Effect of Uric Acid–Lowering Agents on Endothelial FunctionNovelty and Significance
A Randomized, Double-Blind, Placebo-Controlled Trial
Higher levels of serum uric acid are independently associated with endothelial dysfunction, a mechanism for incident hypertension. Overweight/obese individuals are more prone to endothelial dysfunction than their lean counterparts. However, the effect of lowering serum uric acid on endothelial dysfunction in these individuals has not been examined thoroughly. In this randomized, double-blind, placebo-controlled trial of nonhypertensive, overweight, or obese individuals with higher serum uric acid (body mass index ≥25 kg/m2 and serum uric acid ≥5.0 mg/dL), we assigned subjects to probenecid (500–1000 mg/d), allopurinol (300–600 mg/d), or matching placebo. The primary outcome was endothelium-dependent vasodilation measured by brachial artery ultrasound at baseline and 8 weeks. By the end of the trial, 47, 49, and 53 participants had been allocated to receive probenecid, allopurinol, and placebo, respectively. Mean serum uric acid levels significantly decreased in the probenecid (from 6.1 to 3.5 mg/dL) and allopurinol groups (from 6.1 to 2.9 mg/dL) but not in the placebo group (6.1 to 5.6 mg/dL). None of the interventions produced any significant change in endothelium-dependent vasodilation (probenecid, 7.4±5.1% at baseline and 8.3±5.1% at 8 weeks; allopurinol, 7.6±6.0% at baseline and 6.2±4.8% at 8 weeks; and placebo, 6.5±3.8% at baseline and 7.1±4.9% at 8 weeks). In this randomized, double-blind, placebo-controlled trial, uric acid lowering did not affect endothelial function in overweight or obese nonhypertensive individuals. These data do not support the hypothesis that uric acid is causally related to endothelial dysfunction, a potential mechanism for development of hypertension.
Many studies have found an increased risk of hypertension among those with higher serum uric acid levels. In a recent meta-analysis of 25 studies, a 1 mg/dL higher serum uric acid was independently associated with a 15% higher risk of developing hypertension.1 In addition, hyperuricemia was associated with endothelial dysfunction, a known risk factor for hypertension and cardiovascular disease (CVD), in both hypertensive and nonhypertensive individuals.2–5 Animal studies suggest that uric acid produces an increase in blood pressure due in part to endothelial dysfunction.6 Taken together, higher levels of uric acid may be causally associated with the development of hypertension in humans, and this may be mediated by endothelial dysfunction.
Several interventional studies have examined the effect of uric acid lowering on endothelial function in different populations. In a meta-analysis of 9 trials of hypertensive or normotensive individuals, allopurinol significantly improved endothelial function by 2.7% (assessed by either flow-mediated dilation or acetylcholine-mediated blood flow).7 None of these, however, were low risk-of-bias trials, and all but one included patients with substantial comorbidities (eg, heart or renal failure and diabetes mellitus). In addition, the direct effects of allopurinol (known to be a powerful antioxidant) could not be distinguished from the effects of uric acid lowering. Thus, whether or not reducing uric acid levels can improve endothelial function among individuals without hypertension (and thereby potentially prevent hypertension) is uncertain.
We, therefore, conducted a randomized, double-blind, placebo-controlled trial to determine the effect of uric acid lowering on endothelial function among nonhypertensive individuals with uric acid levels ≥5 mg/dL.
Potential subjects were recruited from the metropolitan Boston area using electronic and print advertisements and recruitment systems set up by Brigham and Women’s Hospital and Partners Healthcare. This study aimed to address the potential role of uric acid as a modifiable risk factor for the new development of hypertension, potentially mediated by endothelial dysfunction. Thus, we enrolled participants who were overweight or obese and who are, therefore, likely to have pre-existing endothelial dysfunction8 and who are at high risk for hypertension but without a current diagnosis of hypertension.
Therefore, inclusion criteria included a body mass index ≥25 kg/m2 and serum uric acid level ≥5 mg/dL. Potentially eligible subjects were excluded if they were pregnant or had a history of hypertension (defined as any known history of hypertension, a systolic blood pressure ≥140 mm Hg, a diastolic blood pressure ≥90 mm Hg, or use of antihypertensive drugs), diabetes mellitus, chronic liver disease (or abnormally elevated alanine aminotransferase or aspartate aminotransferase), coronary heart disease, chronic kidney disease (defined as an estimated glomerular filtration rate <60 mL/min/1.73 m2),9 known active malignancy (except nonmelanoma skin cancer), or kidney stones.
After undergoing an initial phone screen, 549 individuals were invited for a screening visit where a study physician explained the study and performed a detailed history and physical examination. Information about medical history including medication use was obtained, as were measurements of blood pressure, serum uric acid, serum creatinine, liver function tests, and electrolytes. All subjects provided written informed consent. The study was approved by the Institutional Review Board at Brigham and Women’s Hospital and conducted in accordance with the principles of the Declaration of Helsinki.
Study Design and Treatments
The study design is depicted in Figure 1. Willing and eligible subjects were scheduled for 3 additional visits after the initial screen, including a baseline visit (visit 1, after which study medication was provided), a check-up visit at 4 weeks (including pill count, symptom check, and laboratory studies to evaluate for adverse effects, such as liver function tests and serum creatinine), and a final visit at 8 weeks (visit 2, after which study medication was discontinued). The same assessments were performed at the baseline visit (visit 1) and the 8-week follow-up visit (visit 2). Assessments of endothelial function were performed (as described below) after an 8-hour fast.
Subjects were randomly assigned to be treated initially with either probenecid 500 mg once daily, allopurinol 300 mg once daily, or placebo once daily. Randomization was computer generated and stratified by race and sex. An unblinded Brigham and Women’s Hospital’s research pharmacist performed the randomization; all other individuals were blinded to treatment assignment, including subjects, investigators, research nurses and research assistants, outcome assessors, and data analyzers.
A 4-week supply of study drugs was provided at the baseline visit by the research pharmacist, with instructions to take 1 pill daily of the assigned drug, to be started after completion of the baseline procedures. All medications were formulated to be physically indistinguishable (ie, by sight, smell, and weight).
Participants returned for a check-up visit at 4 weeks, at which time serum uric acid was remeasured. The unblinded research pharmacist viewed the serum uric acid result and provided the subject with the subsequent 4-week supply of medication. This second 4-week supply of medication included instructions to take 2 pills per day (one from each of two bottles); if the serum uric acid was <3 mg/dL, those in the allopurinol group received 1 bottle of allopurinol (300 mg) and 1 bottle of placebo, whereas if the serum uric acid was ≥3 mg/dL, the research pharmacist provided 2 bottles that each contained allopurinol (300 mg, for a total daily dose of 600 mg). A similar strategy was used for subjects in the probenecid arm (who took either 500 mg or 1000 mg daily during the second 4 weeks of the study, depending on serum uric acid concentration at the 4-week check-up visit). Subjects in the placebo arm were provided with 2 bottles of placebo. Study staff contacted subjects weekly to ascertain side effects and compliance with study medication. Study treatments were discontinued after the visit at 8 weeks was completed.
Serum Uric Acid Levels
Serum uric acid was measured at the outpatient screening to determine study eligibility, as well as at baseline, at the 4-week check-up visit, and at the 8-week follow-up visit. The measurements were performed using the Roche Cobas c analyzer with a colorimetric uricase method (Roche Diagnostics, Indianapolis, IN) with an intra-assay coefficient of variation of 0.6% and an interassay coefficient of variation of 1.3%.
Subjects’ endothelial function was assessed by ultrasound in the morning of the baseline and 8-week visits. Caffeine and alcohol were restricted for at least 12 hours and food for 8 hours before the procedure. Brachial artery diameter was measured by 1 experienced technician under basal conditions and during reactive hyperemia after 5-minute ischemic occlusion, using a methodology recommended by the International Brachial Artery Reactivity Task Force.10 This flow-mediated, endothelium-dependent vasodilation (EDV) of the brachial artery, expressed as a percent change of brachial luminal diameter, occurs at 1 minute of reactive hyperemia and is mediated by NO. Endothelium-independent vasodilation (EIV) was assessed by measuring brachial artery diameter under basal conditions and 3 minutes after administration of sublingual nitroglycerin (0.4 mg); this measurement serves as a control for underlying vascular smooth muscle cell function. Both the technician and reader were blinded to group assignment. Acquisition and analysis of the stored images were performed by the Brigham and Women’s Ultrasound Core Laboratory using software designed for this purpose by Medical Imaging Applications (Coralville, IA).
Homeostatic Model Assessment of Insulin Resistance
Higher levels of uric acid are associated with endothelial dysfunction and development of diabetes mellitus.11 Additionally, in hypertensive participants with normal glucose tolerance tests, uric acid is associated with 1-hour postload glucose.12 Thus, we decided to analyze the association of uric acid lowering on homeostatic model assessment of insulin resistance (HOMA-IR), as a possible intermediate in the development of endothelial dysfunction. Serum glucose (mg/dL) and serum insulin (uIU/mL) were measured at visits 1 and 2 and used to compute the HOMA-IR index.13
We compared participant’s baseline characteristics by treatment assignment to assess the balance of potential confounder. We evaluated the change of EDV and EIV from baseline to 8 weeks as a continuous variable using a paired t test, after we confirmed a normal distribution in endothelial function assessments. We then compared the treatment groups using a 2-group t test. As a secondary analysis, we assessed the change of HOMA-IR from baseline to 8 weeks using a paired t test and a 2-group t test. We also assessed whether the change in EDV between groups differed according to sex, age, body mass index, race, and current smoking status. We also conducted additional sensitivity analyses in which we analyzed the effect of therapy on EDV according to tertile of uric acid. Finally, we evaluated the association of the 8-week change in uric acid level with corresponding changes in EDV and EIV using Pearson correlation coefficients. Statistical significance was set at a 2-tailed P<0.05. All analyses were performed using the SAS statistical package (version 9.4; SAS Institute, Inc, Cary, NC).
A total of 549 subjects were assessed for eligibility (Figure 1). Of these, 149 were randomly assigned to receive probenecid, allopurinol, or placebo. Four subjects discontinued the study before receiving study medication. A total of 120 subjects completed the study (40 assigned to probenecid, 35 to allopurinol, and 45 to placebo). There were no significant differences in serum uric acid levels or any other baseline covariates comparing subjects who completed the study with those who dropped out.
The mean (±SD) age of the randomized population was 41±14 years and mean body mass index was 34±6.3 kg/m2. Mean uric acid level at baseline was 6.1±0.9 mg/dL. There were no significant differences in baseline characteristics of randomized participants among the 3 groups (Table 1).
Serum uric acid levels decreased in both treatment groups as compared with placebo (Figure 2). Uric acid declined from a baseline mean of 6.1 to 4.3 mg/dL at 4 weeks and 3.5 mg/dL at 8 weeks in the probenecid group (P<0.001) and from 6.1 mg/dL to 4.0 mg/dL at 4 weeks and 2.9 mg/dL at 8 weeks in the allopurinol group (P<0.001).
In contrast, the mean serum uric acid level did not fall significantly in the placebo group (6.1 mg/dL at baseline and 5.6 mg/dL at 8 weeks; P=0.48).
There was no statistical difference in the 8-week change in EDV comparing probenecid with placebo (P=0.7) or comparing allopurinol with placebo (P=0.06; Figure 3). EDV in the probenecid group was 7.4% at baseline and 8.3% at 8 weeks; EDV was 7.6% at baseline in the allopurinol group and 6.2% at 8 weeks. The corresponding values in the placebo group were 6.5% and 7.1%. There were also no significant changes in EIV after uric acid lowering with either probenecid or allopurinol as compared with placebo (Figure 3).
We did not observe any significant effect of treatment on 8-week changes of insulin resistance as measured by the HOMA-IR index. HOMA-IR in the probenecid group was 1.9 at baseline and 1.8 at 8 weeks (P=0.5); in the allopurinol group, it was 1.9 at baseline and 8 weeks (P=0.9). There was no difference in the treatment group comparisons in the change in HOMA-IR (P=0.8).
In addition, there were no significant interactions between treatment groups and sex, age, body mass index, race, or current smoking status with respect to endothelial function. Additionally, we did not find an interaction between treatment and tertile of baseline uric acid (P=0.52). We also did not find any associations between 8-week changes in serum uric acid levels and corresponding changes in EDV and EIV (Pearson correlation coefficients of 0.07 for both).
Table 2 depicts adverse events between treatment arms. Abdominal discomfort was the most commonly reported adverse event. Rash was not more frequent with allopurinol than with placebo.
In this randomized, double-blind, placebo-controlled trial of overweight and obese individuals with serum uric acid ≥5 mg/dL, we found that lowering serum uric acid levels did not improve endothelial function. Baseline endothelial function in our study was similar to that in other obese individuals without hypertension and was lower than has been reported in lean subjects.
Animal studies and in vitro studies suggest that hyperuricemia induces endothelial dysfunction. In a rat model of hyperuricemia induced by oxonic acid, for example, the rise in uric acid was associated with a decrease in serum nitrates and nitrites, and this adverse effect was reversed by allopurinol at 1 and 7 days (P<0.001).6 In another experiment of cultured bovine aortic endothelial cells, addition of uric acid to the culture media reduced NO produced in response to vascular endothelial growth factor.6
In addition, numerous human studies reported that higher levels of uric acid are associated with endothelial dysfunction in normotensive individuals.14–16 As a result of this observation, several trials were performed to assess the effect of lowering uric acid on endothelial function. In a meta-analysis of 9 trials by Kanbay et al,7 allopurinol treatment significantly increased endothelium-dependent vasodilation by 2.7% (95% confidence interval, 2.5–2.9%; P<0.001). However, the analyzed studies consisted almost entirely of individuals with type 2 diabetes mellitus, congestive heart failure, or other diseases and, therefore, did not pertain to the primary prevention of hypertension and CVD. In addition, most of these studies were unblinded and, therefore, were subject to potential bias. Only 1 trial included in the meta-analysis enrolled asymptomatic hyperuricemic participants; in that study, those receiving 300 mg of allopurinol per day for 4 months had improved EDV from 7.7% to 8.1% (P=0.003), as compared with hyperuricemic control subjects who received no intervention (EDV of 7.8–7.8 with P=−0.52).17 Although, like our trial, that study included asymptomatic patients with no known CVD, treatment was unblinded, and there were reporting inconsistencies in the number of subjects who received allopurinol. Only one of the trials included in the meta-analysis reported results with probenecid in addition to allopurinol.18 Similar to our trial, that study reported no improvement in EDV when uric acid was lowered with probenecid18; in contrast, they found an improvement in EDV with allopurinol, a finding consistently reported in patients with heart failure, and which the authors attributed to allopurinol’s antioxidant effect on endothelial function in this population.
We found no effect of serum uric acid lowering for 8 weeks on endothelial function, despite effective lowering of uric acid and despite the presence of endothelial dysfunction in our participants at baseline. The endothelial function of our subjects at baseline was comparable to the endothelial function that was reported in other obese populations. As an example, Williams et al8 found lower EDV in obese participants with no known CVD when compared with lean individuals (6.7% versus 9.7%).
Our study has several potential limitations. First, our participants were all overweight or obese, so our findings do not necessarily apply to a lean population. However, by enrolling overweight and obese subjects, we studied a population that had baseline endothelial dysfunction and was at high risk for developing hypertension and other cardiovascular and metabolic disorders. Thus, any beneficial effects of serum uric acid lowering in such individuals would be clinically important. Second, we studied obese adults, whereas profound effects of uric acid lowering (on blood pressure) were noted in obese adolescents,19 and adults may have stiffer arteries that are less amenable to the effects of treatment. However, in various trials of adults discussed earlier,7 allopurinol was found to improve endothelial function, even among adults with significant comorbidities2,17; consequently, it is unlikely that the lack of effect found in our study resulted from inclusion of adults rather than adolescents.
Third, we recruited individuals with serum uric acid levels ≥5.0 mg/dL. It is possible that a population with a higher baseline serum uric acid could have had a benefit on endothelial function from uric acid–lowering therapy. However, epidemiological evidence suggests that the association of higher uric acid levels with worse endothelial function is continuous (as opposed to dichotomous at a certain uric acid threshold) and is present at levels of uric acid within the normal range (continuing below the cutoff we used for eligibility). Fourth, several subjects dropped out before completing the study, particularly in the allopurinol group, and this could have affected our statistical power. However, the change in EDV in the allopurinol group, although not statistically significant, corresponded to a reduction, rather than improvement, in endothelial function; therefore, it is unlikely that dropouts in the allopurinol group limited our ability to find a salutary effect.
Fifth, our study duration was limited to 8 weeks of treatment; it is conceivable that elevated uric acid levels produce endothelial dysfunction that requires a longer period of treatment to recover. However, most of the previous trials that found benefits from uric acid lowering treated for shorter durations than in our study. Our study also has several strengths in its double-blind, placebo-controlled design, and rigorous evaluation of endothelial function using gold-standard techniques.
In conclusion, in this randomized, double-blind, placebo-controlled trial of overweight or obese individuals with no known CVD, lowering serum uric acid with either probenecid or allopurinol did not improve endothelial function. Our findings are not consistent with the hypothesis that uric acid is a modifiable risk factor for endothelial dysfunction.
In summary, although hyperuricemia has been associated with endothelial dysfunction and hypertension, we did not find that lowering uric acid was associated with improvement in endothelial dysfunction in overweight/obese individuals, a population at high risk of developing hypertension. It is important to continue to study possible hypertension-modifiable risk factors given the increasing prevalence of this disease in the United States and in the world.
N.D. Fisher, G.C. Curhan, and J.P. Forman contributed to the conception and design of the study. All authors were involved in the analysis and interpretation of the data. L. Borgi designed and conducted the statistical analysis. L. Borgi worked on the drafting of the article, which was thoroughly reviewed and approved by all authors.
Sources of Funding
This study was funded by research grants 1R01HL105440-01, 1 UL1 TR 00110, 8 UL1 TR000170-05, and 1 UL1 RR025758-04.
- Received September 19, 2016.
- Revision received October 4, 2016.
- Accepted November 29, 2016.
- © 2016 American Heart Association, Inc.
- Zoccali C,
- Maio R,
- Mallamaci F,
- Sesti G,
- Perticone F
- Tomiyama H,
- Higashi Y,
- Takase B,
- Node K,
- Sata M,
- Inoue T,
- Ishibashi Y,
- Ueda S,
- Shimada K,
- Yamashina A
- Corretti MC,
- Anderson TJ,
- Benjamin EJ,
- Celermajer D,
- Charbonneau F,
- Creager MA,
- Deanfield J,
- Drexler H,
- Gerhard-Herman M,
- Herrington D,
- Vallance P,
- Vita J,
- Vogel R
- Perticone F,
- Maio R,
- Tassone JE,
- Perticone M,
- Pascale A,
- Sciacqua A,
- Sesti G
- Perticone F,
- Sciacqua A,
- Perticone M,
- Arturi F,
- Scarpino PE,
- Quero M,
- Sesti G
- Dogan A,
- Yarlioglues M,
- Kaya MG,
- Karadag Z,
- Dogan S,
- Ardic I,
- Dogdu O,
- Kilinc Y,
- Zencir C,
- Akpek M,
- Ozdogru I,
- Oguzhan A,
- Kalay N
- El Solh AA,
- Saliba R,
- Bosinski T,
- Grant BJ,
- Berbary E,
- Miller N
- Kanbay M,
- Huddam B,
- Azak A,
- Solak Y,
- Kadioglu GK,
- Kirbas I,
- Duranay M,
- Covic A,
- Johnson RJ
- George J,
- Carr E,
- Davies J,
- Belch JJ,
- Struthers A
- Soletsky B,
- Feig DI
Novelty and Significance
What Is New?
This is the first study to analyze the effects of uric acid–lowering agents in participants without significant comorbidities.
This is a randomized, double-blind, placebo-controlled trial.
What Is Relevant?
Uric acid lowering did not improve endothelial dysfunction in overweight or obese individuals with no known history of hypertension, diabetes mellitus, or cardiovascular disease.
In a randomized, double-blind, placebo-controlled trial, uric acid lowering was not associated with endothelial dysfunction improvement in nonhypertensive, overweight/obese individuals.