Gota PDF

The n e w e ng l a n d j o u r na l of m e dic i n e
Original Article
Cardiovascular Safety of Febuxostat

or Allopurinol in Patients with Gout
William B. White, M.D., Kenneth G. Saag, M.D., Michael A. Becker, M.D.,
Jeffrey S. Borer, M.D., Philip B. Gorelick, M.D., Andrew Whelton, M.D.,
Barbara Hunt, M.S., Majin Castillo, M.D., and Lhanoo Gunawardhana, M.D., Ph.D.,
for the CARES Investigators*
A BS T R AC T
BACKGROUND
Cardiovascular risk is increased in patients with gout. We compared cardiovascular From the University of Connecticut
outcomes associated with febuxostat, a nonpurine xanthine oxidase inhibitor, with School of Medicine, Farmington (W.B.W.);
the University of Alabama, Birmingham
those associated with allopurinol, a purine base analogue xanthine oxidase in- (K.G.S.); University of Chicago Medicine,
hibitor, in patients with gout and cardiovascular disease. Chicago (M.A.B.), and Takeda Develop-
ment Center Americas, Deerfield (B.H.,
METHODS M.C., L.G.) — both in Illinois; the State
University of New York, Downstate Medi-
We conducted a multicenter, double-blind, noninferiority trial involving patients cal Center, Brooklyn (J.S.B.); Michigan
with gout and cardiovascular disease; patients were randomly assigned to receive State University College of Human Medi-
febuxostat or allopurinol and were stratified according to kidney function. The trial cine, Grand Rapids (P.B.G.); and Johns
Hopkins University School of Medicine,
had a prespecified noninferiority margin of 1.3 for the hazard ratio for the primary Baltimore (A.W.). Address reprint requests
end point (a composite of cardiovascular death, nonfatal myocardial infarction, to Dr. White at the Calhoun Cardiology
nonfatal stroke, or unstable angina with urgent revascularization). Center, University of Connecticut School
of Medicine, 263 Farmington Ave., Farming-
RESULTS ton, CT 06032, or at w white@uchc.edu.
In total, 6190 patients underwent randomization, received febuxostat or allopurinol, *A complete list of the CARES investiga-
and were followed for a median of 32 months (maximum, 85 months). The trial tors is provided in the Supplementary
Appendix, available at NEJM.org.
regimen was discontinued in 56.6% of patients, and 45.0% discontinued follow-up.
In the modified intention-to-treat analysis, a primary end-point event occurred in This article was published on March 12,
2018, at NEJM.org.
335 patients (10.8%) in the febuxostat group and in 321 patients (10.4%) in the
allopurinol group (hazard ratio, 1.03; upper limit of the one-sided 98.5% confi- DOI: 10.1056/NEJMoa1710895
Copyright © 2018 Massachusetts Medical Society.
dence interval [CI], 1.23; P = 0.002 for noninferiority). All-cause and cardiovascular
mortality were higher in the febuxostat group than in the allopurinol group (haz-
ard ratio for death from any cause, 1.22 [95% CI, 1.01 to 1.47]; hazard ratio for
cardiovascular death, 1.34 [95% CI, 1.03 to 1.73]). The results with regard to the
primary end point and all-cause and cardiovascular mortality in the analysis of
events that occurred while patients were being treated were similar to the results
in the modified intention-to-treat analysis.
CONCLUSIONS
In patients with gout and major cardiovascular coexisting conditions, febuxostat was
noninferior to allopurinol with respect to rates of adverse cardiovascular events. All-
cause mortality and cardiovascular mortality were higher with febuxostat than with
allopurinol. (Funded by Takeda Development Center Americas; CARES ClinicalTrials
.gov number, NCT01101035.)
n engl j med nejm.org 1
The New England Journal of Medicine
Downloaded from nejm.org on March 12, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
G
out is a chronic illness character- drafted the manuscript, had full access to the fi-
ized by hyperuricemia, arthropathy, to- nal trial data, and vouch for the accuracy and
phus development, and urolithiasis and completeness of the data and the analyses, as well
is associated with an increased risk of cardiovas- as for the fidelity of the trial to the protocol,
cular and chronic kidney disease.1 The risk of car- which is available with the full text of this article
diovascular events, including death, is substan- at NEJM.org. The appropriate national and insti-
tially higher in people with gout than in those tutional regulatory authorities and ethics com-
without gout.2,3 When the Food and Drug Admin- mittees approved the trial design.
istration (FDA) released a guidance document
outlining specific requirements for the cardio- Patients
vascular safety assessment of antidiabetic thera- Patients were eligible for enrollment in the trial
pies,4 investigators in other therapeutic areas, in- if they had a diagnosis of gout fulfilling the
cluding those studying gout therapies, began to American Rheumatism Association criteria9 and
explore cardiovascular safety with similarly de- a history of major cardiovascular disease before
signed trials. randomization (detailed inclusion and exclusion
Febuxostat, a nonpurine inhibitor of xanthine criteria are provided in Table S1 in the Supple-
oxidase that is used for the management of hyper- mentary Appendix, available at NEJM.org). Addi-
uricemia in patients with gout, inhibits both the tional criteria for inclusion were a serum urate
oxidized and reduced forms of xanthine oxidase level of at least 7.0 mg per deciliter (420 μmol per
and decreases the formation of uric acid.5 Febuxo- liter), or of at least 6.0 mg per deciliter (360 μmol
stat provides highly selective and potent inhibi- per liter) with inadequately controlled gout, after
tion of xanthine oxidase and greater hypouricemic a 1-to-3-week washout period from previous gout
activity than do commonly used doses of allopu- therapies. Patients were regarded as having a
rinol.6 During its development, febuxostat was history of major cardiovascular disease if they
compared with placebo and allopurinol in clini- had had a myocardial infarction, hospitalization
cal trials involving more than 5000 patients with for unstable angina, stroke, hospitalization for
gout5-7; these trials suggested a modestly higher transient ischemic attack, peripheral vascular dis-
rate of cardiovascular events with febuxostat. The ease, or diabetes mellitus with evidence of mi-
Cardiovascular Safety of Febuxostat and Allopu- crovascular or macrovascular disease, as defined
rinol in Patients with Gout and Cardiovascular previously.8 All participants provided written in-
Morbidities (CARES) trial was therefore conduct- formed consent.
ed as an FDA requirement to determine whether
febuxostat was noninferior to allopurinol with Treatment and Procedures
regard to major cardiovascular events in patients Patients were randomly assigned to receive fe-
with gout and cardiovascular disease. buxostat or allopurinol administered in a double-
blind fashion once daily. Randomization was
stratified according to the estimated creatinine
Me thods
clearance at baseline (≥60 ml per minute vs. ≥30
Trial Design but <60 ml per minute).
We conducted a multicenter, randomized, double- Doses of allopurinol were modified according
blind noninferiority trial; details of the design of to kidney function. Patients with an estimated
the trial have been published previously.8 The creatinine clearance of at least 60 ml per minute
funder (Takeda Pharmaceuticals) participated in initially received allopurinol at a dose of 300 mg
the trial design, conduct, and monitoring and in once daily, which was increased in 100-mg in-
data collection, storage, and analyses. An indepen- crements monthly until the patient either had a
dent data and safety monitoring committee moni- serum urate level of less than 6.0 mg per decili-
tored the trial and had access to the unblinded ter or was receiving an allopurinol dose of 600 mg
data. Statistical analyses were performed for the once daily. Patients who had an estimated cre-
data and safety monitoring committee by an in- atinine clearance of at least 30 but less than 60 ml
dependent statistical group (WebbWrites). per minute initially received 200 mg of allopuri-
The academic authors of the present article nol; the dose was increased in 100-mg increments
2 n engl j med nejm.org

Cardiovascular Safety of Febuxostat or Allopurinol in Gout
until the patient either had a serum urate level itored but would not return for study visits, tele-
of less than 6.0 mg per deciliter or was receiving phone contacts were completed, but this was not
an allopurinol dose of 400 mg once daily. preferred or recommended to the sites.
Febuxostat doses were not modified accord-
ing to kidney function. Patients who were ran- End Points
domly assigned to receive febuxostat initially The primary composite end point was the first
received 40 mg once daily and continued to re- occurrence of cardiovascular death, nonfatal myo-
ceive this dose if the serum urate level was less cardial infarction, nonfatal stroke, or urgent re-
than 6.0 mg per deciliter after 2 weeks of therapy. vascularization for unstable angina (definitions
If the serum urate level was higher than 6.0 mg are provided in Table S2 in the Supplementary Ap-
per deciliter at the week 2 visit, the dose of fe- pendix).10 The secondary safety end points included
buxostat was increased to 80 mg once daily for a composite of cardiovascular death, nonfatal
the remainder of the trial. myocardial infarction, or nonfatal stroke as well
The patients’ serum urate levels were revealed as the individual components of the primary end
to the site investigators only during a 10-week point. The consistency of effects on the primary
dose-adjustment period to facilitate dose increas- end point was explored in a variety of subgroups
es that were based on urate response. During (both prespecified and post hoc). Additional safety
that period, the administration of double-blind, end points included death from any cause, urgent
double-dummy trial medications was guided by cerebrovascular revascularization, transient isch-
an interactive voice-response system, a procedure emic attack, hospitalization for heart failure,
that prevented unblinding for patients whose dose arrhythmias not associated with ischemia, and
was not adjusted. After dose adjustments were venous thromboembolic events. An independent
completed, urate levels were concealed from in- central end-points committee, the members of
vestigators and the sponsor, and the interactive which were unaware of the treatment assign-
voice-response system was used to manage treat- ments, adjudicated all suspected end-point events.
ment throughout the trial.
At the screening visit, all urate-lowering Statistical Analysis
therapy was discontinued and, unless the patient Cox proportional-hazards models, stratified ac-
had a history of unacceptable side effects from cording to baseline kidney function, were used
colchicine, treatment with colchicine at a dose of to analyze the time to first occurrence of primary
0.6 mg daily was started for gout flare prophy- and secondary end-point events for all patients
laxis. All the patients received prophylaxis for the who underwent randomization and received treat-
first 6 months of randomly assigned treatment. If ment. A determination of noninferiority of febux-
colchicine treatment resulted in unacceptable side ostat to allopurinol required that the upper bound
effects and the estimated creatinine clearance of the one-sided confidence interval of the hazard
was at least 50 ml per minute, patients received ratio for the primary end point be less than 1.3.
naproxen (250 mg twice daily) with lansoprazole The number and percentage of patients with a
(15 mg once daily). If patients could receive nei- primary end-point event or cardiovascular death
ther colchicine nor naproxen, other nonsteroidal were tabulated for various subgroups. The rela-
antiinflammatory drugs (NSAIDs) or prednisone tive risk (febuxostat vs. allopurinol) was calculated
could be provided as prophylaxis, or the investi- within each subgroup, with homogeneity among
gators could choose to manage gout flares as subgroup levels assessed with the use of the Co-
they occurred. chran–Mantel–Haenszel test. Sensitivity analyses
Outpatient visits were scheduled at screening were performed by excluding events that occurred
and randomization and at 2, 4, 6, 8, 10, 12, and after treatment discontinuation and events that
24 weeks after randomization and every 6 months occurred more than 30 days after treatment dis-
during subsequent years of the trial. Patients with continuation.
reduced kidney function or who were older than The trial was designed to accrue 624 primary
65 years of age at randomization also had visits events for assessing the noninferiority of febuxo-
at 9 months and 15 months to monitor serum stat to allopurinol with regard to cardiovascular
chemical profiles. If patients agreed to be mon- risk, under the assumption of a true hazard ratio
of 1.0 and 90% power. Interim analyses were of the one-sided confidence interval for the haz-
conducted when approximately 25%, 50%, and 75% ard ratio (febuxostat vs. allopurinol) would be cal-
of the events had occurred. For each group-sequen- culated with the use of the critical value from the
tial analysis, it was specified that the upper bound Lan–DeMets–O’Brien–Fleming alpha-spending
Table 1. Baseline Characteristics of the Patients.*
Febuxostat Allopurinol
Characteristic (N = 3098) (N = 3092)
Median age (interquartile range) — yr 64.0 (58–71) 65.0 (58–71)
Age ≥65 yr — no. (%) 1514 (48.9) 1586 (51.3)
Male sex — no. (%) 2604 (84.1) 2592 (83.8)
Duration of gout — yr 11.8±11.4 11.9±11.2
Baseline serum urate level — mg/dl 8.7±1.7 8.7±1.7
Presence of tophi — no. (%) 668 (21.6) 650 (21.0)
Median body weight (interquartile range) — kg 97.7 (84–113) 97.3 (84–113)
Body-mass index† 33.6±7.0 33.4±6.9
Race or ethnic group — no. (%)‡
White 2160 (69.7) 2140 (69.2)
Black 552 (17.8) 593 (19.2)
Asian 92 (3.0) 96 (3.1)
American Indian or Alaska Native 262 (8.5) 234 (7.6)
Native Hawaiian or other Pacific Islander 13 (0.4) 14 (0.5)
Other 19 (0.6) 15 (0.5)
Cardiovascular risk factors and history — no. (%)
Diabetes mellitus with small-vessel disease 1193 (38.5) 1213 (39.2)
Hypertension 2864 (92.4) 2851 (92.2)
Hyperlipidemia 2678 (86.4) 2702 (87.4)
Myocardial infarction 1197 (38.6) 1231 (39.8)
Hospitalization for unstable angina 855 (27.6) 869 (28.1)
Coronary revascularization 1129 (36.4) 1182 (38.2)
Cerebral revascularization 69 (2.2) 54 (1.7)
Congestive heart failure 622 (20.1) 631 (20.4)
Stroke 460 (14.8) 410 (13.3)
Peripheral vascular disease 412 (13.3) 375 (12.1)
Median estimated creatinine clearance — ml/min§
Stage 1 or 2 chronic kidney disease 75.0 73.0
Stage 3 chronic kidney disease 46.0 46.0
Stage of chronic kidney disease — no./total no. (%)
Stage 1 or 2 1456/3092 (47.1) 1459/3090 (47.2)
Stage 3 1636/3092 (52.9) 1631/3090 (52.8)
* Plus–minus values are means ±SD. There were no significant differences between the two groups with regard to any
baseline characteristic. To convert the values for urate to micromoles per liter, multiply by 59.48.
† The body-mass index is the weight in kilograms divided by the square of the height in meters.
‡ Race or ethnic group was reported by the patient.
§ Estimated creatinine clearance was calculated with the use of the Cockcroft–Gault formula and was corrected for ideal
body weight. A value of 60 ml per minute or more indicated stage 1 or 2 chronic kidney disease, and a value of at least
30 but less than 60 ml per minute indicated stage 3 chronic kidney disease.

function, which preserves an overall one-sided al-

Eight patients never received trial medication,
pha of 0.025.11 No other adjustments for multi-which left 6190 patients in a modified intention-
plicity were made. Each of these analyses was to-treat analysis (Fig. S1 in the Supplementary
conducted by an independent statistician and re-
Appendix). The two treatment groups were well
viewed by the data and safety monitoring board.balanced with regard to all baseline characteris-
It was planned that, if the upper bound of the
tics (Table 1, and Table S3 in the Supplementary
one-sided confidence interval of the hazard ratio
Appendix). In the febuxostat group, 61.0% of the
was less than 1.3 at any interim analysis, the trial
patients received 40 mg and 39.0% received 80 mg
would be stopped, since noninferiority of febuxo-
daily as the final adjusted dose. In the allopurinol
stat to allopurinol with regard to cardiovascular
group, on the basis of the protocol-directed crite-
risk would be declared. In April 2016, at the time
ria for estimated creatinine clearance, 21.8% of the
of the 75% interim analysis, the estimated haz-patients received 200 mg, 44.6% received 300 mg,
25.2% received 400 mg, 4.3% received 500 mg, and
ard ratio and adjusted upper bound of the confi-
4.1% received 600 mg.
dence interval for the hazard ratio were 0.99 and
1.23, respectively. However, because of a discrep- Overall, 56.6% of patients discontinued trial
ancy between the hazard ratio for death from treatment prematurely; the rates of premature dis-
continuation were similar in the febuxostat and
any cause in the intention-to-treat analysis and
in the analysis of events that occurred during allopurinol groups (57.3% and 55.9%, respective-
treatment, the data and safety monitoring boardly). The percentage of patients who did not com-
recommended continuing the trial until the pre-plete all trial visits was 45.0% overall — 45.0%
specified 624 primary events had occurred. in the febuxostat group and 44.9% in the allopu-
rinol group. The median duration of exposure to
febuxostat was 728 days, and the median dura-
R e sult s
tion of exposure to allopurinol was 719 days. The
Patients median duration of follow-up was 968 days in the
We enrolled 6198 patients from 320 North Amer- febuxostat group and 942 days in the allopurinol
ican sites from April 2010 through May 2017. group.
Table 2. Major Safety End Points (Modified Intention-to-Treat Analysis).*
Febuxostat Allopurinol Hazard Ratio

End Point (N = 3098) (N = 3092) (95% CI) P Value†
no. of patients (%)

Primary end point: composite of cardiovascular 335 (10.8) 321 (10.4) 1.03 (0.87–1.23)‡ 0.66
death, nonfatal myocardial infarction, non- (0.002)
fatal stroke, or urgent revascularization
due to unstable angina
Secondary end points
Cardiovascular death 134 (4.3) 100 (3.2) 1.34 (1.03–1.73) 0.03
Nonfatal myocardial infarction 111 (3.6) 118 (3.8) 0.93 (0.72–1.21) 0.61
Nonfatal stroke 71 (2.3) 70 (2.3) 1.01 (0.73–1.41) 0.94
Urgent revascularization for unstable angina 49 (1.6) 56 (1.8) 0.86 (0.59–1.26) 0.44
Composite of cardiovascular death, nonfatal 296 (9.6) 271 (8.8) 1.09 (0.92–1.28) 0.33
myocardial infarction, or nonfatal stroke
Death from any cause 243 (7.8) 199 (6.4) 1.22 (1.01–1.47) 0.04
* The modified intention-to-treat analysis included all patients who underwent randomization with the exception of the
8 patients who never received febuxostat or allopurinol.
† The P value in parentheses is for test of the null hypothesis that the hazard ratio is at least 1.3 versus the one-sided
alternative (noninferiority). All other P values are values for the test of superiority of febuxostat to allopurinol and were
calculated with the use of a Cox regression analysis.
‡ The 97% confidence interval is provided here.
A Primary End Point

100 25
Febuxostat
90
20
80
15
Cumulative Rate (%)

70 Allopurinol
60 10
50 5
40
0
30 0 6 12 18 24 30 36 42 48 54 60 66 72
20
10
0
0 6 12 18 24 30 36 42 48 54 60 66 72
Month
No. at Risk
Febuxostat 3098 2784 2493 2111 1854 1589 1369 1165 955 778 573 441 264
Allopurinol 3092 2764 2465 2080 1815 1560 1361 1132 933 767 589 437 258
B Cardiovascular Mortality
100 25
90
20
80
15
Cumulative Rate (%)
70
Febuxostat
60 10
50 5
40 Allopurinol
0
30 0 6 12 18 24 30 36 42 48 54 60 66 72
20
10
0
0 6 12 18 24 30 36 42 48 54 60 66 72
Month
No. at Risk
Febuxostat 3098 2823 2550 2174 1922 1659 1440 1243 1033 838 627 482 288
Allopurinol 3092 2807 2530 2152 1898 1637 1433 1204 1008 838 646 489 287
C All-Cause Mortality
100 25
90
20
80 Febuxostat
15
Cumulative Rate (%)
70
60 10
Allopurinol
50 5
40
0
30 0 6 12 18 24 30 36 42 48 54 60 66 72
20
10
0
0 6 12 18 24 30 36 42 48 54 60 66 72
Month
No. at Risk
Febuxostat 3098 2828 2552 2179 1928 1666 1447 1251 1038 840 631 487 289
Allopurinol 3092 2812 2540 2161 1906 1648 1444 1215 1015 842 650 489 288

Figure 1 (facing page). Cumulative Kaplan–Meier

for noninferiority) (Table 2 and Fig. 1). In the
Estimates of the Time to the First Occurrence of analysis of the nonfatal secondary end points, the
an Adjudicated End-Point Event. hazard ratios were consistent with the overall
Panel A shows the time until the first occurrence of a result. However, the risk of death from any cause
primary end-point event — cardiovascular death, myo- and the risk of cardiovascular death were higher
cardial infarction, stroke, or urgent revascularization in the febuxostat group than in the allopurinol
due to unstable angina — in the febuxostat group and
the allopurinol group. A primary end-point event oc-
group (Table 2). Among the causes of cardiovas-
curred in 10.8% of patients in the febuxostat group cular death, sudden cardiac death was the most
and 10.4% of patients in the allopurinol group after a prevalent classification, occurring in 83 patients
median exposure of 32 months (hazard ratio, 1.03; up- (2.7%) in the febuxostat group and 56 patients
per bound of the one-sided 98.5% CI, 1.23). Panel B (1.8%) in the allopurinol group (Table S7 in the
shows the time until death from a cardiovascular
cause (hazard ratio, 1.34; 95% CI, 1.03 to 1.73), and
Supplementary Appendix). Rates of hospitaliza-
Panel C the time until death from any cause (hazard tion for heart failure, hospital admissions for ar-
ratio, 1.22; 95% CI, 1.01 to 1.47). The insets show the rhythmias not associated with ischemia, venous
same data on an enlarged y axis. thromboembolic events, and hospitalization for
transient ischemic attacks were similar in the two
groups (Table S8 in the Supplementary Appendix).
Biochemical Effects In an analysis according to subgroup, the re-
The proportion of patients with a serum urate sults with regard to the primary end point showed
level of less than 6.0 mg per deciliter was higher no heterogeneity associated with any of the base-
in the febuxostat group than in the allopurinol line factors (Fig. 2). For cardiovascular mortality,
group at week 2; thereafter, higher proportions there was an interaction for NSAID use and the
of patients in the febuxostat group had mainte- absence of use of low-dose aspirin (unadjusted
nance of serum urate levels at less than 6.0 mg P<0.05 for both comparisons) (Fig. S2 in the Sup-
per deciliter at most time points, although the plementary Appendix).
differences between the groups were not large
(Table S4 in the Supplementary Appendix). In Analyses of Events That Occurred during
addition, a larger proportion of patients in the Treatment
febuxostat group than in the allopurinol group In the prespecified analysis of events that occurred
had serum urate levels of less than 5.0 mg per during receipt of the trial drug or within 30 days
deciliter (300 μmol per liter) for the entire trial. after discontinuation of treatment, a primary
Overall, the rates of gout flares were similar in end-point event occurred in 7.8% of patients in
the two treatment groups (0.68 and 0.63 flares the febuxostat group and 7.7% of patients in the
per patient-year in the febuxostat group and allo- allopurinol group (hazard ratio, 1.00; upper bound
purinol group, respectively). There were no signifi- of the one-sided 98.5% CI, 1.22). In this analy-
cant differences in serum levels of electrolytes, sis, the rate of cardiovascular death was higher
glucose, or lipids or in blood pressure between in the febuxostat group than in the allopurinol
the groups during the trial (Table S5 in the Sup- group (hazard ratio, 1.49; 95% CI, 1.01 to 2.22)
plementary Appendix), nor were there differenc- (Table 3). In a post hoc analysis of events that
es in cardiovascular medication use (Table S6 in occurred during treatment, a primary end-point
the Supplementary Appendix). event was also found to occur at similar rates in
the febuxostat group and the allopurinol group
Safety (6.2% and 6.4% of patients, respectively; hazard
After the accrual of 624 events that initiated trial ratio, 0.94; upper bound of the one-sided 98.5%
closeout and before database lock, 32 additional CI, 1.17) (Table S9 in the Supplementary Appen-
primary end-point events occurred. In the com- dix). The risk of death from any cause and the
plete analysis, a primary end-point event occurred risk of cardiovascular death were higher in the
at similar rates in the febuxostat group and the febuxostat group than in the allopurinol group.
allopurinol group (10.8% and 10.4% of patients,
respectively, at a median period of 32 months; Other Analyses
hazard ratio, 1.03; upper bound of the one-sided The baseline characteristics were balanced among
98.5% confidence interval [CI], 1.23; P = 0.002 the patients who did not complete all the trial
Subgroup Febuxostat Allopurinol Relative Risk (95% CI) P Value for

no. of patients with primary end point/total no. (%) Interaction
Baseline renal function 0.40
Moderately reduced 207/1636 (12.7) 212/1631 (13.0) 0.97 (0.81–1.16)
Mildly reduced 110/1217 (9.0) 92/1231 (7.5) 1.21 (0.93–1.58)
Normal 17/239 (7.1) 17/228 (7.5) 0.95 (0.50–1.82)
Age 0.38
<65 yr 133/1584 (8.4) 130/1506 (8.6) 0.97 (0.77–1.23)
≥65 yr 202/1514 (13.3) 191/1586 (12.0) 1.11 (0.92–1.33)
Sex 0.61
Female 42/494 (8.5) 45/500 (9.0) 0.94 (0.63–1.41)
Male 293/2604 (11.3) 276/2592 (10.6) 1.06 (0.90–1.23)
BMI 0.89
<30 110/1045 (10.5) 106/1063 (10.0) 1.06 (0.82–1.36)
≥30 225/2053 (11.0) 215/2024 (10.6) 1.03 (0.86–1.23)
NSAID use 0.10
Yes 112/856 (13.1) 95/908 (10.5) 1.25 (0.97–1.62)
No 223/2242 (9.9) 226/2184 (10.3) 0.96 (0.81–1.15)
Low-dose aspirin use 0.09
Yes 163/1496 (10.9) 175/1481 (11.8) 0.92 (0.75–1.13)
No 172/1602 (10.7) 146/1611 (9.1) 1.18 (0.96–1.46)
Smoking history 0.91
Current smoker 38/390 (9.7) 38/415 (9.2) 1.06 (0.69–1.63)
Nonsmoker or former smoker 297/2708 (11.0) 283/2677 (10.6) 1.04 (0.89–1.21)
Baseline serum urate (mg/dl) 0.78
<9.0 162/1778 (9.1) 166/1815 (9.1) 1.00 (0.81–1.22)
9.0 to <10.0 83/666 (12.5) 71/646 (11.0) 1.13 (0.84–1.53)
≥10.0 90/654 (13.8) 84/631 (13.3) 1.03 (0.78–1.36)
History of diabetes 0.72
Yes 193/1710 (11.3) 180/1699 (10.6) 1.07 (0.88–1.29)
No 142/1388 (10.2) 141/1393 (10.1) 1.01 (0.81–1.26)
History of hypertension 0.73
Yes 318/2864 (11.1) 306/2851 (10.7) 1.03 (0.89–1.20)
No 17/234 (7.3) 15/241 (6.2) 1.17 (0.60–2.28)
History of nonfatal myocardial infarction 0.37
Yes 185/1197 (15.5) 171/1231 (13.9) 1.11 (0.92–1.35)
No 150/1901 (7.9) 150/1861 (8.1) 0.98 (0.79–1.22)
History of nonfatal stroke 0.15
Yes 63/460 (13.7) 67/410 (16.3) 0.84 (0.61–1.15)
No 272/2638 (10.3) 254/2682 (9.5) 1.09 (0.93–1.28)
Race 0.65
White 268/2160 (12.4) 260/2140 (12.1) 1.02 (0.87–1.20)
Nonwhite 67/938 (7.1) 61/952 (6.4) 1.11 (0.80–1.56)
Years since gout diagnosis 0.46
<5 130/1150 (11.3) 118/1091 (10.8) 1.05 (0.83–1.32)
5–10 48/580 (8.3) 60/615 (9.8) 0.85 (0.59–1.22)
>10 156/1367 (11.4) 143/1386 (10.3) 1.11 (0.89–1.37)
History of cardiac revascularization 0.16
Yes 187/1129 (16.6) 169/1182 (14.3) 1.16 (0.96–1.40)
No 148/1969 (7.5) 152/1910 (8.0) 0.94 (0.76–1.17)
Initial gout flare prophylaxis 0.91
Colchicine 295/2604 (11.3) 283/2591 (10.9) 1.04 (0.89–1.21)
Noncolchicine 40/494 (8.1) 38/501 (7.6) 1.07 (0.70–1.63)
Colchicine use during trial 0.15
Yes 104/699 (14.9) 84/694 (12.1) 1.23 (0.94–1.61)
No 231/2399 (9.6) 237/2398 (9.9) 0.97 (0.82–1.16)
History of hyperlipidemia 0.33
Yes 297/2678 (11.1) 294/2702 (10.9) 1.02 (0.88–1.19)
No 38/420 (9.0) 27/390 (6.9) 1.31 (0.81–2.10)
At least one dose adjustment 0.97
Yes 153/1207 (12.7) 176/1485 (11.9) 1.07 (0.87–1.31)
No 182/1891 (9.6) 145/1607(9.0) 1.07 (0.87–1.31)
Insulin use during trial 0.91
Yes 116/620 (18.7) 111/607 (18.3) 1.02 (0.81–1.29)
No 219/2478 (8.8) 210/2485 (8.5) 1.05 (0.87–1.25)
History of congestive heart failure 0.86
Yes 115/622 (18.5) 114/631 (18.1) 1.02 (0.81–1.29)
No 220/2476 (8.9) 207/2461 (8.4) 1.06 (0.88–1.27)
0.1 1.0 10.0
Febuxostat Better Allopurinol Better

Figure 2 (facing page). Risk Ratios for the Primary

tion in our trial included patients who were at
End Point According to Subgroup. considerably higher cardiovascular risk than those
All subgroup analyses were prespecified with the ex- included in other assessments of the cardiovascu-
ception of the analyses of race, years since gout diag- lar safety of various gout therapies,15,16 with event
nosis, history of cardiac revascularization, initial gout rates during our trial of more than 10%. The
flare prophylaxis, colchicine use during the trial, histo- safety outcomes in this trial were prespecified
ry of hyperlipidemia, dose adjustment during the trial,
insulin use during the trial, and history of congestive
and adjudicated by members of a cardiovascular
heart failure, which were post hoc. end-point committee who were unaware of the
treatment assignments; therefore, our safety out-
comes may be more reliable than data based on
visits and those who completed all the visits conventional adverse-event reporting.
(Table S10 in the Supplementary Appendix). The Unexpectedly, all-cause mortality was higher
proportions of patients who did not complete all in the febuxostat group than in the allopurinol
the trial visits were larger in the United States group, because of an excess of cardiovascular
than in Canada or Mexico (Table S11 in the Sup- deaths. Findings were similar in the modified
plementary Appendix). There were 199 additional intention-to-treat analysis and in the prespeci-
patients, identified by a search company (Omni- fied analysis that included events that occurred
Trace), who died (Table S12 in the Supplemen- during treatment and within 30 days after treat-
tary Appendix). The rates of death from any cause ment discontinuation. The mechanism underly-
during treatment (incorporating the additional ing this risk of death is unclear. Preclinical car-
deaths) were consistent with those in the analy- diovascular studies of febuxostat have shown no
ses of events that occurred during treatment de- toxic effects related to cardiac rhythm, function,
scribed above. The rates of death from any cause or metabolism.17-21 In addition, the rates of adju-
after discontinuation of trial medication were dicated nonfatal events, including myocardial in-
similar in the two treatment groups. farction, coronary revascularization, arrhythmias,
and hospitalization for heart failure, were similar
in the febuxostat group and the allopurinol group.
Discussion
The only heterogeneity in the analyses of car-
In the CARES trial, treatment with febuxostat diovascular mortality occurred in two subgroups
resulted in overall rates of major cardiovascular — patients with concomitant administration of
events that were similar to those associated with aspirin or NSAIDs. These drugs may be associated
allopurinol treatment among patients with gout with more frequent gout flares, which, in turn,
who had coexisting cardiovascular disease. How- could lead to increases in cardiovascular events.22
ever, cardiovascular death and deaths from any However, we did not find a large difference in the
cause were more frequent in the febuxostat group reduction in urate level between the treatment
than in the allopurinol group. groups, nor did we detect differences in flare
Although xanthine oxidase inhibitors are in rates. Furthermore, the occurrence and intensity
widespread clinical use for the treatment of pa- of gout flares are difficult to capture accurately
tients with gout,12 data on the cardiovascular in clinical trials. Finally, these findings may have
safety of these drugs from large, randomized been due to chance, given the large number of
clinical trials are limited. During a development tests performed and the small numbers of events
program involving more than 5000 patients, the in each subgroup.
rate of cardiovascular events was higher among Important limitations of this trial are the
patients treated with febuxostat (0.74 per 100 large number of participants who discontinued
patient-years; 95% CI, 0.36 to 1.37) than among the trial treatment and the large number of par-
those treated with allopurinol (0.60 per 100 pa- ticipants who did not complete follow-up. Dis-
tient-years; 95% CI, 0.16 to 1.53).6-8 In contrast, continuation of treatment would be expected to
observational evaluations have suggested benefi- bias the analyses toward the null hypothesis,
cial cardiovascular outcomes after treatment with which could have resulted in missing a significant
febuxostat or allopurinol in patients with gout and difference between the groups in the primary or
coexisting cardiorenal conditions.13,14 The popula- nonfatal secondary outcomes. The effect of the
Table 3. Events That Occurred during Treatment or within 30 Days after Discontinuation of Treatment.*
Febuxostat Allopurinol Hazard Ratio

End Point (N = 3098) (N = 3092) (95% CI) P Value
no. of patients (%)

Primary end point: composite of cardiovascular 242 (7.8) 238 (7.7) 1.00 (0.82–1.22)† 0.99
death, nonfatal myocardial infarction,
nonfatal stroke, or urgent revasculariza-
tion due to unstable angina
Secondary end points
Cardiovascular death 62 (2.0) 41 (1.3) 1.49 (1.01–2.22) 0.047
Nonfatal myocardial infarction 93 (3.0) 106 (3.4) 0.87 (0.66–1.15) 0.32
Nonfatal stroke 59 (1.9) 62 (2.0) 0.94 (0.66–1.34) 0.72
Urgent revascularization for unstable angina 45 (1.5) 44 (1.4) 1.00 (0.66–1.52) 0.98
Composite of cardiovascular death, nonfatal 205 (6.6) 200 (6.5) 1.01 (0.83–1.22) 0.93
myocardial infarction, or nonfatal stroke
Death from any cause 92 (3.0) 72 (2.3) 1.26 (0.93–1.72) 0.14
* This analysis was prespecified in the statistical analysis plan.

† The 97% confidence interval is provided here.
high rate of loss to follow-up is less easy to predict, resulted in overall rates of major adverse cardio-
since it may not have been random; however, vascular events similar to those associated with
approximately equal numbers of patients discon- allopurinol. Higher all-cause mortality, resulting
tinued follow-up in the two treatment groups, from an imbalance in cardiovascular deaths, was
and the baseline characteristics of these partici- observed with febuxostat than with allopurinol.
pants were similar to those of participants who
completed follow-up. Supported by Takeda Development Center Americas.
Disclosure forms provided by the authors are available with
In conclusion, among patients with gout and the full text of this article at NEJM.org.
cardiovascular disease, treatment with febuxostat We thank all the patients who participated in the trial.
References
1. Roddy E, Doherty M. Epidemiology of 6. Becker MA, Schumacher HR Jr, Wort- American College of Cardiology/Ameri-
gout. Arthritis Res Ther 2010;12:223-34. mann RL, et al. Febuxostat compared can Heart Association Task Force on
2. Choi HK, Curhan G. Independent im- with allopurinol in patients with hyper- Clinical Data Standards (Writing Com-
pact of gout on mortality and risk for uricemia and gout. N Engl J Med 2005; mittee to Develop Cardiovascular End-
coronary heart disease. Circulation 2007; 353:2450-61. points Data Standards). Circulation 2015;
116:894-900. 7. Schumacher HR Jr, Becker MA, Lloyd 132:302-61.
3. Krishnan E, Svendsen K, Neaton JD, E, MacDonald PA, Lademacher C. Febuxo- 11. Lan KKG, DeMets DL. Discrete se-
Grandits G, Kuller LH. Long-term cardio- stat in the treatment of gout: 5-yr find- quential boundaries for clinical trials.
vascular mortality among middle-aged men ings of the FOCUS efficacy and safety Biometrika 1983;70:659-63.
with gout. Arch Intern Med 2008; 168: study. Rheumatology (Oxford) 2009; 48: 12. Khanna D, Khanna PP, Fitzgerald JD,
1104-10. 188-94. et al. 2012 American College of Rheuma-
4. FDA announces new recommenda- 8. White WB, Chohan S, Dabholkar A, tology guidelines for the management of
tions on evaluating cardiovascular risk in Hunt B, Jackson R. Cardiovascular safety gout. Part 2: therapy and anti-inflamma-
drugs intended to treat type 2 diabetes. of febuxostat and allopurinol in patients tory prophylaxis of acute gouty arthritis.
Pharmaceutical Online. December 18, 2008 with gout and cardiovascular comorbidi- Arthritis Care Res 2012;63:1447-61.
(https://w ww.pharmaceuticalonline.com/ ties. Am Heart J 2012;164:14-20. 13. Foody J, Turpin RS, Tidwell BA, Law-
doc/fda-announces-new-recommendations 9. Wallace SL, Robinson H, Masi AT, rence D, Schulman KL. Major cardiovas-
-on-0001). Decker JL, McCarty DJ, Yü TF. Preliminary cular events in patients with gout and as-
5. Schumacher HR Jr, Becker MA, Wort- criteria for the classification of the acute sociated cardiovascular disease or heart
mann RL, et al. Effects of febuxostat ver- arthritis of primary gout. Arthritis failure and chronic kidney disease initiat-
sus allopurinol and placebo in reducing Rheum 1977;20:895-900. ing a xanthine oxidase inhibitor. Am
serum urate in subjects with hyperurice- 10. Hicks KA, Tcheng JE, Bozkurt B, et al. Health Drug Benefits 2017;10:393-401.
mia and gout: a 28-week, phase III, ran- 2014 ACC/AHA key data elements and 14. Singh JA, Ramachandaran R, Yu S,
domized, double-blind, parallel-group tri- definitions for cardiovascular endpoint Curtis JR. Allopurinol use and the risk of
al. Arthritis Rheum 2008;59:1540-8. events in clinical trials: a report of the acute cardiovascular events in patients

with gout and diabetes. BMC Cardiovasc LR. Febuxostat: drug review and update. dase inhibition with febuxostat attenu-
Disord 2017;17:76. Expert Opin Drug Metab Toxicol 2014;10: ates systolic overload-induced left ven-
15. Wei L, Mackenzie IS, Chen Y, Struthers 747-58. tricular hypertrophy and dysfunction in
AD, MacDonald TM. Impact of allopuri- 18. Sanchez-Lozada LG, Tapia E, Bautis- mice. J Card Fail 2008;14:746-53.
nol use on urate concentration and car- ta-García Pet al. Effects of febuxostat on 21. Hou M, Hu Q, Chen Y, Zhao L, Zhang
diovascular outcome. Br J Clin Pharmacol metabolic and renal alterations in rats J, Bache RJ. Acute effects of febuxostat, a
2011;71:600-7. with fructose-induced metabolic syn- nonpurine selective inhibitor of xanthine
16. Bardin T, Keenan RT, Khanna PP, et drome. Am J Physiol Renal Physiol 2008; oxidase, in pacing induced heart failure.
al. Lesinurad in combination with allopu- 294:F710-718. J Cardiovasc Pharmacol 2006;48:255-63.
rinol: a randomised, double-blind, place- 19. Zhao L, Roche BM, Wessale JL, et al. 22. Krishnan E. Inflammation, oxidative
bo-controlled study in patients with gout Chronic xanthine oxidase inhibition fol- stress and lipids: the risk triad for athero-
with inadequate response to standard of lowing myocardial infarction in rabbits: sclerosis in gout. Rheumatology (Oxford)
care (the multinational CLEAR 2 study). effects of early versus delayed treatment. 2010;49:1229-38.
Ann Rheum Dis 2017;76:811-20. Life Sci 2008;82:495-502. Copyright © 2018 Massachusetts Medical Society.
17. Grewal HK, Martinez JR, Espinoza 20. Xu X, Hu X, Lu Z, et al. Xanthine oxi-

Gota PDF

Uploaded by

Copyright:

Available Formats

Gota PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Gota PDF

Uploaded by

Copyright:

Available Formats

The n e w e ng l a n d j o u r na l of m e dic i n e

Cardiovascular Safety of Febuxostat

The New England Journal of Medicine

Table 1. Baseline Characteristics of the Patients.*

The New England Journal of Medicine

function, which preserves an overall one-sided al-

Table 2. Major Safety End Points (Modified Intention-to-Treat Analysis).*

Febuxostat Allopurinol Hazard Ratio

no. of patients (%)

A Primary End Point

Cumulative Rate (%)

The New England Journal of Medicine

Figure 1 (facing page). Cumulative Kaplan–Meier

Subgroup Febuxostat Allopurinol Relative Risk (95% CI) P Value for

Febuxostat Better Allopurinol Better

The New England Journal of Medicine

Figure 2 (facing page). Risk Ratios for the Primary

Febuxostat Allopurinol Hazard Ratio

no. of patients (%)

* This analysis was prespecified in the statistical analysis plan.

The New England Journal of Medicine

You might also like

* This analysis was prespecified in the statistical analysis plan.