Gout and Cardiovascular Disease: Crystallized Confusion

Aryeh M. Abeles; Michael H. Pillinge


Curr Opin Rheumatol. 2019;31(2):118-124. 

In This Article

The Treatment of Gout: A Test Case for the Impact of Gout on Cardiovascular Disease

If gout promotes cardiovascular risk, it would seem logical that treatment of gout would lower cardiovascular risk. Current treatment of gout includes two overarching strategies: lowering serum urate and managing gouty inflammation. Here we examine the recent literature for each of these strategies.

Urate-lowering Therapy and Cardiovascular Risk: Multiple Studies, no Consensus

Allopurinol. Recent studies conflict as to whether urate-lowering therapy (ULT) reduces the incidence of cardiovascular disease among gout patients. For instance, one case–control study from France showed that allopurinol lowered the risk of nonfatal myocardial infarction (MI) by 20% among patients with gout (although the adjusted OR of 0.8 had a confidence interval crossing 1), whereas a retrospective cohort study from Taiwan concluded that gout patients on allopurinol were 20% more likely to experience incident cardiovascular events.[13,14] Differences in study populations render direct comparison challenging, however. Most significantly, 67% of the French cohort took at least 200 mg/day allopurinol, whereas 67% in the Taiwanese study were on unusually low doses (≤100 mg). On the other hand, escalating allopurinol dose failed to reduce cardiovascular mortality in one 10-year observational, active-comparator study.[15] However, in that study, the group undergoing dose escalation differed from a fixed low-dose allopurinol control group in ways that may have confounded cardiovascular outcomes. For example, significantly more individuals in the dose-escalation group were managed by rheumatologists (18 vs. 7%), suggesting the possibility of more severe disease at study entry. Additionally, the dose-escalation group had higher baseline serum urate levels and greater use of prednisone, factors, which may have conferred cardiovascular risk and mitigated any possible effect of allopurinol. Finally, the maximum dose of allopurinol achieved for individuals in both groups was lower than the mean dose of allopurinol reported to be needed to achieve the ACR/EULAR target serum urate: nearly 80% of nonescalators received 100 mg/day or less, and 99% or less of dose escalators received 300 mg/day. Thus, both groups may have received doses of allopurinol too low to achieve a meaningful cardiovascular effect.

A population-based, case–control study from Spain reported that current use of allopurinol was associated with decreased risk of MI (adjusted OR 0.52). The effect was time-dependent and dose-dependent – patients taking allopurinol for more than 180 days had an MI risk that was even more reduced (OR 0.21). Higher doses of allopurinol were also potentially more protective; at least 300 mg daily yielded an OR of 0.30, whereas less than 300 mg yielded an OR = 0.67 (but NS, with 95% CI 0.37–1.23). Additionally, patients with a prior history of acute MI experienced significant risk reduction for nonfatal acute MI recurrence (adjusted OR = 0.16).[16] Like the Spanish study, a Taiwanese retrospective cohort study of gout patients showed a time-dependent allopurinol benefit. Although allopurinol did not decrease incident CAD during the initial period of observation, it was associated with a decreased risk of CAD if it had been taken for at least 270 days.[17]

In contrast, a Danish group found no time-dependent or dose-dependent protective effect of allopurinol, in a propensity-matched cohort study using Denmark's national database. Neither the dose of allopurinol nor the degree of serum urate lowering, affected incident cardiovascular disease (nonfatal MI, nonfatal stroke, and cardiovascular death), and achieving a target serum urate [<0.36 mmol/l (<6.0 mg/dl)] did not lead to better cardiovascular outcomes (OR = 1.01).[18] An observational study performed in the United States examined the effect of significant serum urate lowering (≥3 mg/dl) on incident cardiovascular disease in a cohort of gout patients covered by private health insurers and concurred with the Danish study: significantly lowering serum urate resulted in no decline in incident cardiovascular disease (adjusted hazard ratio of 1.01).[19] To confuse matters, however, in a different study, the Danish group matched all incident allopurinol users from a single county with nonusers. In that study, they observed that any allopurinol use slightly decreased the risk of a composite cardiovascular endpoint (nonfatal CVA, nonfatal MI, cardiovascular death; hazard ratio 0.89). Allopurinol also decreased the risk of death (hazard ratio 0.68).[20] Although most patients taking allopurinol in the community have a history of gout, neither of the Danish studies provided data regarding gout history. Thus, it is possible that the investigators were comparing a group of allopurinol-treated gout patients with a group of individuals with untreated asymptomatic hyperuricemia. As several studies have reported a higher risk of cardiovascular disease in gout patients vs. those with hyperuricemia alone, the small relative reduction in risk among the allopurinol group may have underestimated the actual protective effect.

Further examples of studies with diametrically opposed findings include a prospective case-matched cohort study from Taiwan and a propensity-score matched retrospective US study using a Medicare database. The Taiwanese study showed that gout patients taking ULT (allopurinol or benzbromarone) had lower rates of cardiovascular disease (hazard ratio 0.29) and all-cause mortality (hazard ratio 0.47) than those who did not.[21] In contrast, the United States study investigated the effect of xanthine oxidase inhibitor (XOI) initiation on incident cardiovascular disease and death and found no difference in outcomes between gout patients who began XOI and those who did not (though the study's authors acknowledge that low adherence in the XOI group may have affected the outcome).[22]

Two recently published meta-analyses have examined the effect of urate lowering on incidence of cardiovascular disease. The first included only randomized control trials of ULTs in 7757 patients with established gout, nearly all of whom had been enrolled in trials investigating XOIs (allopurinol or febuxostat). The authors concluded that urate lowering did not reduce incidence of cardiovascular events. Unfortunately, most of the included studies were of short duration, included low-risk populations, and were underpowered to detect between-group differences.[23] The other meta-analysis was larger and more selective, including only studies assessing XOI use, mainly in high-risk populations. That study suggested that whereas allopurinol, a purine-like XOI, reduced incident cardiovascular disease, the nonpurine-like XOI febuxostat did not. Paradoxically, the protective benefit of allopurinol dissipated with higher (>300 mg/day) doses. Importantly, this meta-analysis included mostly studies of hyperuricemic patients without gout (only 12 of 93 studies had majority-gout patients; most had either no patients with gout or the presence of gout was not recorded). Thus, whether this second meta-analysis provides any insights into allopurinol impact on cardiovascular risk in gout patients is unclear. Moreover, even in established gout patients, modifying hyperuricemia may not favorably impact upon what may be a gout-associated biological phenotype associated with gout that independently leads to adverse cardiovascular outcomes.[24]

In summary, although numerous retrospective studies have investigated whether allopurinol reduces incident cardiovascular disease in gout patients, we remain without clarity. Some studies show that allopurinol markedly reduces incident cardiovascular disease; other that it affords no benefit. Some studies show that higher doses of allopurinol are essential for a better outcome; others that any dose at all is good enough (or provides no benefit whatsoever). Heterogeneity in study design, and the limitations inherent to retrospective studies, contribute to the confusion.

The ALL-HEART study is an ongoing, prospective, randomized trial investigating whether allopurinol reduces cardiovascular outcomes in patients with established ischemic heart disease and asymptomatic hyperuricemia. The primary outcome is the composite end-point of nonfatal MI, nonfatal stroke or cardiovascular death. Allopurinol is being given at different doses (100– 600 mg daily) based on renal function at time of study entry. However, as one exclusion criterion is gout, the study may have limited bearing on patients in rheumatology practices, as gout may confer different cardiovascular risks than asymptomatic hyperuricemia.[25]

Other urate-lowering therapies and cardiovascular outcome. Whether urate-lowering therapies other than allopurinol may provide cardiovascular benefit remains an open question.

Probenecid: In one Medicare cohort study, patients taking probenecid were less likely to suffer myocardial infarction or stroke than those taking allopurinol (hazard ratio 0.8). Rather than implying relative safety of probenecid over allopurinol; however, this finding may reflect baseline differences between the patient groups, and the study lacked important comparative data, most notably pretreatment serum urate levels. On the basis of prescribing guidelines, patients receiving probenecid were less likely to have CKD, which could have affected cardiovascular outcomes. Nevertheless, in the subcohort of patients without chronic kidney disease, probenecid was still associated with a reduced risk of major adverse cardiovascular events (hazard ratio 0.84). As probenecid is uncommonly prescribed in practice, available data was limited; in the accessed Medicare database, probenecid was prescribed at roughly 3% of the rate of allopurinol.[26]

Febuxostat vs. allopurinol: cardiovascular safety: A recent cohort study of Medicare patients (average age 76) compared outcomes of 24 936 patients beginning on febuxostat to 74 808 beginning on allopurinol. No differences emerged between the groups regarding risk of MI, stroke, coronary revascularization, or all-cause mortality.[27] However, these findings are at odds with the results of the prospective CARES trial, comparing the effect of febuxostat vs. allopurinol on incident cardiovascular disease. Inclusion criteria included a history of major cardiovascular disease and gout. Although the study's primary outcome – new cardiovascular events – did not differ between the two treatment groups, two secondary outcomes – cardiovascular and all-cause death – favored allopurinol. However, the study had significant limitation, including the fact that differences in mortality between the two treatment arms may have been explained by effects of NSAIDs and aspirin (ASA). Patients in the febuxostat group who were co-administered NSAIDs experienced a significantly higher rate of cardiovascular mortality than those taking allopurinol and NSAIDs [41/856 (4.8%) of those co-administered febuxostat and NSAIDs died because of cardiovascular causes vs. 20/908 (2.2%) taking allopurinol and NSAIDs]. Although it is unclear why NSAIDs should have differing effects on patients taking different urate-lowering medications, NSAIDs are known to increase the risk of cardiovascular disease and death, and in a clinical trial involving high-risk cardiovascular patients, NSAIDs should probably have been excluded from the trial protocol. Another factor that appears to have impacted differences between the treatment groups is whether patients took ASA. Patients randomized to receive febuxostat but who did not take ASA had a significantly higher risk of cardiovascular mortality than those patients randomized to allopurinol not taking ASA (85/1602 [5.3%], vs. 43/1611 [2.7%]). Therefore, the mortality difference between the groups might be largely attributable to the adverse effects of NSAIDs and the salutary effects of low-dose ASA in a high-risk cardiovascular population, rather than any risk specifically conferred by febuxostat or benefit conferred by allopurinol. The authors obliquely address this in the discussion, when they write, 'The only heterogeneity in the analyses of cardiovascular mortality occurred in two subgroups – patients with concomitant administration of ASA or NSAIDs. These drugs may be associated with more frequent gout flares, which, in turn, could lead to increases in cardiovascular events.' At the very least, these data suggest that appropriate cardiovascular risk management in high-risk patients (use of ASA and avoidance of NSAIDs) may mitigate any possible effects favoring allopurinol over febuxostat. Another difference between the two groups is that febuxostat was more effective at lowering serum urate. Given the possibility that overly aggressive urate lowering increases the likelihood of cardiovascular events, this potential confounder could also contribute to the difference in cardiovascular outcomes between the two groups (for instance, at month 36 of the study, 33% of the patients on allopurinol had UA <5, vs. 46.1% of patients taking febuxostat) – and could potentially be remedied by slower titration of febuxostat. Importantly, the dropout rate for the study was nearly 60%; when a post hoc analysis included data on as many dropped-out patients as could be identified, the adverse cardiovascular signal for febuxostat was no longer significant. Additionally, most of the cardiovascular events (and deaths) occurred after drug discontinuation, making it difficult to impute causation. Finally, as the CARES trial lacked an untreated control group, even granting the difference in outcomes between the two treatments, it is impossible to determine with certainty whether a greater rate of events in the febuxostat group represents an actual increase in events, or possibly a less dramatic risk reduction when compared with allopurinol.[28]

Clouding the picture further, a recent retrospective cohort study of commercially insured gout patients with preexisting cardiovascular disease (or heart failure) and stage 3 or 4 CKD showed that febuxostat led to favorable outcomes relative to allopurinol. Patients initiating febuxostat were less likely to experience major cardiovascular events than those initiating allopurinol (hazard ratio 0.52). Shortcomings to this study included lack of cohort matching, lack of propensity-score matching, short follow-up (9 months), absence of mortality data, and absence of data on serum urate levels of the patients. Additionally, much of the difference between the outcomes of the two medications was driven by differences in rates of peripheral vascular events.[29]

Possibly the results of another ongoing study, the FAST trial – enrolling in Denmark and the UK – will prove more definitive in its findings. This randomized controlled trial of patients with 'symptomatic hyperuricemia' (patients with acute gouty attacks or those with at least one confirmed tophus in the absence of history of acute gouty arthritis) is testing the relative cardiovascular safety of allopurinol and febuxostat. As with the CARES trial, the study's protocol mandates that patients' medications be titrated to a specific serum urate target. If needed, patients receiving febuxostat will receive higher doses than in the CARES trial, as febuxostat is approved at 80 and 120 mg doses in Europe. This cohort will also differ from CARES in terms of cardiovascular risk. Where the CARES patients had established cardiovascular disease, the FAST patients only had to have at least one cardiovascular risk factor.[30]

Nonurate-lowering Therapy Medications and Cardiovascular Risk

Although the bulk of attention has been paid to ULTs' potential impact on cardiovascular risk, there are other classes of medications that might offer favorable cardiovascular risk reduction, including colchicine and HMG-CoA reductase inhibitors (i.e. statins).

Gout, antiinflammatories and cardiovascular risk. The argument that inflammation may promote cardiovascular disease raises the possibility that the acute and chronic inflammation seen in gout could confer cardiovascular risk, and that anti-inflammatory agents used to treat gout could potentially reverse that risk. Of course, gout mainstay drugs, such as NSAIDS and glucocorticoids are well established to raise cardiovascular risk, but these effects are presumed to act through mechanisms independent of their anti-inflammatory effects.

Colchicine: A handful of studies have assessed whether the antiinflammatory gout drug colchicine could lower cardiovascular risk.[31,32,33] One single-center retrospective cross-sectional study showed that gout patients who had received at least one prescription of colchicine had a lower prevalence of myocardial infarctions (1.2 vs. 2.6%, hazard ratio 0.46), but the duration of colchicine was unknown, and the study was not adjusted for confounders.[31] Nevertheless, a subsequent retrospective Medicare database study showed similar findings. Among gout patients, individuals taking colchicine had markedly lower rates of cardiovascular events (hazard ratio 0.51) compared with nonusers, as well as a 73% reduction in all-cause mortality (hazard ratio 0.27).[32] Subsequent studies, reported so far only in abstract form, suggest that the ability of colchicine to lower both hsCRP (an MI risk biomarker) and overall MI risk may be observed only with its active use.[33] Finally, a large pragmatic trial studied the effect of daily colchicine among nongout patients for secondary prevention of cardiac outcomes and showed a significant benefit on a scale similar to that of the two colchicine studies reported in gout patients.[34]

Anti-IL-1β therapy: It is now appreciated that IL-1β is a central player in gouty inflammation, and IL-β has also been appreciated as an inflammatory cytokine associated with cardiovascular disease. Thus, it was interesting to observe the recent CANTOS trial, in which the antigout, anti-IL-1β biologic canakinumab was shown to reduce major cardiovascular events in a large cohort of patients.[35] This trial provides the first definitive proof-of-principle that an anti-inflammatory strategy can reduce cardiovascular risk. As the benefit was modest, and the expense of canakinumab is high, it is unlikely to advance to common use. Perhaps not surprisingly, a recently published secondary analysis revealed that the use of canakinumab in CANTOS also reduced the rate of incident gout (e.g. first gouty attack). Whether other anti-IL1β therapies (e.g. anakinra) would have similar effects is not known. Of note, some studies suggest that colchicine itself may reduce IL-1β levels by blocking its cellular production.[36]

Statins and Cardiovascular Risk in Gout Patients

Many gout patients require statins to treat hyperlipidemia, raising the question of whether statins can reduce cardiovascular risk in gout patients. Indeed, the understanding that statins may exert a protective effect not only by reducing serum lipids but also through an anti-inflammatory effect, suggests the possibility of an additional cardioprotective role in patients with gout.[37] In one incident user cohort study, patients with gout who initiated statins had significantly reduced all-cause mortality compared with a propensity score-matched cohort of patients with gout not taking statins (hazard ratio 0.84). The benefit was even more pronounced for those patients without known cardiovascular disease (hazard ratio 0.65).[38] In a prior meta-analysis investigating clinical trials of statins for patients at low risk of vascular disease, statin use conferred a more modest reduction in mortality, with a relative risk of 0.91.[39] More data is needed, but these initial data suggest that statins might be effective at reducing mortality in the gout population.