The US Food and Drug Administration has recently released a safety announcement  reporting an increased risk of heart-related deaths and death from all causes in the treatment arm assigned to febuxostat in the CARES (CArdiovascular safety of febuxostat and allopurinol in patients with gout and cardiovascular comoRbiditiES) trial . This large (over 6,000 patients with prevalent cardiovascular [CV] disease) double-blind randomized trial has been run over a 5-year follow-up period to compare cardiovascular risk in patients with gout treated with febuxostat compared with those randomized to allopurinol as comparator. The primary CV end point for this trial is a composite of CV death, nonfatal myocardial infarction, nonfatal stroke, and unstable angina requiring urgent coronary revascularization. To date, no difference has been shown between randomization arms for the primary composite outcome, but febuxostat showed an increased risk of the secondary outcomes of heart-related death and death from all causes.
Two other prospective, randomized, open label with blinded endpoint (PROBE) studies are currently ongoing: FAST (Febuxostat versus Allopurinol Streamlined Trial) , an open-label trial recruiting in British and Danish gout populations, and FREED (Febuxostat for cerebral and caRdiorenovascular events prEvEntion stuDy) , which was conducted in hyperuricemic Japanese subjects. To date, neither of these studies has reported any signal of differences in CV risk between treatment arms. The FAST and CARES studies are a post-approval requirement by the US Food and Drug Administration (FDA) and European Medicines Agency (EMA) in response to an increased, although not statistically significant, number of CV events in the population assigned to febuxostat.
Regardless of whether these two studies are in agreement with the CARES trial regarding safety signals, it is important to know that there are important differences in the design of these trials, a summary of which can be viewed below:
Table 1: Summary of key differences between CARES, FAST and FREED trials
Allopurinol dose range (mg/day)
Febuxostat dose range (mg/day)
Male patients >50 years old
Female patients >55 years old with a prevalent CV event or diabetic vascular involvement
Composite: CV death, non-fatal MI, non-fatal stroke, unstable angina requiring revascularization
United Kingdom, Sweden, and Denmark
Gout patients >60 years old with a prevalent CV event, or at risk of having a CV event
Previous CV event
Up-dosing lead-in phase** followed by washout period
Composite: any event included in the Antiplatelet
Trialists’ Collaboration (APTC) composite endpoint, until 456 events have occurred for the non-inferiority comparison
Hyperuricemia ***patients >65 years old at risk of, or with a prevalent CV event, or chronic kidney disease
Multiple variables including gender, renal function, diabetes, recent CV event, baseline serum urate level, and institution
Composite: CV-renal death, new or recurring cerebrovascular or ischemic heart event, progression of renal disease or end stage renal disease, heart failure requiring in-hospital care, peripheral vascular disease requiring treatment, new atrial fibrillation
*Corrected by renal function, to reach sUA< 6 mg/dL (0.36 mmol/L). ** to reach sUA < 6 mg/dL (0.36 mmol/L)in all patients, then randomized. ***sUA between 7 and 9 mg/dL (between 0.42 mmol/L and 0.54 mmol/L); CV: cardiovascular; MI: myocardial infarction; PROBE: prospective, randomized, open label with blinded endpoint ; sUA: serum uric acid level.
In addition to the differences shown in Table 1, there may be concerns regarding bias in these trials. In the FAST trial, patients randomized to allopurinol received the dose used during the lead-in phase, which could infer a selection bias. In addition, none of these trials included a placebo arm, as this would not be considered to be ethical in the long-term treatment of gout patients. The ALL-HEART study in patients with ischemic heart disease compared high dose allopurinol with placebo , while the TROFEO (TopiROxostat and FEbuxostat in a randomized, Open-label, cross-over trial for hyperuricemia with cardiovascular disease) trial will compare febuxostat and topiroxostat in Japan . These trials may help to ascertain whether there is a class or drug effect in patients with CV disease.
The clinical question raised is how best to treat patients with CV disease who either do not tolerate, or suffer adverse events while on, allopurinol or are at high risk of developing severe to lethal adverse events with allopurinol, such as subjects with the HLA B*5801 allele . Benzbromarone or pegloticase are alternatives, but are not available in many countries. Treating flares alone would also be an alternative to consider in patients with recent onset gout or infrequent attacks .
Data from recent papers has shown increased risk of episodes of heart failure due to NSAID use in patients with prevalent heart disease , in addition to an increased number of hospital admissions due to the use of NSAIDs, which are drug and dose dependent . A recent recommendation from the European Society of Cardiology suggests a stepwise approach to the pharmacological treatment of musculoskeletal pain in patients with, or at high risk of, CV disease with the first step being to optimize treatment of the underlying disease in order to avoid the need to prescribe NSAIDs . It is important to remember when prescribing NSAIDs that most patients who survive acute kidney injury (AKI) have an indication of arthritis and over half of users were taking NSAIDs regularly both before and after an AKI event. In addition, heart failure was present in 44% and chronic kidney disease (CKD) in 36% of NSAID users .
Until new data are released from further analysis of any of the trials cited above, following the FDA  and EMA  labeling regarding prescription of febuxostat to patients with ischemic heart disease or heart failure is important. Nevertheless, using high dose NSAIDs, which have also been shown to be associated with increased risk of CV events, to treat recurrent flares, appears to not be a plausible option for some patients with CV disease where there is no other alternative to allopurinol.
About the author