2024 年 31 巻 6 号 p. 864-875
Aims: The anti-inflammatory effects of the xanthine oxidase inhibitor, febuxostat, a urate-lowering agent, have been reported in animal studies. However, the anti-inflammatory effects of urate-lowering therapy and its associated cardiovascular protective effects have not been fully determined in actual clinical practice. This study aimed to investigate the effect of febuxostat on white blood cell (WBC) count in patients with asymptomatic hyperuricemia and to assess for potential correlations between changes in WBC count and inflammatory biomarkers and atherosclerosis in this patient population.
Methods: This was a post hoc subanalysis of the PRIZE study, a multicenter, prospective, randomized, open-label clinical trial. In the PRIZE study, asymptomatic hyperuricemia patients were randomized to febuxostat group or control group with non-pharmacological therapy and evaluated the effect on vascular. The primary endpoints of this study were the assessment of the time course of WBC count over 24 months and its changes from baseline. Correlations of WBC count with high-sensitivity C-reactive protein (hs-CRP) and mean common carotid artery (CCA)-IMT were also exploratorily examined in the febuxostat group.
Results: A total of 444 patients (febuxostat group, n=223; control group, n=221) with WBC measurements available at baseline and at least one of the follow-up time points of 12 or 24 months, were enrolled. Febuxostat modestly, but significantly, reduced WBC counts at 12 and 24 months compared with the baseline levels (P=0.002 and P=0.026, respectively). Notably, the WBC count in the febuxostat group at 12 and 24 months was significantly lower than that in the control group (P=0.007 and P=0.023, respectively). The changes in WBC count were associated with those of hs-CRP (P=0.038), but not with CCA-IMT (P=0.727).
Conclusions: Febuxostat therapy for 24 months modestly, but significantly, decreased WBC count in patients with asymptomatic hyperuricemia. This might potentially reflect a modest anti-inflammatory action of febuxostat in clinical settings.
See editorial vol. 31: 861-863
Inflammation is widely known to play a central role in the development of atherosclerosis1). Hyperuricemia was recently reported to induce inflammation in vascular endothelial cells, resulting in endothelial dysfunction and subsequent atherosclerosis2, 3). Since the report of Mizuno et al. regarding the action of the xanthine oxidase inhibitor and urate-lowering agent, febuxostat, which suppresses the expression of inflammatory cytokines in obese–diabetic model mice4), the anti-inflammatory effect of urate-lowering agents has attracted attention. However, the anti-inflammatory effects of urate-lowering therapy and its associated cardiovascular protective effects have not been fully determined in actual clinical practice. Moreover, the detailed effects of febuxostat on inflammatory markers remain to be fully uncovered in clinical settings.
White blood cell (WBC) count, a common marker of inflammation, has been reported to be associated with various cardiovascular diseases, including coronary artery disease5, 6), and several clinical trials have used it as a surrogate marker to estimate possible therapeutic effect on inflammation7, 8).
In the PRIZE (program of vascular evaluation under uric acid control by xanthine oxidase inhibitor, febuxostat: multicenter, randomized controlled) study, a 24-month treatment with febuxostat did not suppress the increase in carotid intima-media thickness (IMT) in patients with asymptomatic hyperuricemia and carotid plaques9). To date, we have conducted additional sub-analyses examining whether febuxostat exerts off-target effects on non-urate parameters, such as vascular function10, 11), left ventricular diastolic function12), and oxidative stress13). However, the effect of febuxostat on inflammation was not fully elucidated. In the present post hoc subanalysis of the PRIZE study, we sought to investigate the effect of febuxostat on WBC counts in patients with asymptomatic hyperuricemia and assess potential correlations between WBC count and another inflammation-related marker, high-sensitivity C-reactive protein (hs-CRP), and carotid IMT, representing atherosclerosis.
This study was a post hoc subanalysis of the PRIZE study, a multicenter, prospective, randomized, open-label, blinded-endpoint clinical trial. Details of the design of the PRIZE study have been reported previously9, 14). Briefly, recruitment and follow-up of patients were conducted from May 2014 to August 2018 at 48 clinical sites throughout Japan. All participants provided written informed consent before screening and randomization. Eligible patients were equally and randomly assigned to either the febuxostat or control (non-pharmacologic treatment of hyperuricemia) groups via the enrollment website, and were followed up for 24 months after their baseline visit. Carotid artery ultrasonography was performed at baseline and after 12 and 24 months (or at premature termination) at each local site. The PRIZE study was conducted in strict compliance with the Declaration of Helsinki and the Ethical Guidelines for Clinical Research issued by the Ministry of Health, Labour, and Welfare (University Hospital Medical Information Network Clinical Trial Registry UMIN000012911). After publication of the main results of the PRIZE study9), a series of subanalyses, including this work, was approved by the Ethics Committee of the Saga University Hospital (2020-05-R01) and was subsequently registered (UMIN000041322).
Study ParticipantsDetailed inclusion and exclusion criteria for the PRIZE trial have been previously reported9, 14). Briefly, adults (age ≥ 20 years, regardless of sex) with asymptomatic hyperuricemia (serum uric acid levels [SUA] ＞7.0 mg/dL) and a maximum IMT of ≥ 1.1 mm measured at the time of eligibility confirmation were eligible for the study. For the main exclusion criteria, we excluded patients who were receiving medications for hyperuricemia within 8 weeks before the eligibility assessment, had gouty nodules or subjective symptoms of gouty arthritis within 1 year before the eligibility assessment, had cerebro-cardiovascular events and revascularization within 3 months before the eligibility assessment, and those who were pre- or postoperative or having a severe infection or severe trauma during the time of the eligibility assessment. In this analysis, we included patients whose WBC count data were available at baseline and at least one of the follow-up time points of 12 or 24 months.
TreatmentDuring the study period, all subjects were given lifestyle guidance, focusing on a healthy diet, limitation of alcohol consumption, and exercise recommendations. Based on the protocol, patients in the febuxostat group received an initial febuxostat dose of 10 mg/day, titrated to 20 mg/day after 1 month and 40 mg/day after 2 months, with a target maintenance dose of 40 mg/day. After 3 months, the maximum dose was increased to 60 mg/day if possible. If SUA levels decreased to below 2.0 mg/dL during the study period, the dose was reduced by 20 mg. Treatment with antidiabetic, antiplatelet, antihypertensive, and lipid-lowering agents was kept unchanged during the study period to the extent possible, based on the patient’s clinical conditions.
EndpointsThe primary endpoints of this study were the assessment of the time course of WBC count over 24 months and its changes from baseline. Correlations of WBC count with hs-CRP and mean common carotid artery (CCA)-IMT were also exploratorily examined in the febuxostat group. For these endpoints, subgroup analyses according to several clinical background characteristics were also performed.
Measurement of Mean CCA-IMTThe details of the protocols and methods of measuring carotid IMT have been previously described9, 14). Based on the standardized protocol15), a trained sonographer performed a high-resolution carotid ultrasonography using a standard system equipped with a ＞7.5-MHz linear transducer in a blinded manner. The imaging data were stored and sent to the core imaging laboratory, where an expert analyzer measured the IMT values in a blinded manner using an automated measurement software program (Vascular Research Tools 5, Medical Imaging Applications LLC, Coralville, IA). Longitudinal B-mode images perpendicular to the ultrasound beam, with a 3–4 cm imaging depth, were recorded in the distal CCA 10 mm from the carotid bulb. The absolute changes from baseline to 24 months were calculated for each side and then averaged.
Statistical AnalysisContinuous variables and categorical data were described as medians (interquartile ranges) and numbers (percentages), respectively, for expressions of baseline demographics and clinical characteristics. WBC counts and their 95% confidence intervals (CIs) in the febuxostat and control groups at each time point, i.e., at 12 and 24 months, were estimated in a mixed-effects model adjusted for age, sex, and baseline WBC count. Subgroup analyses were performed in the febuxostat group to examine the interactions according to demographic factors on the WBC count. Subgroups were defined based on the following factors: sex, age (＜65 vs. ≥ 65 years), history of hypertension, diabetes, dyslipidemia, and atherosclerotic cardiovascular disease, body mass index (BMI ＜25 kg/m2 vs. ≥ 25 kg/m2), statin use, and WBC count at baseline (＜overall median of 6200 cells/µL vs. ≥ overall median). We examined the association between uric acid and WBC count changes from baseline to 24 months using a mixed-effect model adjusted for age, sex, and baseline WBC count. We also examined the association between changes in WBC count and changes in hs-CRP and mean CCA-IMT in the febuxostat group using a mixed-effects model adjusted for age, sex, and hs-CRP or mean CCA-IMT baseline values. Although these models included the interaction between the changes in WBC count and each time point, our analysis focused solely on the 24-month time point for the result. Subgroup analyses were also conducted using the same criteria in the main analysis of the WBC trend.
All significant levels were set at 5% and were two-sided. All statistical analyses were performed using R 4.2.2 software (R Foundation for Statistical Computing, Vienna, Austria).
The flow of participants in the PRIZE study has been previously reported. Among a total of 483 subjects defined as the full analysis set (febuxostat, n=239; control, n=244), 444 subjects (febuxostat, n=223; control, n=221) whose WBC count data were available at baseline and 12 or 24 months were included in this subanalysis (Fig.1). Baseline demographic and clinical characteristics of the included participants were almost balanced between the allocated groups, as shown in Table 1. Median patient age was 71.0 years (interquartile range: 63.0–77.0), and one-fifth were female. Almost all patients had a history of hypertension, about 60% had dyslipidemia, and one-third had a history of atherosclerotic cardiovascular disease.
WBC, white blood cells.
| Variables | Overall (n = 444) | Febuxostat (n = 223) | Control (n = 221) |
|---|---|---|---|
| Age, y | 71.0 (63.0–77.0) | 70.0 (63.0–76.0) | 71.0 (63.0–78.0) |
| Females | 90 (20.3) | 47 (21.1) | 43 (19.5) |
| Body mass index, kg/m2 | 24.6 (22.5–27.0) | 24.3 (22.3–26.7) | 24.8 (22.9–27.2) |
| Current smoking | 46 (10.4) | 26 (11.7) | 20 (9.0) |
| Serum uric acid, mg/dL | 7.6 (7.3–8.2) | 7.6 (7.3–8.3) | 7.7 (7.3–8.2) |
| eGFR, mL/min/1.73 m2 | 55.0 (45.6–66.7) | 55.3 (44.8–66.3) | 54.8 (46.8–66.7) |
| Medical history | |||
| Hypertension | 393 (88.5) | 197 (88.3) | 196 (88.7) |
| Diabetes | 162 (36.5) | 79 (35.4) | 83 (37.6) |
| Dyslipidemia | 264 (59.5) | 132 (59.2) | 132 (59.7) |
| ASCVD | 144 (32.4) | 71 (31.8) | 73 (33.0) |
| Medications | |||
| ARB | 255 (57.4) | 125 (56.1) | 130 (58.8) |
| ACEI | 48 (10.8) | 24 (10.8) | 24 (10.9) |
| Diuretic | 127 (28.6) | 63 (28.3) | 64 (29.0) |
| Statin | 219 (49.3) | 107 (48.0) | 124 (50.7) |
| Aspirin | 151 (34.0) | 72 (32.3) | 79 (35.7) |
Data are expressed as median (interquartile range) or number (percentage).
ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; ASCVD, atherosclerotic cardiovascular disease; eGFR, estimated glomerular filtration rate.
Fig.2 shows the time course of predicted WBC counts over 24 months in both groups. The predicted WBC count at baseline was 6190 (95% CI, 6057 to 6323) cells/µL in the febuxostat group and 6175 (95% CI, 6042 to 6307) cells/µL in the control group, indicating a group difference of 15 (95% CI, −166 to 197) cells/µL (P=0.865). At 12 months, the predicted WBC count in the febuxostat group (5919 [95% CI, 5783 to 6055] cells/µL) was lower than that in the control group (6176 [95% CI, 6041 to 6310] cells/µL; group difference: −257 [95% CI, −442 to −72] cells/µL, P=0.007). Similarly, at 24 months, the predicted WBC count in the febuxostat group (5998 [95% CI, 5862 to 6135] cells/µL) was also lower than that in the control group (6217 [95% CI, 6079 to 6356] cells/µL; group difference: −219 [95% CI, −407 to −30] cells/µL, P=0.023). In the febuxostat group, the change in WBC count from baseline was significant at both time points:12 months (−272 [95% CI, −440 to −104] cells/µL, P=0.002) and 24 months (−192 [95% CI, −361 to −23] cells/µL, P=0.026); whereas, the control group showed no significant change (Table 2).
The time course of predicted WBC counts in over 24 months in the febuxostat and control groups is shown. At 12 months, the predicted WBC count in the febuxostat group was lower than that in the control group (P=0.007). Similarly, at 24 months, the predicted WBC count in the febuxostat group was also lower than that in the control group (P=0.023).
WBC, white blood cells.
| Time | Febuxostat | Control | ||||
|---|---|---|---|---|---|---|
| Change from baseline | 95% CI | P-value | Change from baseline | 95% CI | P-value | |
| 12 months | −272 | −440 to −104 | 0.002 | 1 | −167 to 169 | 0.991 |
| 24 months | −192 | −361 to −23 | 0.026 | 43 | −128 to 214 | 0.624 |
CI, confidence interval.
Fig.3 shows the WBC count trends in the subgroups stratified according to background clinical characteristics in the febuxostat group. There were no significant differences in WBC counts at 24 months between the groups, except in patients subgrouped according to BMI (25 kg/m2) (P=0.020, Fig.3G).
The WBC count trends over 24 months in the subgroups stratified according to background clinical characteristics in the febuxostat group are shown. Patients were stratified for analyses according to the following clinical factors: sex (A), age (B), hypertension (C), diabetes (D), dyslipidemia (E), ASCVD (F), BMI (G), statin use (H), and baseline WBC count (I).
ASCVD, atherosclerotic cardiovascular disease; BMI, Body mass index; WBC, white blood cell.
WBC, white blood cells.
Fig.4 shows the association between the WBC count and clinical parameters from baseline to 24 months in the febuxostat group. When WBC count was analyzed as an outcome, there was no significant association with SUA levels (P=0.142, Supplementary Fig.1). Next, we analyzed hs-CRP and mean CCA-IMT separately as outcomes and found that a change in the WBC count was significantly associated with a change in hs-CRP (P=0.038, Fig.4A), but not with mean CCA-IMT (P=0.727, Fig.4B). The same analysis was performed in subgroups stratified by clinical characteristics (Supplementary Figs.2 and 3). In the analysis of WBC count and hs-CRP, while there were differences in some of the subgroups (age ≥ 65 years, with hypertension, no diabetes, statin usage, WBC at baseline ＜median), there were no differences in the analysis of mean CCA-IMT. There was also no significant interaction in the association between changes in WBC count and hs-CRP, except for subgroups stratified according to the presence of diabetes and WBC count at baseline (Supplementary Fig.2).
CCA-IMT, common carotid artery intima-media thickness; hs-CRP, high-sensitivity C-reactive protein; WBC, white blood cells.
SUA, serum uric acid; WBC, white blood cells.
Patients were stratified for analyses according to the following clinical factors: Sex (A), age (B), hypertension (C), diabetes (D), dyslipidemia (E), ASCVD (F), BMI (G), statin use (H), and baseline WBC count (I).
ASCVD, atherosclerotic cardiovascular disease; BMI, Body mass index; hs-CRP, high-sensitivity C-reactive protein; WBC, white blood cell.
Patients were stratified for analyses according to the following clinical factors: Sex (A), age (B), hypertension (C), diabetes (D), dyslipidemia (E), ASCVD (F), BMI (G), statin use (H), and baseline WBC count (I).
ASCVD, atherosclerotic cardiovascular disease; BMI, Body mass index; CCA-IMT, common carotid artery intima-media thickness; WBC, white blood cell.
In this subanalysis of the PRIZE study, compared with baseline levels, febuxostat modestly, but significantly reduced WBC counts at 12 months and 24 months, and the predicted WBC count in the febuxostat group was significantly lower than that in the control group. The change in WBC count was associated with the change in hs-CRP, but it was not significantly associated with the change in mean CCA-IMT. Our findings might potentially reflect a modest anti-inflammatory action of febuxostat in a clinical setting.
Inflammation is deeply involved in the pathogenesis of atherosclerosis and cardiovascular diseases16). Several medications, including statins, have been attracting attention for their pleiotropic anti-inflammatory effects17, 18). Recently, the Canakinumab Anti-inflammatory Thrombosis Outcomes Study (CANTOS)19) and the Colchicine Cardiovascular Outcomes Trial (COLCOT)20) have shown that these drugs, on top of statin treatment, reduced the risk of cardiovascular events. However, clinical trials of febuxostat showed no clear benefit in terms of the risk of atherosclerotic cardiovascular diseases21) and atherosclerosis9). The same is true for urate-lowering therapy, including allopurinol22, 23). These results suggest that a reduction in SUA levels might not necessarily contribute to a reduction in the risk of such events. One reason for this observation might be that although hyperuricemia causes an increased inflammatory response2), at the practical clinical level, urate-lowering therapy might not be sufficient to suppress inflammation or detect the effects of treatment. In fact, to date, there has been no clear clinical evidence of inflammation suppression by urate-lowering drugs.
Our findings might represent one of the subtle clinical effects of febuxostat on subclinical inflammatory responses in patients with asymptomatic hyperuricemia. In contrast, no apparent effect on hs-CRP, a common marker of inflammation, was observed in the PRIZE study24). The reasons for the different effects on both these inflammatory markers are unknown, and the mechanism and clinical significance of the decrease in WBC count observed in this study needs further verification. At the same time, it is necessary to establish a treatment strategy for patients with hyperuricemia that not only lowers SUA levels but also effectively suppresses inflammation and improves cardiovascular prognosis.
In this study, WBC counts decreased significantly after two years of febuxostat treatment, suggesting its anti-inflammatory effect. Reportedly, uric acid taken up by adipocytes and vascular endothelial cells activates NADPH oxidase, which increases intracellular reactive oxygen species and radicals, and intracellular uric acid is an oxidation-promoting substance and is considered one of the causes of metabolic syndrome and atherosclerosis25, 26). Therefore, we hypothesized that febuxostat might induce an anti-inflammatory effect by decreasing SUA levels. However, although febuxostat significantly reduced SUA levels in the PRIZE study9), our analysis showed that there was no correlation between changes in SUA levels and changes in WBC counts. Intracellularly, urate is a pro-oxidant, whereas extracellularly, it buffers reactive oxygen species27, 28). Thus, because urate has opposing effects of antioxidation and oxidation, decreasing SUA levels might not necessarily reduce inflammation. Further studies are needed to determine whether febuxostat exerts its anti-inflammatory effect by lowering SUA levels.
We conducted several subgroup analyses to investigate possible factors other than SUA levels that might be associated with febuxostat-induced reduction in WBC count. Since adipose tissue itself is known to undergo inflammatory changes in obesity29), we expected that febuxostat might lower WBC counts more effectively in subjects who are obese. However, the WBC count at 24 months was significantly lower in subjects who were not obese (BMI ≤ 25 kg/m2) than in those who were obese (BMI ＞25 kg/m2). Although we currently have no exact reason for this unexpected result, the effect of febuxostat on WBC counts might be mediated by different pathways from obesity-derived inflammation. Although other subgroups stratified according to factors that might potentially influence inflammatory changes were also analyzed, none of them showed significant differences.
In another subanalysis of the PRIZE study, febuxostat did not decrease hs-CRP24), while in the present subanalysis, a modest but significant reduction in WBC count was observed in patients treated with febuxostat. We also found an association between changes in WBC count and hs-CRP at 24 months in the febuxostat group (Fig.4A). These observations might indicate the possibility that WBC count reflects the anti-inflammatory effect of febuxostat more sensitively than hs-CRP. In the subgroup analysis, most of the stratification factors showed no interaction in the associations between changes in WBC count and hs-CRP; however, two stratification factors, diabetes and WBC count at baseline, showed interaction. Interestingly, no significant association was found between patients with diabetes and ≥ median WBC count, whereas a significant association was found in corresponding counterpart subgroups (Supplementary Fig.2). Although the precise reason for the difference in the association between WBC count and hs-CRP only for those two factors, diabetes and WBC count at baseline, is still unclear, febuxostat treatment may differentially affect the impact on individual inflammatory markers and their associations in patients with diabetes or a higher WBC count, which indicates a potential presence of a systemic inflammatory status. Since WBC count is reportedly closely related to the progression of atherosclerosis30), we further examined whether there was any relationship between WBC count and mean CCA-IMT and found none. This subgroup analysis also showed similar results. This might be, at least in part, because febuxostat did not exert a sufficient anti-inflammatory effect to suppress atherosclerosis in this study’s patient population9) since the observed reduction in WBC count was also modest.
The possibility that febuxostat-induced myelosuppression might have caused the decrease in WBC count also requires discussion. Xanthine oxidase inhibitors with a purine skeleton, such as allopurinol, inhibit nucleic acid synthesis and are known to cause myelosuppression as one of the serious side effects. However, febuxostat does not have a purine skeleton and might cause less myelosuppression compared with allopurinol. In the PRIZE study, clinically apparent pancytopenia as an adverse effect was not reported by local investigators. At the least, the decreased WBC counts observed in the febuxostat group were unlikely to have been attributable to clinically apparent febuxostat-induced myelosuppression.
There are several limitations to this study. First, the inflammatory markers, especially WBC count, are susceptible to infection and other clinical statuses, potentially making them unsuitable endpoints. Although there were no relevant adverse events such as infections to suggest an obvious bias between the febuxostat and control groups in the PRIZE study9), we cannot deny the possibility that subclinical or unreported clinical events could have affected the WBC count. Second, the number of each fraction of WBCs was not measured in this study. The neutrophil-to-lymphocyte ratio has attracted attention as a marker of systemic inflammation, and its usefulness as a prognostic factor in cardiovascular and cerebrovascular diseases has been reported31, 32). Hence, the effects of febuxostat on WBC fractions should also be addressed in the future. Third, a 2-year observation period might have been too short to prevent carotid IMT progression and to examine the effects of febuxostat-induced WBC count reduction on atherosclerosis. Extending the observation period might help further assessment of the effects of febuxostat on inflammatory status and even atherosclerosis. Finally, since the present subanalysis only included patients with asymptomatic hyperuricemia, their baseline inflammatory status might have been less activated compared with that of patients with active gout. Patients with symptomatic hyperuricemia are considered to be in a more advanced pro-inflammatory state via the precipitation and activation of accumulated urate crystals33). Inclusion of patients with symptomatic hyperuricemia might help in clarifying the anti-inflammatory effects of febuxostat.
Twenty-four months of febuxostat treatment modestly, but significantly, decreased WBC counts in patients with asymptomatic hyperuricemia. Our findings might potentially reflect a modest anti-inflammatory action of febuxostat in clinical settings. Further research is needed to elucidate the anti-inflammatory effect of febuxostat and its clinical significance in greater detail.
The authors thank all the participants and staff for their contributions to the PRIZE study.
The PRIZE study was funded by Teijin Pharma Limited, Japan, with the funding received by K.N. The funder of the trial had no role in the study design, data collection, data analysis, data interpretation, or writing of the report.
Conceptualization, M.T., A.T., K.N.; methodology, M.T., A.T.; data curation, M.T.; investigation, A.T., I.N., Y.S., S.H., A.K., M.O., T.I., K.N.; formal analysis, H.Y.; writing – original draft preparation, M.T., A.T.; writing – review & editing, I.N., Y.S., S.H., A.K., M.O., T.I., K.N.; supervision, A.T., K.N. All authors have read and agreed to the published version of the manuscript.
A.T. received honoraria from Boehringer Ingelheim, and research funding from GlaxoSmithKline, Takeda, Bristol-Myers Squibb, and Novo Nordisk. H.Y. reported receiving lecture fee from Kyowa Kirin. N.K. has received honoraria from AstraZeneca, Bayer Yakuhin, Boehringer Ingelheim Japan, Daiichi Sankyo, Eli Lilly Japan, Kowa, Mitsubishi Tanabe Pharma, MSD, Novartis Pharma, Novo Nordisk Pharma, Ono Pharmaceutical, Otsuka, and Takeda Pharmaceutical; research grant from Asahi Kasei, Astellas, Boehringer Ingelheim Japan, Fuji Yakuhin, Mitsubishi Tanabe Pharma, Mochida Pharmaceutical, Novartis Pharma, and Teijin Pharma; scholarship from Bayer Yakuhin, Medtronic, and Teijin Pharma. A.T. is an Editorial Board member of the journal, and they were not involved in handling this manuscript during the submission and review processes. All other authors declare no competing interests.
