Comparative Cardiovascular Risks of Febuxostat and Allopurinol in Patients with Diabetes Mellitus and Chronic Kidney Disease

Introduction

Hyperuricemia is characterized by a serum urate level of >6 mg/dL. Gout, which is a type of inflammatory arthritis, manifests as an inflammatory response within the joints due to hyperuricemia [1,2]. Hyperuricemia has been associated with various comorbidities, such as hypertension, diabetes mellitus (DM), cardiovascular (CV) disease, and chronic kidney disease (CKD) [3,4]. Hyperuricemia is closely linked to the onset of diabetes and its chronic complications. Numerous animal and human experiments have verified that uric acid predominantly influences diabetes and its complications through inflammation, oxidative stress, endothelial dysfunction, and other mechanisms [5,6].

The 2020 American College of Rheumatology Guidelines for the management of gout strongly recommended starting urate-lowering therapy in patients with ≥1 subcutaneous tophus, radiographic evidence of gout-related damage, or ≥2 flares per year. For those who have experienced gout attacks 1–2 times per year, urate-lowering therapy is recommended. Additionally, for patients experiencing their first gout attack with concomitant CKD stage ≥3, serum uric acid >9 mg/dL, or urinary stones, urate-lowering therapy initiation is recommended [7]. Febuxostat (a nonpurine xanthine oxidase inhibitor) and allopurinol (a purine base xanthine oxidase inhibitor) are used as first-line urate-lowering agents for the management of hyperuricemia and gout [8].

Allopurinol has been reported to increase the risk of HLA-B*58: 01-mediated cutaneous adverse drug reactions in specific Asian populations [9]. This drug is also associated with the development of various adverse effects in patients with CKD [10–12]. Thus, genotyping is recommended prior to the initiation of allopurinol therapy [13,14]. Furthermore, the drug should be initiated at a low dosage, which should be gradually increased to achieve the target serum uric acid level [15–17]. Despite the limitations, allopurinol remains an effective medication for reducing uric acid levels; it can also mitigate CV risk and decelerate CKD progression [18–20].

Febuxostat is associated with a low risk of allergic reactions [21,22]. Patients generally exhibit good tolerability to this drug and do not require dose modifications. A growing body of evidence supports febuxostat’s efficacy in reducing urate levels and its potential renal-protective properties, highlighting it as an attractive therapeutic option for patients with CKD [23–27]. However, the choice between febuxostat and allopurinol for the treatment of hyperuricemia remains debatable.

Some studies have reported no difference in CV outcomes between febuxostat users and allopurinol users [18,19,23,28–34], whereas other studies have suggested that febuxostat increases the risk of CV-related mortality [26,35–37]. The rates of all-cause mortality were also higher among febuxostat users compared to those using allopurinol. Therefore, patient monitoring should include detection of signs and symptoms indicative of myocardial infarction, stroke, or other possible cardiovascular disease [37,38].

Another meta-analysis revealed that febuxostat neither increases nor reduces the risk of major adverse CV events in patients with hyperuricemia; however, this drug may increase the risk of CV-related mortality in patients with gout and a history of CV events [36]. Hence, considering the potential risk of febuxostat in vulnerable populations, the current study analyzed real-world data to explore its cardiovascular effects in patients with both CKD and DM.

Material and Methods

ETHICS STATEMENT:

The beneficial claim records were all de-identified before analysis; therefore, the requirement for informed consent was waived by the Ethics Review Board of Shin Kong Wu Ho-Su Memorial Hospital (protocol number: 20230714R).

DATA SOURCE:

This study was conducted using CKD thematic data from the National Health Insurance Research Database (NHIRD) [39]. The NHIRD is a vital population-level data repository for generating real-world evidence to inform clinical decisions and healthcare policy-making [40]. It contains the data of approximately 23 million individuals in Taiwan, with records dating back to 1995. This database contains information on patients’ date of birth, sex, diagnostic codes, surgical interventions, and medication use. Before 2015, diseases were coded in the NHIRD by using the diagnostic codes outlined in the International Classification of Diseases (ICD), Ninth Revision, Clinical Modification; subsequently, the codes outlined in the ICD, Tenth Revision, Clinical Modification were used. To ensure patient privacy, the NHIRD data are encrypted before release.

STUDY DESIGN AND COHORT:

This study was designed as a population-based, retrospective cohort study. Figure 1 presents a flowchart showing patient selection. From the NHIRD, we identified patients who received a CKD diagnosis between January 2012 and December 2017, from which we identified patients who received a diagnosis of gout or hyperuricemia between 2012 and 2017 and were prescribed metformin. From this patient population, we identified individuals who had taken febuxostat or allopurinol from the day hyperuricemia was diagnosed up to 180 days after the diagnosis (Figure 2). The index date was determined as the day of hyperuricemia diagnosis, plus 180 days.

We excluded patients who were younger than 18 years or older than 85 years, lacked information on sex, had used both allopurinol and febuxostat between the diagnosis and index dates, received a cancer diagnosis before the index date, were hospitalized for any reason within 3 months before the index date, died before the index date, underwent hemodialysis before the index date, or used erythropoietin within 90 days after the index date (this excluded patients with advanced CKD). Finally, this study included 15 898 patients, of whom 2997 used febuxostat and 12901 used allopurinol. Propensity score matching involves creating matched sets of the febuxostat group and allopurinol group with similar propensity score values. The most common method of propensity score matching is 1: 1 or pair matching [41]. We performed 1: 1 propensity score matching based on sex, age, and Charlson comorbidity index (CCI) scores; each resultant subgroup comprised 2997 patients.

According to the medication usage regulations of Taiwan’s National Health Insurance, the use of allopurinol is limited to treating conditions such as gout, gouty arthritis, uric acid stones, and hyperuricemia resulting from cancer chemotherapy. Febuxostat is restricted to managing hyperuricemia in patients with chronic gout. Therefore, our study population should all have hyperuricemia or gout.

In Taiwan, metformin is the primary pharmacological intervention for managing DM; it is recommended only for patients with an eGFR of ≥30 mL/min/1.73 m². This criterion enabled us to effectively exclude patients with severely compromised renal function (eGFR <30 mL/min/1.73 m²). Our study primarily focused on patients with CKD stages 1 to 3 (eGFR of ≥30 mL/min/1.73 m²), considering the concurrent use of metformin in our study population.

COVARIATES:

Baseline comorbidities, including hypertension, ischemic heart disease, congestive heart failure (HF), atrial fibrillation, peripheral artery disease, stroke, hyperlipidemia, kidney calculi, and liver cirrhosis, were recorded when a patient received at least 3 outpatient diagnoses or 1 inpatient diagnosis within 1 year before the index date. The ICD codes for diseases are displayed in Table 1. The use of other drugs such as angiotensin-converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), beta-blockers, calcium channel blockers (CCBs), antithrombotic agents (eg, anticoagulant and antiplatelet agents), non-steroidal anti-inflammatory drugs (NSAIDs), fenofibrate, statins, and diuretics was recorded when a patient used a given drug for at least 3 months within 1 year before the index date. Table 2 displays the anatomical therapeutic chemical codes.

OUTCOMES:

We estimated incidence rates and HRs for all-cause hospitalization, hospitalization for HF, and hospitalization for CV interventions (including those for cardiac catheterization, percutaneous coronary intervention, or coronary bypass surgery) and compared them between febuxostat users and allopurinol users.

STATISTICAL ANALYSIS:

Continuous data are presented in terms of the mean±SD values, whereas categorical data are presented in terms of the number and percentage values. Baseline characteristics were compared between febuxostat users and allopurinol users by using the t test for continuous variables and the chi-square test for categorical variables.

Competing risk regression was performed for all-cause hospitalization, hospitalization for HF, and hospitalization for CV interventions. The Fine and Gray method [42] was used to estimate the stratified proportional sub-distribution HRs with 95% confidence intervals (CIs) for the effects of exposure and covariates on the cumulative incidence of hospitalization outcomes. An intention-to-treatment basis was adopted according to the patients’ initial medication use status without consideration any subsequent regimen changes. Each individual was followed from the index date until the event occurred or was censored due to death (if it was not the outcome under analysis) or the end of the study period on December 31, 2017.

For all tests, statistical significance was set at a 2-sided P value of <.05. Statistical analyses were conducted using SAS (version 9.4; SAS Institute, Cary, NC, USA).

SENSITIVITY AND SUBGROUP ANALYSES:

To evaluate the robustness of our findings, the index date was extended from 180 days to 360 days after diagnosis of hyperuricemia or gout. Moreover, we used a separate regression model to identify the association between the use of febuxostat and allopurinol. We used a modified stepwise approach with 3 modeling steps, wherein demographic data, comorbidities, and medications were incrementally included in the models. This approach allowed for comprehensive assessment of the association between the indicated variables.

Furthermore, we conducted a subgroup analysis focusing on the outcomes of HF admission and the experience of CV interventions. The subgroups were formed based on age; sex; a history of hypertension, ischemic heart disease, congestive HF, atrial fibrillation, peripheral vascular disease, stroke, hyperlipidemia, kidney calculi, or cirrhosis; and the use of ACEIs/ARBs, beta-blockers, CCBs, and antithrombotic agents.

Results

PATIENT CHARACTERISTICS:

A total of 15 898 patients were included in this study. Following 1: 1 propensity score matching, 2997 patients were included in each group. The baseline characteristics of both groups before and after matching are outlined in Table 3. Before matching, the mean ages of febuxostat (n=2997) and allopurinol (n=12901) users were 65.6±12.0 and 63.5±12.3 years, respectively. In both groups, the proportion of women was higher than that of men. The most common baseline comorbidities among both febuxostat and allopurinol users were hypertension, hyperlipidemia, and ischemic heart disease, which were diagnosed in 58.1%, 41.7%, and 14.6% of all febuxostat users, and in 63.1%, 45.9%, and 14.6% of all allopurinol users, respectively.

Febuxostat users were more likely to receive ACEI/ARB, beta-blockers, antithrombotic agents, statins, and diuretics, whereas allopurinol users were more likely to receive NSAIDs and fenofibrate.

After matching, the incidence rates of most comorbidities appeared similar between the groups, except for hypertension and hyperlipidemia. However, significant differences persisted in the rates of prescribed medications, except for CCB and fenofibrate.

HOSPITALIZATION AND CV OUTCOMES:

Before propensity score matching, febuxostat users exhibited higher incidence rates and HRs for all outcomes than did allopurinol users (all P<.001; Table 4). After propensity score matching, the HRs for all-cause hospitalization, hospitalization for HF, and hospitalization for CV interventions were still significantly higher for febuxostat users than for allopurinol users (all P<.001; HR: 1.33 [95% CI: 1.24 to 1.42], 1.60 [95% CI: 1.42 to 1.81], and 1.52 [95% CI: 1.32 to 1.74], respectively).

In multivariate analyses performed using 3 models (Table 5), the risk of all-cause hospitalization was significantly higher in febuxostat users than in allopurinol users (HR: 1.33 [95% CI: 1.24 to 1.41], 1.33 [95% CI: 1.25 to 1.41], and 1.33 [95% CI: 1.25 to 1.42], respectively). The risk of hospitalization for HF was significantly higher in febuxostat users than in allopurinol users (HR: 1.60 [95% CI: 1.42 to 1.81], 1.59 [95% CI: 1.41 to 1.80], and 1.62 [95% CI: 1.43 to 1.83], respectively). The risk of hospitalization for CV interventions was still significantly higher in febuxostat users than in allopurinol users (HR: 1.52 [95% CI: 1.32 to 1.74], 1.51 [95% CI: 1.32 to 1.73], and 1.51 [95% CI: 1.32 to 1.74], respectively).

The Gray test was performed to estimate the cumulative incidence rates of the 3 outcomes (shown in Figure 3). For all outcomes of all-cause hospitalization (Figure 3A), hospitalization for heart failure (Figure 3B), and hospitalization for CV intervention (Figure 3C), the cumulative incidence rates were significantly higher in febuxostat users than in allopurinol users (P<.0001 for each outcome).

SENSITIVITY ANALYSIS:

When we extended the index date from 180 days to 360 days, almost all outcomes consistently supported the obtained results (Table 6). Only the hospitalization for CV interventions in model 3 exhibited no significant between-group difference but still revealed the same trend (HR: 1.37 [95% CI: 1.09 to 1.71], 1.28 [95% CI: 1.02 to 1.60], and 1.24 [95% CI: 0.99 to 1.55], respectively).

SUBGROUP ANALYSIS:

We performed a series of stratified analyses to validate the reliability of our findings (shown in Figure 4). In almost all subgroups – except for patients aged <50 years, those with peripheral vascular disease, those with kidney calculi, and those receiving fenofibrate – significant differences were noted between febuxostat users and allopurinol users in terms of hospitalization for HF (Figure 4A) and that for CV interventions (Figure 4B). However, no significant between-group difference was noted in the groups of patients with stroke who were hospitalized for HF and those with congestive HF and atrial fibrillation who were hospitalized for CV interventions.

Discussion

STRENGTHS AND LIMITATIONS:

The current study has several notable strengths. First, previous studies evaluating the effects of febuxostat have predominantly focused on the general population, leading to inconsistent CV results. However, our study focused specifically on patients with DM and stage 1 to 3 CKD. We investigated whether febuxostat increases the risk of CV events in the susceptible population compared to allopurinol. Second, the national insurance claims database provided us with a sufficient sample size and study outcomes to draw meaningful inferences in our study. Third, the sensitivity analysis and subgroup analysis showed consistent results, suggesting the robustness of the study findings. Finally, the study groups were effectively balanced using propensity score methods and adjusted for baseline covariates to eliminate the disparities between groups. Otherwise, the outcomes and baseline comorbidities in present study were delineated through validated diagnostic codes, ensuring accuracy and reliability in their definition [52,53].

Although these strengths are impressive, this study has certain limitations. To begin with, essential potential confounding factors like body mass index, blood pressure, lifestyle, and laboratory data were lacking in the NHI database, all of which are linked to CVD. Secondly, this was not a randomized controlled trial (RCT), leading to a significant concern about the uneven baseline between the 2 groups, which could have introduced bias due to confounding by indication. Nevertheless, the use of propensity score matching and multivariable adjustments in the regression model can help mitigate these effects. Lastly, our study focused solely on Taiwanese patients. Consequently, the results require further confirmation before broader application to other populations worldwide.

Conclusions

Our findings indicate that febuxostat may increase CV risk, particularly in patients with preexisting risk factors for CV events, such as CKD and DM. Thus, in addition to careful consideration before prescribing febuxostat, close monitoring of potential adverse events is imperative. Additional randomized controlled trials are needed to validate our results.