用者有顯著較高的心衰竭住院、心房顫動住院及心血管相關死亡發生風險，尤以增加心衰竭住院風險之結果更具穩健性；此外，febuxostat的使用劑量和心血管事件風險具劑量反應效果。我們建議臨床決策者在處方febuxostat時應追蹤心血管相關症狀和檢驗數值，未來需進一步的研究以了解febuxostat增加心血管事件風險之機轉。

Objective: Febuxostat and allopurinol are both xanthine oxidase inhibitors (XOIs) and recommended as the first-line pharmacological therapy for patients with hyperuricemia or gout. However, the U.S. Food and Drug Administration (FDA) Adverse Event Reporting System indicated potential signals of febuxostat-associated cardiovascular (CV) thromboembolic events, and one clinical trial suggested a higher rate of all-cause and CV mortality with febuxostat than with allopurinol. This study aimed to assess whether febuxostat, compared to allopurinol, is associated with a higher risk of mortality and CV events. Also, the dose-response effects of febuxostat on the CV outcomes were analyzed. Methods: We conducted a retrospective cohort study using the National Health Insurance Research Database (NHIRD) in Taiwan. We identified patients who were first prescribed XOIs between 2012/4/1 and 2015/12/31 and classified them into groups of febuxostat users and allopurinol users. The two groups were 1:1 matched by the propensity score (PS), index date, history of renal impairment and benzbromarone use to balance the patient characteristics. Outcomes of interest were major adverse cardiovascular event (MACE, including myocardial infarction, ischemic stroke, and CV death), venous thromboembolism (VTE), CV-related hospitalization, CV death, and all-cause mortality. Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) of the outcomes. In addition, to study the dose-response relationship, mean daily dose of XOIs users were estimated, categorized into low dose (≤0.5 defined daily dose, DDD), medium dose (>0.5 DDD, ≤1 DDD), and high dose (>1 DDD), and included as time-dependent variables in the Cox models. Results: We included 44,111 patients in each group after matching, and all the covariates were balanced between the two groups. In the as-treated analyses, febuxostat, compared with allopurinol, was associated with a significantly higher risk of heart failure (HR, 1.22; 95% CI, 1.13 to 1.33; incidence, 60.1 vs. 51.0/1000 person-years), atrial fibrillation (HR, 1.19; 95% CI, 1.05 to 1.36; incidence, 23.7 vs. 20.7/1000 person-years), and CV death (HR, 1.19; 95% CI, 1.03 to 1.36; incidence, 19.2 vs. 17.2/1000 person-years). However, a similar risk was found in MACE composite endpoint, VTE, myocardial infarction, ischemic stroke, and all-cause mortality. The elevated risk of heart failure was consistent throughout the main analysis and sensitivity analyses. The mean daily dose of febuxostat users and allopurinol users were 53.1 mg (0.66 DDD) and 133.3 mg (0.33 DDD), respectively. The use of medium- and high-dose febuxostat, compared with low-dose febuxostat, showed a hierarchically increased risk of CV outcomes. Conclusions: In the real-world setting, the use of febuxostat, compared with allopurinol, was associated with an increased risk of heart failure, atrial fibrillation, and CV death. The elevated risk of heart failure was found to be robust throughout the analyses. In addition, higher dose of febuxostat may result in a higher rate of CV outcomes. We recommend that health care professionals should carefully access CV-related symptoms and lab data in their patients when prescribing febuxostat. Further studies are needed to investigate the underlying mechanism of the CV adverse outcomes.