The direct association between gout and the increased risk of CVD is still unclear. However, many studies have demonstrated that uric acid, which is strongly associated with the pathogenesis of gout, increases the risk of CV-related comorbidities, including hypertension, type 2 diabetes mellitus, renal disease, and metabolic syndrome [8910141516]. Uric acid-related comorbidities may be associated with the increased risk of CVD. Furthermore, uric acid itself is a pro-oxidant in endothelial tissue. In long-standing hyperuricemia, excessive reactive oxygen species (ROS) production impairs nitric oxide production and induces endothelial dysfunction, which subsequently increase CVD risk [151617]. Furthermore, uric acid uptake by vascular smooth muscle cells and adipocytes leads to increased nicotinamide adenine dinucleotide phosphate (NADPH) oxidase activity, stimulating NADPH oxidase-dependent ROS production [1718]. Uric acid also acts as a danger-associated molecular pattern that provokes systemic inflammation. Inflammation itself also affects CVD risk [19]. On the basis of these observations, hyperuricemia is a major independent risk factor for not only gout but also CVD and its related mortality.

However, evidence of a close association between hypouricemia and increased CVD risk is increasing. Recent studies have reported a relationship between lower serum uric acid levels and CVD, showing a U-shaped distribution [2021222324]. Lee et al. [24] revealed that both lower and higher serum uric acid levels were associated with an increased 10-year predicted CV risk in men based on data from the Korea National Health and Nutrition Examination Survey (2016~2017). Using data from a single university, Kang et al. [23] also demonstrated that a lower serum uric acid level was an independent risk factor of all-cause mortality in men with normal kidney function. Although the association between increased CVD risk and hypouricemia is not fully understood, some possible explanations have been proposed. First, uric acid has antioxidant properties in the extracellular environment, including biologic fluid [1525]. Lee et al. [24] suggested that, as a powerful antioxidant, uric acid scavenges oxygen radicals and protects low-density lipoprotein cholesterol from oxidation, resulting in a protective effect against atherosclerosis [2627]. The loss of antioxidant effect due to low serum uric acid level might increase CVD risk. Second, hypouricemia is a potential predictor of muscle weakness, including sarcopenia. Several studies reported a positive association between serum uric acid levels and handgrip strength and sarcopenia [2228293031]. The increases in the circulating markers of oxidative stress under the sarcopenic state are also related to CVD risk. On the basis of the above-mentioned data on the relationship between serum uric acid level and CVD, controlling uric acid level within its physiological level using ULTs has been hypothesized to improve CV outcomes.

While it is important to treat and prevent acute flare with colchicine, glucocorticoids, and nonsteroidal anti-inflammatory drugs (NSAIDs), the ultimate strategy to control the symptoms and complications of gout is long-term treatment of hyperuricemia. Controlling hyperuricemia can prevent uric acid from depositing in tissues and normalize serum uric acid levels, which can help prevent the recurrence of articular symptoms and extra-articular complications. Therefore, current worldwide guidelines [323334] recommend the active use of ULTs to reach and maintain the target uric acid level (Table 1).

The urate-lowering drugs available in Korea are allopurinol and febuxostat, both of which inhibit uric acid formation, and benzbromarone, which promotes urinary excretion of uric acid. Benzbromarone has the advantage of being safe to use in patients with mildly decreased renal function [35]. However, renal stones and fatal hepatotoxicity have been reported in some patients receiving benzbromarone [353637]. Hence, XOIs are the primary choice of ULT in Korea.

For most patients in whom urate-lowering therapy is indicated, previously we used allopurinol as first-line ULT, typically 100 mg daily with a dose escalation to 100 mg every 2 to 4 weeks to reach the uric acid level below 6 mg/dL [11]. However, skin side effects such as skin rash and hives were frequently reported, and some patients, especially Asians, experienced fatal side effects such as allopurinol-induced severe cutaneous adverse reaction (SCAR). In addition, allopurinol is one of the drugs that frequently induce SCAR and death in Korea [38]. Therefore, since July 2016, febuxostat, another XOI, has also been approved as a first-line urate-lowering drug in Korea. Unlike allopurinol, febuxostat does not have a purine base, and is metabolized mainly in the liver and is excreted equally through the feces and urine. Therefore, the dose in patients with moderate hepatic and renal dysfunction need not be adjusted [39]. Furthermore, febuxostat has rarely been reported to cause hypersensitivity and skin problems, including SCAR. Clinical studies have shown that febuxostat reduces uric acid level more rapidly and effectively than allopurinol [4041424344]. Considering its fast effect and low side effect, febuxostat is currently the most preferred drug for hyperuricemia in Korea. However, the CARES trial, a recent large-scale randomized controlled trial (RCT), insisted that febuxostat increased CV-related and all-cause mortality in gout patients with CV risk as compared with allopurinol [45]. Owing to the unexpected and frightening results of febuxostat, many rheumatologists are confused as to which drug is suitable as first-line ULT in Korea.

As febuxostat reduced elevated uric acid levels, a risk factor of CVD, more significantly than allopurinol, febuxostat was not expected to increase CV events more than allopurinol does [46]. Although the results were not statistically significant, the febuxostat-treated patients reported a slightly higher number of CVD events than the allopurinol-treated patients in three published RCTs, as follows: APEX (febuxostat, allopurinol and placebo-controlled study in gout patients) [41], FOCUS (safety and efficacy study of 40, 80, and 120 mg of febuxostat for 5 years) [44], and EXCEL (long-term safety study between febuxostat 80 and 120 mg and allopurinol in gout patient) [43].

Definitive conclusions regarding CV safety, however, could not be drawn from the earlier trials because the number of events was too small for drawing meaningful conclusions. Thus, a phase 3, double-blind RCT was performed to confirm the safety of febuxostat. The CONFIRMS study [47] examined the comparative urate-lowering effect and safety between febuxostat (40 or 80 mg) and allopurinol (300 mg) in a larger number of patients with gout (N = 2,269). This study analyzed the occurrence of adverse events that met the antiplatelet trialists collaboration (APTC) criteria, such as CV death, nonfatal myocardial infarction, and nonfatal stroke, and non-APTC CV adverse events, including coronary revascularization, transient ischemic attack, congestive heart failure, and arrhythmia. The differences in the rates of the adjudicated non-APTC CV events between the two groups were not statistically significant (1.32% for 40 mg febuxostat, 80 mg 1.19% for febuxostat, and 0.93% for allopurinol). The assessed APTC CV event rates were 0.0% (95% confidence interval [CI] 0.000~0.486) 40 mg for febuxostat, 0.4% (95% CI 0.082~1.155) for 80 mg febuxostat (one patient experienced nonfatal myocardial infarction and two patients experienced nonfatal stroke), and 0.4% (95% CI 0.082~1.155) for allopurinol (two CV-related deaths) [1347]. Although this was a short-term stability test, after the CONFIRMS trial, the US Food and Drug Administration (FDA) approved febuxostat in 2009 for the regulation of uric acid level in patients with gout; however, additional studies comparing the CV safety of febuxostat and allopurinol in gout patients with CV risk factors were requested [1346]. Therefore, a multicenter, randomized, double-blind, non-inferiority post-marketing trial was conducted from April 2010 to May 2017 to compare the CV outcomes of febuxostat therapy with those of allopurinol therapy in patients with gout and CVD.

In the CARES study, 6,190 gout patients with established CV diseases were included and followed up for a median of 32 months. The baseline characteristics, including disease duration, CV risk factors, and CV history in two groups (febuxostat and allopurinol), were well balanced. No differences in primary end-points (first occurrence of CV death, nonfatal myocardial infarction, nonfatal stroke, or urgent revascularization for unstable angina) were observed (hazard ratio [HR], 1.03; 95% CI, 0.87~1.23). However, all-cause (HR, 1.22; 95 CI, 1.01~1.47) and C mortality (HR, 1.34; 95% CI, 1.03~1.73) were higher in the febuxostat group than in the allopurinol group. Moreover, sudden cardiac death occurred more frequently in the febuxostat group (2.7%) than in the allopurinol group (1.8%). On the basis of these results, the authors of the CARES trial insisted that febuxostat has a higher risk of CV mortality than allopurinol [45].

Unlike previous observational cohort studies of XOIs and CV outcomes, the CARES study was the largest prospective double-blind, RCT to directly compare CV outcomes between allopurinol and febuxostat. Furthermore, this study evaluated the effectiveness and safety of high-dose allopurinol (above 300 mg per day) in patients with renal dysfunction, in which 33.6% of patients in the allopurinol group received more than 400 mg per day (400 mg; 25.2%, 500 mg; 4.3%, and 600 mg; 6.1%). The results indicated that, with kidney function monitoring, cautiously increasing allopurinol dose might be safe and effective in patients with renal impairment [13].

Although the CARES RCT is the largest trial that directly compared CV outcome between febuxostat and allopurinol, several uncertainties and biases remain in this study. First, an attrition bias is possible. In the study, more than half of the participants discontinued the trial regimen (56.6%) and follow-up (45.0%) during the study. In addition, an additional 199 deaths (allopurinol, n=110 and febuxostat, n=89) were reported by a search company (Omni Trace) after termination of the study. Post hoc analysis that included these additional deaths showed that the HR of all-cause mortality between the two groups was not statistically significant (HR, 1.09; 95% CI, 0.94~1.28) [12134548]. Second is the possibility of risks of exaggeration and belief that allopurinol is good for CVD, whereas febuxostat has adverse effects on CVD. As no placebo-controlled group was included in the study, we cannot judge whether increased mortality is directly related to febuxostat or not. If both allopurinol and febuxostat have a protective effect to prevent CVD as compared with placebo or other urate-lowering drugs, the increased CV-related mortality risk of febuxostat in the study might not be a big issue [1344]. Third is selection bias. Post hoc subgroup analyses of CV mortality revealed a more frequent use of NSAIDs and less frequent use of low-dose aspirin among patients who died in the febuxostat group than among those who died in the allopurinol group (NSAID use: 4.8% vs. 2.2%, p=0.032; low-dose aspirin use: 3.3% vs. 3.8%, p=0.001). An approximately twofold higher CV mortality rate was observed in patients randomly assigned to febuxostat without aspirin or taking NSAIDs than in those who were receiving allopurinol without aspirin or with low-dose aspirin. No difference in CV mortality was observed among the febuxostat and allopurinol users not taking NSAIDs and taking low-dose aspirin [12]. Lastly, the causal relationship between XOIs and CV mortality is vague. Among the 442 all-cause deaths, only 63 events developed during the on-demand period and 85% of events occurred after discontinuation. Johnson et al. [49] reported a mean incidence rate of CV deaths among patients at risk of 0.17% per month for febuxostat and 0.085% per month during the CARES trial, which increased 18.7- and 18.4-fold to 2.19% and 1.56% per month, respectively, for febuxostat and allopurinol, respectively, during the first 30 days after treatment discontinuation. The reason for the increased rate of major adverse cardiac events after discontinuation of XOI treatment remains unclear; however, the possible explanations include “xanthine oxidase withdrawal syndrome,” in which the cardioprotective effects disappear upon XOI withdrawal [49]. Furthermore, uric acid levels may rebound after XOI discontinuation, which leads to excessive production of ROS related to CVD. A second possible explanation is the “sick-stopper effect” based on a sharply increased mortality rate within 30 days after drug discontinuation in which worsening of other comorbidities may lead to study drug discontinuation and cause poor clinical outcomes [50]. These findings indicate that XOIs are not the only culprit in CV-related deaths.

Following the CARES trial, several studies have compared the CV safety of febuxostat and allopurinol (Table 2) [515253545556575859]. Similar to the CARES trial, one cohort study of gout patients given allopurinol (n=44,111) and febuxostat (n=44,111) that used Taiwan National Health insurance data indicated that febuxostat increased the CV mortality as compared with allopurinol (HR, 1.19; 95% CI, 1.03~1.36) [51]. However, several other studies showed no such association. A retrospective study using US Medicare claims data reported no significant difference in the risk of myocardial infarction, stroke, new-onset heart failure, and coronary revascularization between 24,936 febuxostat initiators and 74,808 allopurinol initiators among a cohort of 77,774 older Medicare patients with gout. Moreover, all-cause mortality also did not differ between the two groups (HR, 0.95; 95% CI, 0.89~1.02) [53]. Recently, data from another prospective RCT, Febuxostat for Cerebral and Cardiorenovascular Events Prevention Study (FREED), were published. This trial recruited elderly patients aged 65 years or older who had hyperuricemia and high CV risk to compare the incidences of cerebral, CV, renal events, and all-cause deaths (the primary composite event) between patients treated with febuxostat (n=537) and those treated with drugs besides febuxostat, including allopurinol (n=533). The primary composite event rate was significantly lower in the febuxostat group than in the non-febuxostat treatment group (HR, 0.75; 95% CI, 0.59~0.95; p=0.017) [59]. Despite the controversial results of febuxostat-related CV risk, both the US and Korean FDAs issued a public safety alert in February 2019 concerning the higher risk of CV death with febuxostat [6061]. In addition, the 2020 American College of Rheumatology (ACR) draft guidelines strongly recommend allopurinol over other ULTs as the preferred first-line agent for all gout patients, including those with CKD stage 3 or worse. They also conditionally recommended that patients receiving febuxostat with a history of CVD or a new CVD-related event switch to an alternative oral ULT if available [62].

We recommend a “wait-and-see” approach. The most significant impact on the FDA recommendations was largely based on the results of the CARES trial; however, Asians comprised only 3% of the CARES trial participants [45]. In this regard, direct implementation of the findings to Korean patients may result in problems due to ethnic differences. Therefore, urate-lowering drugs should be selected on the basis of the results of studies performed in Korean patients. Kang et al. [55] used 2002~2015 Korean National Health Insurance Service data for the entire Korean population to conduct a comparative study of CV occurrence and all-cause mortality between the two drugs, with a 4:1 propensity score matching of 39,640 allopurinol initiators and 9,910 febuxostat initiators. Even though middle-aged patients (mean age, 59 years) were mostly enrolled, this Korean population-based study indicated no difference in the risk of non-fatal CV events and all-cause mortality between allopurinol and febuxostat initiators. Second, the selection of urate-lowering drugs in Korea must consider the risk of allopurinol-induced SCAR. In 2019, the Korean SCAR consortium conducted a retrospective study of 123 patients enrolled in the Korean SCAR registry between 2011 and 2016 to identify the drugs that induced drug rash with eosinophilia and systemic symptoms (DRESS) syndrome. The results showed that allopurinol induced DRESS syndrome most frequently (n=28) [38]. HLA-B*5801, which is closely related to SCAR, is highly expressed especially in Korean, Han Chinese, and Thai populations [3463]. In one study of Koreans, 12.2% of the total study population was positive for HLA-B*5801 and 92% of allopurinol-induced SCARS cases were positive for HLA-B*5801, indicating the association between SCAR and HLA-B*5801 (odds ratio [OR]=98.0) [64]. In a retrospective cohort analysis using HLA genotype data in patients with chronic renal failure among Koreans taking allopurinol, SCAR occurred in 18% of patients with chronic renal failure who were positive for HLA-B*5801 [65]. A 2019 review of 13 RCTs analyzed the differences in mortality and adverse skin reaction between febuxostat and allopurinol in patients with gout or asymptomatic hyperuricemia (sample size: 13,539 patients). This review found no significant difference in the increase of cardiac-related mortality between febuxostat and allopurinol (OR, 0.72; 95% CI, 0.24~2.13; p=0.55) and a significantly lower incidence of adverse skin reactions in the febuxostat group (OR, 0.72; 95% CI, 0.55~0.96; p=0.02). In a subgroup analysis, they separately analyzed Asian and Western data. The non-Asian subgroup of patients who were receiving febuxostat showed higher CV-related mortality than those who were receiving allopurinol (OR, 1.31; 95% CI, 1.01~1.71). However, we found no significant difference in CV-related mortality between the two drugs (OR, 0.73; 95% CI, 0.17~3.23). The occurrence of cutaneous adverse reactions was significantly lower in the Asian patients who were using febuxostat (OR, 0.25; 95% CI, 0.11~0.58) [66].