Abstract
Previous studies do not widely support hyperuricemia as a risk factor for stroke and other cardiovascular diseases (CVD). We assessed the relationship between hyperuricemia and ischemic stroke (~ 900 cases) using a large dataset from the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study. We employed a case-cohort design (incident stroke cases and randomly-selected cohort participants) and weighted Cox-proportional-hazard-models to estimate the association of serum urate level ≥6.8 mg/dl (i.e., hyperuricemia) and 6.0 to <6.8 mg/dl versus <6.0 mg/dl (reference) with incident stroke. Analyses were stratified by race, gender, and age. Mediation of cardiovascular disease comorbidities on the serum urate-stroke association was tested. Hyperuricemia was associated with stroke (HR 1.40 [95%CI: 1.10–1.78]) after adjustment for demographic variables and systolic and diastolic blood pressure (SBP and DBP, respectively). This association was substantially attenuated (HR 1.17 [95%CI: 0.90–1.51]) by additional covariate adjustment. In particular, apparent treatment-resistant hypertension (aTRH: SBP≥140 mmHg or DBP≥90 mmHg on 3 antihypertensive medications, or use of ≥4 antihypertensive medications) and the count of antihypertensive medication classes (nAHT) significantly reduced the effect of hyperuricemia on ischemic stroke. Specifically, aTRH and nAHT, respectively, mediate 45% and 43% of the association. There was no effect modification in the association between hyperuricemia and stroke by age, race, or gender. We conclude that hyperuricemia may be a risk factor for stroke. The substantial attenuation of this association by aTRH and nAHT suggests that severe hypertension may be a mediator.
Keywords: uric acid, urate, stroke, hypertension, high blood pressure, epidemiology, treatment resistant hypertension
Graphical Abstract
INTRODUCTION
The prevalence of hyperuricemia (serum urate ≥ 6.8 mg/dl) is increasing in the United States of America and across the world.1, 2 According to the 2007–2008 National Health and Nutrition Examination Survey (NHANES), the prevalence of stroke, myocardial infarction (MI) and hypertension are 1.7, 1.5 and 2.5 times higher, respectively, among those with versus without hyperuricemia,3 but evidence for hyperuricemia as an independent risk factor for cardiovascular disease (CVD) has been equivocal.4–6 Given the association of hyperuricemia and CVD comorbidities, it has been challenging to establish an indisputable role of serum urate as an independent risk factor for CVD outcomes, especially stroke.
Two meta-analytical reviews of prospective observational studies suggest that hyperuricemia is significantly associated with stroke and mortality.7, 8 However, the largest of the component studies, and the only one based on a US population (the Atherosclerosis Risk in Communities study (ARIC)), found that hyperuricemia was not independently associated with ischemic stroke after adjustment for diuretic-treated hypertension.9 Hypertension is one of the leading risk factors for stroke and multiple lines of evidence from epidemiological10, 11 and animal12, 13 studies as well as from clinical trials14 suggest that serum urate may increase blood pressure. Therefore, hypertension may be an intermediate in the pathway (i.e., a mediator) between hyperuricemia and stroke.
Conclusive evidence of an association between serum urate and stroke has been previously limited by the confounding influence of comorbidities and by small sample size, problems that magnify in the analysis of substrata (e.g. analysis by race and gender).7–9 Nevertheless, given observed racial and gender differences in serum urate levels, other reports have suggested the association of serum urate and stroke may differ by race and by gender.15–19 In the current study we examined the association of serum urate with incident stroke in Caucasians and African Americans from the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study. Here we present a comprehensive analysis of the relationship between serum urate and incident ischemic stroke, quantifying the role of individual comorbidities (e.g., blood pressure) in mediating the effect of serum urate on stroke as well as examining differences in the main effect by race, gender, and age.
METHODS AND MATERIALS
Data and Material Disclosure Statement:
Data and materials can be requested by contacting the REGARDS policies and procedures committee at regardsadmin@uab.edu.
Study design and population:
REGARDS, one of the largest population-based prospective cohort studies of Caucasians and African Americans in the US, was designed to assess geographic and racial differences in stroke mortality.20 A total of 30,239 community dwelling participants ≥45 years of age were enrolled between 2003 and 2007. The study collected clinical, demographic and lifestyle information from all participants during a telephone interview and a baseline in-person visit (where urine and blood samples were collected). Participants were then contacted on a semi-annual basis to ascertain hospitalizations and health care encounters for stroke, with medical records retrieved for physician adjudication of suspected events. Written consent was obtained during the in-person visit; the institutional review boards of all participating institutions approved the study methodology.
We used a case-cohort design whereby all incident adjudicated stroke cases and a randomly selected subset of participants were sampled from the full cohort. In a case-cohort design, the randomly selected sub-cohort of participants can develop the outcome of interest and become cases. This is accounted for by the use of weighted proportional hazards analysis to assess the relationship between the exposure and the outcome of interest.21 This sampling design has been used in prior studies in REGARDS.22–25 The current case-cohort study consisted of all participants with incident ischemic stroke during the period from January 1, 2003 to April 1, 2016. The sub-cohort was randomly selected using an age-race-sex-stratified sampling design to ensure the representation of high-risk groups. Figure S1 contains a schematic of the study population.
Exposure:
Serum uric acid was measured in the case-cohort sample using the Roche colorimetric assay (coefficient of variation = 0.87). We categorized levels into clinically relevant categories of < 6; 6 to < 6.8, and ≥ 6.8 mg/dl (hyperuricemia). We used this definition of hyperuricemia as a threshold to allow comparisons to other studies in the literature.8, 26, 27 The reference group of < 6 mg/dl was defined based on current recommendations to maintain serum urate level below 6 mg/dl among individuals treated for gout.28, 29 Hereafter, we refer to the category < 6 mg/dl as the referent, 6.0 to < 6.8 mg/dl as intermediate and ≥ 6.8 mg/dl as hyperuricemia. To assess the robustness of results, we performed sensitivity analyses using tertiles based on the observed distribution of serum urate in the random cohort, < 5.2, 5.2 to < 6.6, and ≥ 6.6 mg/dl, and considering serum urate as a continuous variable.
Outcome:
Ischemic stroke was defined as focal neurologic deficit lasting > 24 hours or non-focal neurological symptoms consistent with stroke upon neuroimaging according to the World Health Organization definition.30 Events were adjudicated by a team of trained neurologists. The stroke sub-types were defined using TOAST (Trial of Org10172 in Acute Stroke Treatment) criteria, and categorized as cardio-embolic, large vessel disease, small vessel disease, unknown sub-type, and other sub-types.31 After excluding 90 participants who had a history of stroke at baseline, 109 participants who had hemorrhagic stroke and 1 participant who had missing follow up information, there were 903 adjudicated incident stroke cases available for analysis in the current study (Figure S1).
Covariates:
Age, race, and smoking status were determined by self-report. Systolic and diastolic blood pressure (SBP and DBP) were obtained during the in-home visit, and defined as the average of 2 seated measures taken after a 5-minute rest. Medication usage (antihypertensive medications, lipid-lowering medications, aspirin, and warfarin) was obtained by study personnel during the in-home visit after reviewing the pharmacy labels/pill bottles for all medications taken by participants in the prior two weeks. Diabetes was defined as self-reported use of insulin or oral glucose lowering agents, a fasting blood glucose level of 126 mg/dl, or a non-fasting glucose concentration of ≥200 mg/dl. Coronary artery disease was defined as self-reported history of myocardial infarction, coronary revascularization, or baseline evidence of a prior myocardial infarction on the study electrocardiogram. Left ventricular hypertrophy and atrial fibrillation were determined from medical history or positive findings on the study electrocardiogram. To investigate the effect of hypertension on the association between uric acid and the risk for stroke, we used SBP and DBP as well as other indicators of hypertension severity. Number of antihypertensive (nAHT) agents was defined as the count of different antihypertensive medication classes. Eligible classes were angiotensin converting enzyme (ACE)-inhibitors, mineralocorticoid receptor antagonists, alpha blockers, angiotensin receptor blockers, beta blockers, calcium channel blockers, central acting agents, diuretics, and direct vasodilators. Apparent treatment resistant hypertension (aTRH) was defined as being taking 4 or more AHT medication classes regardless of blood pressure level, or 3 medication classes with SBP ≥140 mm Hg or DBP ≥90 mm Hg.32–34 The aTRH variable was categorized as the following: a) aTRH; b) individuals on 1 or 2 antihypertensive medications classes regardless of blood pressure level or those with controlled hypertension (SBP/DBP <140/90 mm Hg) on 3 medication classes); c) those not on antihypertensive treatment (reference group).
Statistical Analysis:
We compared participant characteristics across serum urate level categories using a chi-square test for categorical variables and analysis of covariance for continuous variables among participants from the sub-cohort. We fitted a series of weighted (by N of the sample substrata) Cox-proportional hazard models for stroke modeling baseline serum urate as the primary exposure. Model 1 adjusted for age, gender, race and age-race interaction.35 The interaction effect accounts for higher observed stroke risk among African Americans in younger compared to older REGARDS participants. Model 2 accounted for variables in model 1 plus SBP and DBP. To evaluate potential mediators of the association between serum urate and ischemic stroke, we then constructed 14 models (collectively labeled Model 3) that included the covariates in Model 2 in addition to each of the following risk factors, added one at a time: current use of hypertensive medication, nAHT, aTRH, diabetes, left ventricular hypertrophy, high density lipoprotein cholesterol level, atrial fibrillation, coronary artery disease, aspirin use, lipid-lowering medication use, warfarin use, diuretic use, smoking and estimated glomerular filtration rate (eGFR). Model 4 was a fully adjusted model and included adjustment for the variables in model 2 plus aTRH (representing the strongest statistical mediator among hypertension related variables in model 3 (antihypertensive use, aTRH, diuretic use and nAHT)), smoking, diabetes, atrial fibrillation, left ventricular hypertrophy, coronary artery disease, aspirin, lipid medications, warfarin, eGFR and high-density lipoprotein cholesterol level. Further analysis was stratified by race, gender, and age (< 65 years/≥ 65 years).
A sensitivity analysis was performed by fitting models with serum urate as a standardized continuous variable (by subtracting the mean from each observation and dividing by the standard deviation) and categorized by tertile category. Additional analyses were performed excluding individuals on a diuretic given the known serum urate increasing effects of thiazide diuretics.36 We also conducted a sensitivity analysis of aTRH using the new definition of 2018 AHA Scientific Statement on Resistant Hypertension using the blood pressure threshold of SBP/DBP 130/80 mmHg.32 We also limited our analysis to those who have normal renal function (eGFR ≥60 ml/min/1.73m2) to account for possible increase in serum urate due to physiological interference in urate excretion in response to low renal function. In secondary analyses, models 1–4 were fit to each stroke subtype, separately: cardio-embolic, large vessel disease, small vessel disease, other etiology, and undetermined etiology.
We conducted a mediation analysis as described by Judd and Kenny for variables that attenuated the relationship between serum urate and stroke in model 3 and could be a plausible mediator.37 Mediation analysis is based on the rationale that the relationship between the risk factor and outcome is presumed to be working through an intermediate factor or biological mechanism.38–40 Statistically, the intermediate factor should attenuate the significant association between the risk factor (serum urate) and outcome (stroke). Through a mediation analysis, the effect of the exposure (serum urate) can be partitioned into an indirect effect (which is represented by the change in beta-coefficient after including intermediate factor) and a direct effect (which is not explained by the mediator).In brief, we estimated the change in the coefficient of serum urate after addition of the mediator in the model and subsequently obtained a 95% confidence interval for the change (indirect effect) using a bootstrapping technique. We also considered the product of coefficient approach for mediation.41, 42 As we found that the results for the product of coefficient approach were comparable, we present the difference method based on prior literature showing it is the most conservative approach for binary outcomes.39, 43 Finally, for completeness, we assessed the association between the intermediate factors and incident stroke. All the analyses were conducted using SAS 9.4 version.
RESULTS
Baseline characteristics:
Overall, 490 (51.5%), 165 (17.4%) and 296 (31.1%) of the 951 REGARDS study participants in the random sub-cohort had serum urate levels < 6.0 mg/dL, 6.0 to < 6.8 mg/dL and ≥ 6.8 mg/dl (i.e., hyperuricemia), respectively. Descriptive characteristics of the random sub-cohort population by serum urate category are presented in Table 1. Mean SBP was significantly higher in the group with hyperuricemia compared to those with serum urate levels 6.0 to < 6.8 mg/dl or < 6.0 mg/dl. Additionally, the prevalence of comorbidities such as diabetes and low kidney function (defined as estimated glomerular filtration rate < 60 ml/min/1.73 m2) was higher in the high serum urate category, but was relatively similar among those in the low and intermediate category.
Table 1:
Baseline Characteristics of REGARDS participants across serum urate levels in the random sub-cohort
| Characteristics | Overall | Less than 6 mg/dl | 6 to < 6.8 mg/dl | ≥ 6.8 mg/dl | P |
|---|---|---|---|---|---|
| Total | (n=951) | (n=490) | (n=165) | (n=296) | |
| DEMOGRAPHICS | |||||
| Age (years) | 64.6 (12.2) | 63.6 (12.3) | 65.0 (12.3) | 66.0 (11.7) | 0.002 |
| Males N (%) | 471 (49.5) | 197 (40.2) | 83 (50.3) | 191 (64.5) | <0.001 |
| African-Americans N (%) | 465 (48.9) | 208 (42.4) | 93 (56.4) | 164 (55.4) | <0.001 |
| Current Smoker N (%) | 133 (14.0) | 78 (15.9) | 20 (12.1) | 35 (11.8) | <0.001 |
| COMORBIDITIES | |||||
| Diabetes N (%) | 195 (20.5) | 88 (18.0) | 32 (19.4) | 75 (25.3) | 0.001 |
| Atrial Fibrillation N (%) | 78 (8.2) | 42 (8.6) | 9 (5.5) | 27 (9.1) | <0.001 |
| Triglycerides Mean (SD) | 131.4(71.0) | 121.6 (58.6) | 132.2(77.1) | 148.4(81.8) | <0.001 |
| Total Cholesterol Mean (SD) | 191.7(39.2) | 194.2 (39.3) | 191.0(36.7) | 187.7(40.1) | 0.08 |
| LDL Cholesterol Mean (SD) | 113.3(33.8) | 114.0 (33.4) | 115.3(30.9) | 111.1(36.2) | 0.38 |
| HDL Cholesterol Mean (SD) | 51.6 (16.7) | 55.1 (17.5) | 49.0 (15.0) | 47.0 (14.7) | <0.001 |
| SBP Mean (SD) | 127.1(17.1) | 125.5 (17.4) | 126.8(16.5) | 130.1(16.8) | <0.001 |
| DBP Mean (SD) | 76.4 (10.2) | 76.0 (10.2) | 76.0 (9.8) | 77.3 (10.4) | 0.18 |
| LVH N (%) | 84 (8.8) | 34 (6.9) | 8 (4.8) | 42 (14.2) | <0.001 |
| Low kidney function N (%) | 131 (13.8) | 31 (6.3) | 16 (9.7) | 84 (28.4) | <0.001 |
| Albuminuria N (%) | 138 (14.5) | 51 (10.4) | 24 (14.5) | 63 (21.3) | <0.001 |
| Prevalent CAD N (%) | 154 (16.2) | 63 (12.9) | 26 (15.8) | 65 (22.0) | <0.001 |
| MEDICATIONS | |||||
| #Current Use-AHT N (%) | 480 (50.5) | 200 (40.8) | 90 (54.5) | 190 (64.2) | <0.001 |
| aTRH N (%) | <0.001 | ||||
| Untreated | 370 (38.9) | 241 (49.2) | 61 (37.0) | 68 (23.0) | |
| Treated, but not meeting the definition of aTRH | 499 (52.5) | 225 (45.9) | 88 (53.3) | 186 (62.8) | |
| Resistant | 42 (4.4) | 9 (1.8) | 8 (3.6) | 25 (8.4) | |
| AHT Medication classes N (%) | <0.001 | ||||
| 0 | 370 (38.9) | 241 (49.2) | 61 (37.0) | 68 (23.0) | |
| 1 | 210 (22.1) | 100 (20.4) | 88 (53.3) | 70 (23.6) | |
| 2 | 218 (22.9) | 101 (20.6) | 8 (4.8) | 77 (26.0) | |
| ≥3 | 134 (14.1) | 39 (8.0) | 6 (3.6) | 77 (26.0) | |
| Lipid-lowering medication N (%) | 290 (30.5) | 121 (24.7) | 56 (33.9) | 113 (38.2) | <0.001 |
| Diuretics N (%) | 311 (32.7) | 109 (22.2) | 54 (32.7) | 148 (50.0) | <0.001 |
| Aspirin N (%) | 385 (40.5) | 170 (34.7) | 70 (42.4) | 145 (49.0) | <0.001 |
| Warfarin N (%) | 7 (0.7) | 5 (1.0) | 1 (0.6) | 1 (0.3) | <0.001 |
LDL = Low density lipoproteins levels, HDL = High-Density Lipoprotein levels, SBP = Systolic blood pressure, DBP = Diastolic blood pressure, ACR = Albumin-to-Creatinine Ratio; CAD = Coronary artery disease
Current use-AHT is self-reported
Low kidney function = estimated glomerular filtration rate <60 mg/dl; Albuminuria: ACR >30 mg/g; LVH = Left Ventricular Hypertrophy; AHT = Anti-hypertensive; aTRH = apparent Treatment resistant hypertension, categorized as: 0 = not on any hypertensive medications, 1 = individuals on 1 or 2 medications, or have controlled hypertension on 3 medications, 2 = Individuals on 4 or more medications or have uncontrolled hypertension on 3 medications. Number represents raw numbers, and percent indicates weighted percents.
Association between serum urate and stroke:
Hazard ratios for the association between serum urate and incident ischemic stroke in the case-cohort sample are presented in Table 2. The unadjusted estimate for the association of hyperuricemia with ischemic stroke was HR [95% CI] = 1.70 [1.35–2.15] compared to referent group <6 mg/dl. After adjustment for age, gender, race and the age-race interaction, hyperuricemia was associated significantly with a higher risk for stroke (model 1: HR [95% CI] = 1.42 [1.12–1.80]). The association persisted upon further adjustment for SBP and DBP (model 2: HR [95% CI] = 1.40 [1.10–1.78]). In model 3, when individually adjusting for additional risk factors, the HRs were similar to model 2 except when adjusting for nAHT medication classes (HR [95% CI] = 1.17 [0.90–1.51]), aTRH (HR [95% CI] = 1.21 [0.94–1.56]) and eGFR (HR [95% CI] = 1.24 [0.95–1.62]). The relationship between hyperuricemia and stroke was attenuated in the fully adjusted model (model 4: HR [95% CI] = 1.06 [0.78–1.45]). The association between the intermediate serum urate category (i.e., 6.0 to < 6.8 mg/dL) and stroke was statistically significant in the unadjusted model, but it was not significant upon adjustment (models 1–4). Sensitivity analyses demonstrated similar results when serum urate was considered as a continuous variable, when categorized into tertiles, when participants on diuretics were removed, and when the analysis was restricted to participants with normal renal function (Tables S1–S4). As per the new aTRH definition from the 2018 AHA Scientific Statement, 86 participants were reclassified from controlled hypertension category to treatment resistant hypertension category. The results of the sensitivity analysis using the new aTRH definition are almost identical to the main analysis (model 3: HR [95%CI] = 6.0 to 6.8 mg/dL: 1.19 [0.88–1.62]; ≥ 6.8 mg/dL = 1.20 [0.93–1.54]).
Table 2:
Association between baseline serum urate and stroke in REGARDS Case-Cohort Study.
| Models | Events /Total | Less than 6 mg/dl (n=889) | 6 to < 6.8 mg/dl (n=325) | ≥ 6.8 mg/dl (n=640) |
|---|---|---|---|---|
| All Case-Cohort | ||||
| Unadjusted | 903/1854 | Ref | 1.38 (1.03–1.85) | 1.70 (1.35–2.15) |
| Model 1 | 903/1854 | Ref | 1.23 (0.92–1.64) | 1.42 (1.12–1.80) |
| Model 2 | 899/1845 | Ref | 1.25 (0.93–1.68) | 1.40 (1.10–1.78) |
| Model 3 | ||||
| BP + Current use-AHT | 874/1786 | Ref | 1.20 (0.89–1.63) | 1.33 (1.04–1.72) |
| BP + AHT Medication (0,1,2, ≥3) | 878/1805 | Ref | 1.18 (0.87–1.61) | 1.17 (0.90–1.51) |
| BP + aTRH | 878/1805 | Ref | 1.18 (0.87–1.60) | 1.21 (0.94–1.56) |
| BP + Diabetes | 891/1828 | Ref | 1.29 (0.96–1.73) | 1.39 (1.09–1.78) |
| BP + LVH | 886/1816 | Ref | 1.31 (0.97–1.77) | 1.33 (1.04–1.71) |
| BP + HDL | 890/1824 | Ref | 1.24(0.91–1.67) | 1.35 (1.06–1.73) |
| BP + AF | 879/1799 | Ref | 1.28 (0.95–1.73) | 1.40 (1.09–1.80) |
| BP + CAD | 880/1813 | Ref | 1.25 (0.92–1.70) | 1.34 (1.04–1.72) |
| BP + Aspirin | 880/1811 | Ref | 1.27 (0.94–1.71) | 1.37 (1.07–1.75) |
| BP + Lipid-lowering Medications | 869/1794 | Ref | 1.26 (0.93–1.70) | 1.38 (1.08–1.77) |
| BP + Warfarin | 869/1794 | Ref | 1.28 (0.94–1.73) | 1.38 (1.07–1.77) |
| BP + Diuretic | 899/1845 | Ref | 1.25 (0.93–1.68) | 1.39 (1.08–1.80) |
| BP + Smoking | 894/1836 | Ref | 1.26 (0.931.69) | 1.42 (1.11–1.81) |
| BP + eGFR | 893/1832 | Ref | 1.19 (0.88–1.60) | 1.24 (0.95–1.62) |
| Model 4 | 819/1685 | Ref | 1.31 (0.93–1.85) | 1.06 (0.78–1.45) |
BP = systolic blood pressure + diastolic blood pressure, AHT = antihypertensive; aTRH = apparent Treatment resistant hypertension is categorized as: 0 = not on any hypertensive medications, 1= individuals on 1 or 2 medications, or have controlled hypertension on 3 medications, 2 = Individuals on 4 or more medications or have uncontrolled hypertension on 3 medications. LVH = left ventricular hypertrophy, HDL = High Density Lipoproteins, AF = Atrial Fibrillation, CAD = Coronary Artery Disease, eGFR = estimated glomerular filtration rate; Associations represented as hazards ratios (95%) using weighted cox models; Model 1: Adjusted for age, gender, race, interaction between age and race; Model 2: Adjusted for model 1 + hypertension (systolic blood pressure + diastolic blood pressure); Model 3: Adjusted for model 2 + (number of hypertensive medications or treatment resistant hypertension or diabetes or LVH or HDL or AF or CAD or Aspirin or Lipid-lowering Medications or Warfarin or Diuretic or Smoking or eGFR, added one at a time); Model 4: Adjusted for model 2 + aTRH, smoking, diabetes, atrial fibrillation, left ventricular hypertrophy, coronary artery disease, aspirin, lipid-lowering medications, warfarin, estimated glomerular filtration rate and high-density lipoprotein level.
Mediation analysis demonstrated that the effect size of hyperuricemia on stroke using model 2 variables plus the single variable aTRH was significantly reduced (change in beta-coefficient [bootstrap 95%CI]: 0.15 [0.05, 0.26]). Compared to the effect observed from model 2 without aTRH, this represents a 45% reduction in the magnitude of the relationship between hyperuricemia and stroke. A similar pattern was observed when nAHT medication classes was added to the model instead of aTRH: (change in beta-coefficient [bootstrap 95%CI]: 0.14 [0.06 to 0.24]), representing a 43% change in the magnitude of the relationship between hyperuricemia and stroke. aTRH did not significantly mediate the effect of intermediate serum urate levels on stroke [change in beta-coefficient (bootstrap 95%CI): 0.07 (−0.02,0.16)]. In fulfillment of mediation requirements, hyperuricemia is associated with aTRH category (and with increasing nAHT medication classes [Table 1]) and aTRH and nAHT was also significantly associated with stroke in the current data (crude HR for aTRH: 4.10[95%CI 2.69–6.25]; nAHT: 1.35[95%CI1.23–1.48]).
Age, gender and race-specific association between serum urate and stroke:
Results of the race, gender and age stratified analysis are presented in Tables 3–6, S5 and S6. When stratifying by race, the association of hyperuricemia with stroke from the fully adjusted model was similar to the full case-cohort among both African Americans (HR [95% CI] = 1.23 [0.80–1.92]) and Caucasians (HR [95% CI] = 0.96 [0.60–1.53]) (Tables 3 and 4). Results were also similar when stratifying by gender and age (Tables 5–6 and S5–S6). P-values for tests of two-way interaction effects of hyperuricemia by race, gender or age on stroke were all > 0.05, indicating no significant effect modification of the association between hyperuricemia and stroke. Results for the intermediate serum urate category showed some differences by race and gender where risk was higher among males and Caucasians however the stroke counts are low in those substrata.
Table 3:
Association between baseline serum urate and stroke among African American participants.
| Models | Events /Total | Less than 6 mg/dl (n=361) | 6 to < 6.8 mg/dl (n=151) | ≥ 6.8 mg/dl (n=321) |
|---|---|---|---|---|
| All Case-Cohort | ||||
| Unadjusted | 368/833 | Ref | 0.85 (0.55–1.31) | 1.58 (1.12–2.24) |
| Model 1 | 368/833 | Ref | 0.77 (0.51–1.18) | 1.42 (1.02–1.99) |
| Model 2 | 367/829 | Ref | 0.80 (0.52–1.23) | 1.41 (1.00–1.99) |
| Model 3 | ||||
| BP + Current use-AHT | 356/805 | Ref | 0.74 (0.47–1.15) | 1.35 (0.94–1.92) |
| BP + AHT Medication (0,1,2, ≥3) | 356/806 | Ref | 0.76 (0.48–1.20) | 1.26 (0.87–1.82) |
| BP + aTRH | 356/806 | Ref | 0.76 (0.49–1.20) | 1.32 (0.92–1.90) |
| BP + Diabetes | 364/824 | Ref | 0.83 (0.53–1.28) | 1.39 (0.98–1.98) |
| BP + LVH | 360/815 | Ref | 0.81 (0.52–1.26) | 1.30 (0.91–1.86) |
| BP + HDL | 362/819 | Ref | 0.80 (0.51–1.25) | 1.37 (0.96–1.96) |
| BP + AF | 357/803 | Ref | 0.79 (0.51–1.24) | 1.41 (0.99–2.00) |
| BP + CAD | 358/814 | Ref | 0.78 (0.50–1.23) | 1.41 (0.98–2.01) |
| BP + Aspirin | 358/814 | Ref | 0.77 (0.50–1.21) | 1.38 (0.97–1.96) |
| BP + Lipid-lowering Medications | 354/806 | Ref | 0.81 (0.52–1.26) | 1.40 (0.99–2.00) |
| BP + Warfarin | 354/806 | Ref | 0.82 (0.53–1.26) | 1.40 (0.99–1.98) |
| BP + Diuretic | 367/829 | Ref | 0.83 (0.53–1.29) | 1.48 (1.04–2.12) |
| BP + Smoke | 365/826 | Ref | 0.83(0.54–1.27) | 1.40(0.99–1.99) |
| BP + eGFR | 364/824 | Ref | 0.75(0.48–1.17) | 1.28(0.88–1.87) |
| Model 4 | 331/752 | Ref | 0.78 (0.46–1.30) | 1.23 (0.80–1.92) |
BP = systolic blood pressure + diastolic blood pressure, AHT = antihypertensive; aTRH =apparent treatment resistant hypertension, categorized as: 0 = not on any hypertensive medications, 1= individuals on 1 or 2 medications, or have controlled hypertension on 3 medications, 2 = Individuals on 4 or more medications or have uncontrolled hypertension on 3 medications. LVH = left ventricular hypertrophy, HDL = High-Density Lipoproteins, AF = Atrial Fibrillation, CAD = Coronary Artery Disease, eGFR = estimated glomerular filtration rate; Associations represented as hazards ratios (95%) using weighted cox models; Model 1: Adjusted for age, gender; Model 2: Adjusted for model 1 + hypertension (systolic blood pressure + diastolic blood pressure); Model 3: Adjusted for model 2 + (number of hypertensive medications or treatment resistant hypertension or diabetes or LVH or HDL or AF or CAD or Aspirin or Lipid-lowering Medications or Warfarin or Diuretic or Smoking or eGFR, added one at a time); Model 4: Adjusted for model 2 + aTRH, smoking, diabetes, atrial fibrillation, left ventricular hypertrophy, coronary artery disease, aspirin, lipid-lowering medications, warfarin, estimated glomerular filtration rate and high-density lipoprotein level
Table 6:
Association between baseline serum urate and stroke among women
| Models | Events /Total | Less than 6 mg/dl (n=532) | 6 to < 6.8 mg/dl (n=145) | ≥ 6.8 mg/dl (n=240) |
|---|---|---|---|---|
| All Case-Cohort | ||||
| Unadjusted | 437/917 | Ref | 1.11 (0.73–1.68) | 1.80 (1.28–2.53) |
| Model 1 | 437/917 | Ref | 0.92 (0.61–1.39) | 1.50 (1.06–2.12) |
| Model 2 | 435/912 | Ref | 0.86 (0.56–1.32) | 1.38 (0.97–1.98) |
| Model 3 | ||||
| BP + Current use-AHT | 421/881 | Ref | 0.79 (0.51–1.24) | 1.24 (0.84–1.81) |
| BP + AHT Medication (0,1,2, ≥3) | 423/889 | Ref | 0.76 (0.48–1.18) | 1.11 (0.74–1.66) |
| BP + aTRH | 423/889 | Ref | 0.80 (0.50–1.22) | 1.21 (0.82–1.78) |
| BP + Diabetes | 431/903 | Ref | 0.86 (0.56–1.32) | 1.33 (0.92–1.92) |
| BP + LVH | 427/894 | Ref | 0.91 (0.59–1.40) | 1.23 (0.84–1.78) |
| BP + HDL | 430/898 | Ref | 0.82 (0.53–1.28) | 1.31 (0.90–1.89) |
| BP + AF | 424/885 | Ref | 0.84 (0.54–1.29) | 1.37 (0.95–1.98) |
| BP + CAD | 422/888 | Ref | 0.82 (0.51–1.32) | 1.29 (0.88–1.90) |
| BP + Aspirin | 422/887 | Ref | 0.87 (0.56–1.34) | 1.32 (0.91–1.91) |
| BP + Lipid-lowering Medications | 418/879 | Ref | 0.90 (0.58–1.39) | 1.36 (0.94–1.98) |
| BP + Warfarin | 418/879 | Ref | 0.87 (0.56–1.34) | 1.38 (0.95–2.01) |
| BP + Diuretic | 435/912 | Ref | 0.87 (0.57–1.34) | 1.42 (0.98–2.07) |
| BP + Smoke | 432/907 | Ref | 0.87(0.57–1.33) | 1.41(0.97–2.04) |
| BP + eGFR | 432/904 | Ref | 0.80(0.51–1.25) | 1.24(0.83–1.86) |
| Model 4 | 389/818 | Ref | 0.78 (0.46–1.34) | 1.04 (0.62–1.73) |
BP = systolic blood pressure + diastolic blood pressure, AHT = antihypertensive; aTRH = apparent treatment resistant hypertension, categorized as: 0 = not on any hypertensive medications, 1= individuals on 1 or 2 medications, or have controlled hypertension on 3 medications, 2 = Individuals on 4 or more medications or have uncontrolled hypertension on 3 medications. LVH = left ventricular hypertrophy, HDL = High-Density Lipoproteins, AF = Atrial Fibrillation, CAD = Coronary Artery Disease, eGFR = estimated glomerular filtration rate; Associations represented as hazards ratios (95%) using weighted cox models; Model 1: Adjusted for age, race, interaction between age and race; Model 2: Adjusted for model 1 + hypertension (systolic blood pressure + diastolic blood pressure); Model 3: Adjusted for model 2 + (number of hypertensive medications or treatment resistant hypertension or diabetes or LVH or HDL or AF or CAD or Aspirin or Lipid-lowering Medications or Warfarin or Diuretic or Smoking or eGFR, added one at a time); Model 4: Adjusted for model 2 + aTRH, smoking, diabetes, atrial fibrillation, left ventricular hypertrophy, coronary artery disease, aspirin, lipid-lowering medications, warfarin, estimated glomerular filtration rate and high-density lipoprotein level
Table 4:
Association between baseline serum urate and stroke among Caucasian participants.
| Models | Events /Total | Less than 6 mg/dl (n=528) | 6 to < 6.8 mg/dl (n=174) | ≥ 6.8 mg/dl (n=319) |
|---|---|---|---|---|
| All Case-Cohort | ||||
| Unadjusted | 535/1021 | Ref | 1.93 (1.30–2.87) | 1.73 (1.26–2.37) |
| Model 1 | 535/1021 | Ref | 1.77 (1.19–2.62) | 1.41 (1.01–1.97) |
| Model 2 | 532/1016 | Ref | 1.78 (1.19–2.66) | 1.38 (0.98–1.94) |
| Model 3 | ||||
| BP + Current use-AHT | 518/981 | Ref | 1.77 (1.16–2.69) | 1.29 (0.91–1.84) |
| BP + AHT Medication (0,1,2, ≥3) | 522/999 | Ref | 1.66 (1.09–2.52) | 1.07 (0.74–1.54) |
| BP + aTRH | 522/999 | Ref | 1.63 (1.07–2.48) | 1.09 (0.76–1.57) |
| BP + Diabetes | 527/1004 | Ref | 1.82 (1.21–2.73) | 1.38 (0.98–1.94) |
| BP + LVH | 526/1001 | Ref | 1.92 (1.27–2.91) | 1.34 (0.95–1.90) |
| BP + HDL | 528/1005 | Ref | 1.73 (1.15–2.61) | 1.33 (0.94–1.88) |
| BP + AF | 522/996 | Ref | 1.82 (1.20–2.74) | 1.39 (0.97–1.97) |
| BP + CAD | 522/999 | Ref | 1.84 (1.20–2.80) | 1.30 (0.91–1.87) |
| BP + Aspirin | 522/997 | Ref | 1.86 (1.23–2.80) | 1.36 (0.96–1.93) |
| BP + Lipid-lowering Medications | 515/988 | Ref | 1.80 (1.19–2.72) | 1.35 (0.96–1.92) |
| BP + Warfarin | 515/988 | Ref | 1.83 (1.20–2.80) | 1.34 (0.93–1.92) |
| BP + Diuretic | 532/1016 | Ref | 1.78 (1.19–2.66) | 1.28 (0.89–1.84) |
| BP + Smoke | 529/1010 | Ref | 1.71(1.13–2.58) | 1.42 (1.00–2.01) |
| BP + eGFR | 529/1008 | Ref | 1.72(1.14–2.58) | 1.17 (0.80–1.70) |
| Model 4 | 488/933 | Ref | 2.00 (1.24–3.21) | 0.96 (0.60–1.53) |
BP = systolic blood pressure + diastolic blood pressure; AHT = antihypertensive; aTRH = apparent treatment resistant hypertension, categorized as: 0 = not on any hypertensive medications, 1= individuals on 1 or 2 medications, or have controlled hypertension on 3 medications, 2 = Individuals on 4 or more medications or have uncontrolled hypertension on 3 medications. LVH = left ventricular hypertrophy, HDL = High-Density Lipoproteins, AF = Atrial Fibrillation, CAD = Coronary Artery Disease, eGFR = estimated glomerular filtration rate; Associations represented as hazards ratios (95%) using weighted cox models; Model 1: Adjusted for age, gender; Model 2: Adjusted for model 1 + hypertension (systolic blood pressure + diastolic blood pressure); Model 3: Adjusted for model 2 + (number of hypertensive medications or treatment resistant hypertension or diabetes or LVH or HDL or AF or CAD or Aspirin or Lipid-lowering Medications or Warfarin or Diuretic or Smoking or eGFR, added one at a time); Model 4: Adjusted for model2 + aTRH, smoking, diabetes, atrial fibrillation, left ventricular hypertrophy, coronary artery disease, aspirin, lipid medications, warfarin, and high density lipoprotein level; Model 4: Adjusted for model 2 + number of hypertensive medications, smoking, diabetes, atrial fibrillation, left ventricular hypertrophy, coronary artery disease, aspirin, lipid-lowering medications, warfarin, diuretic, estimated glomerular filtration rate and high-density lipoprotein level
Table 5:
Association between serum urate and stroke among men
| Models | Events /Total | Less than 6 mg/dl (n=357) | 6 to < 6.8 mg/dl (n=180) | ≥ 6.8 mg/dl (n=400) |
|---|---|---|---|---|
| All Case-Cohort | ||||
| Unadjusted | 466/937 | Ref | 1.60 (1.06–2.42) | 1.57 (1.12–2.20) |
| Model 1 | 466/937 | Ref | 1.58 (1.05–2.38) | 1.43 (1.03–1.98) |
| Model 2 | 464/933 | Ref | 1.72 (1.13–2.63) | 1.47 (1.05–2.05) |
| Model 3 | ||||
| BP + Current use-AHT | 453/905 | Ref | 1.74 (1.12–2.70) | 1.47 (1.04–2.08) |
| BP + AHT Medication (0,1,2, ≥3) | 455/916 | Ref | 1.74 (1.12–2.70) | 1.28 (0.90–1.82) |
| BP + aTRH | 455/916 | Ref | 1.71 (1.10–2.65) | 1.32 (0.93–1.87) |
| BP + Diabetes | 460/925 | Ref | 1.84 (1.20–2.84) | 1.52 (1.08–2.13) |
| BP + LVH | 459/922 | Ref | 1.79 (1.16–2.76) | 1.48 (1.05–2.07) |
| BP + HDL | 460/926 | Ref | 1.75 (1.14–2.69) | 1.46 (1.04–2.05) |
| BP + AF | 455/914 | Ref | 1.86 (1.21–2.87) | 1.44 (1.01–2.04) |
| BP + CAD | 458/925 | Ref | 1.75 (1.14–2.68) | 1.45 (1.03–2.04) |
| BP + Aspirin | 458/924 | Ref | 1.76 (1.15–2.70) | 1.46 (1.04–2.05) |
| BP + Lipid-lowering Medications | 451/915 | Ref | 1.73 (1.12–2.66) | 1.47 (1.05–2.06) |
| BP + Warfarin | 451/915 | Ref | 1.79 (1.16–2.75) | 1.39 (0.99–1.96) |
| BP + Diuretic | 464/933 | Ref | 1.71 (1.12–2.61) | 1.42 (0.99–2.03) |
| BP + Smoke | 462/929 | Ref | 1.76(1.14–2.70) | 1.50(1.07–2.11) |
| BP + eGFR | 461/928 | Ref | 1.69(1.10–2.60) | 1.26(0.88–1.81) |
| Model 4 | 430/867 | Ref | 2.11(1.29–3.45) | 1.14(0.75–1.73) |
BP = systolic blood pressure + diastolic blood pressure; aTRH =apparent treatment resistant hypertension, categorized as: 0 = not on any hypertensive medications, 1= individuals on 1 or 2 medications, or have controlled hypertension on 3 medications, 2 = Individuals on 4 or more medications or have uncontrolled hypertension on 3 medications. LVH = left ventricular hypertrophy, HDL = High-Density Lipoproteins, AF = Atrial Fibrillation, CAD = Coronary Artery Disease, eGFR = estimated glomerular filtration rate; Associations represented as hazards ratios (95%) using weighted cox models; Model 1: Adjusted for age, race, interaction between age and race; Model 2: Adjusted for model 1 + hypertension (systolic blood pressure + diastolic blood pressure); Model 3: Adjusted for model 2 + (number of hypertensive medications or treatment resistant hypertension or diabetes or LVH or HDL or AF or CAD or Aspirin or Lipid-lowering Medications or Warfarin or Diuretic or Smoking or eGFR, added one at a time); Model 4: Adjusted for model 2 + aTRH, smoking, diabetes, atrial fibrillation, left ventricular hypertrophy, coronary artery disease, aspirin, lipid-lowering medications, warfarin, estimated glomerular filtration rate and high-density lipoprotein level
Association between serum urate and stroke subtypes:
As shown in Table S7, there were no significant associations between hyperuricemia and stroke sub-types. However, those with serum urate 6 to <6.8 mg/dL had significantly higher risk of cardio-embolic stroke in the fully adjusted model compared to those with serum urate levels < 6 mg/dl (HR [95% CI] =2.18 [1.16–4.09]).
DISCUSSION
In the current study we present results from a large, biracial analysis of the association between hyperuricemia and incident ischemic stroke in the Reasons for Geographic and Racial Differences in Stroke study. The results show that hyperuricemia is a risk factor for incident ischemic stroke. In consideration of well-known biological links between hyperuricemia and hypertension as well as hypertension and stroke, we further conducted mediation analyses that showed that indicators of hypertension severity (apparent treatment resistant hypertension and separately count of antihypertensive medication classes) are statistical mediators of the hyperuricemia stroke association. We also report that hyperuricemia is associated with hypertension severity and that hypertension severity is associated with stroke in fulfillment of model assumptions. Our novel findings suggest that severe hypertension may be intermediate in the pathway between hyperuricemia and stroke in a community-based population. Age, gender and race were not important modifiers of the hyperuricemia-stroke association in the REGARDS population.
Many lines of evidence suggest a strong role for hyperuricemia in the etiology of hypertension and hypertension is the risk factor with the largest population-attributable risk for stroke. 10, 44–46 Likewise, aTRH has been independently associated with stroke in prior studies.34, 47 Our data expand this evidence by identifying nAHT medication classes and aTRH as the main blood pressure related factors mediating the serum urate-stroke association in a population based sample. The effect of nAHT medication classes was maintained when participants treated with a diuretic were excluded further supporting a role for hypertension severity (and not diuretic-related increase in serum urate) in the pathway.
Low kidney function is also a risk factor for incident stroke and eGFR attenuated the relationship between hyperuricemia and stroke in our data.47, 48 Serum urate has been reported to be associated with incident CKD,49 and experimental evidence suggests urate may cause renal injury in rat models.50 A recent Mendelian randomization study did not support a causal association of serum urate and eGFR.51 Furthermore, reduced renal function results in reduced renal excretion of urate and higher serum urate levels.52 Thus, the mechanisms of the relationship between kidney function and hyperuricemia are still under study. Our results showed that eGFR attenuated the effect of hyperuricemia on stroke to the same degree as nAHT medication classes and aTRH. However, the lack of consensus in the literature about the pathogenic mechanism of serum urate on CKD makes it premature to consider eGFR as a mediator of the effect of hyperuricemia on stroke. Finally, when analysis of participants with eGFR > 60 ml/min/1.73m2 was performed, the results were consistent with the main analysis.
More than 15 published studies have evaluated the serum urate-stroke relationship. In the most recent meta-analysis, the overall RR (95% CI) for the association of hyperuricemia with stroke was 1.33 (1.11–1.58) after restricting the analysis to 5 studies that had >100 stroke cases.8 In the ARIC study, over 13,000 participants were followed for incident stroke over 12.6 years.9 Similar to our study, there was no statistically significant association of hyperuricemia with stroke after full adjustment (HR [95% CI] = 1.25 [0.91, 1.73]) despite notable differences between the two study populations (the ARIC population is ~10 years younger in comparison to REGARDS and recruited across 4 US metropolitan areas). Additionally, the ARIC study noted effect modification with diuretic use such that the association was absent in the presence of diuretics. This is not consistent with our study that suggested the association is attenuated by nAHT medication classes and aTRH.
Our subgroup analyses by gender demonstrated some substantial differences from prior studies. In the meta-analysis described above women (RR [95%CI] = 1.25 [1.04–1.51]) but not men (RR [95%CI] = 1.08 [0.85–1.38]) showed a higher risk for incident stroke associated with hyperuricemia, and the authors suggested gender differences could underlie the heterogeneity observed in the results.8 Another meta-analysis indicated very little variation among estimates for males and females within thirteen studies.16 In our study, and the ARIC study, gender was not an important modifier of the hyperuricemia-stroke relationship.9 Interestingly, moderately elevated serum urate levels (6 to <6.8 mg/dL) were a risk factor for stroke in males but not females. This result needs to be interpreted with caution since there were only 85 cases in males. Future studies should examine this result in additional populations.
We also investigated age and race as potential effect modifiers of the hyperuricemia-stroke relationship and did not find differences by these variables in the fully adjusted models. Similar to results from REGARDS, the ARIC study concluded that stroke risk associated with hyperuricemia was not different among African American versus Caucasian adults.9 Additionally, a study of incident CVD that included stroke in the Coronary Artery Risk Development in Young Adults (CARDIA) study also found no evidence of a significant race interaction.15 In the current study, the intermediate serum urate category was significantly associated with stroke in Caucasians but not African Americans, though this result requires further study. Overall, in this large, older national sample there were no racial or age differences in the relationship between hyperuricemia and stroke.
The main strength of this study is the use of a general population-based sample where stroke events were carefully adjudicated using medical records by trained physicians with etiologic subtyping. Some limitations include that results only generalize to the two ethnic groups represented in REGARDS. Our findings between intermediate serum urate category and stroke are limited by a smaller number of events especially in substrata. Additionally, error from misclassification of individuals with hyperuricemia may have two sources: 1) an individual at baseline with hyperuricemia may become normouricemic and 2) individuals normouricemic at baseline may become hyperuricemic before stroke incidence. It is expected that these misclassification errors would attenuate the association between ischemic stroke and serum urate.53 Thus we expect the potential bias to be in the direction of the null. Finally, though our rationale for the mediation analysis was based on biological evidence, the study design does not allow to exclude the possibility that hypertension severity may be a confounder of the association between serum urate and stroke. We do encourage that these associations be confirmed in future studies, including animal models.
Supplementary Material
Novelty and Significance:
What is new?
Serum urate is an independent risk factor for ischemic stroke. The current study attempts to quantify the role of individual comorbidities in the relationship between serum urate and stroke. The association between serum urate and stroke is mediated by the presence of hypertension severity
What is Relevant?
We found that the number of antihypertensive medication classes and apparent treatment resistant hypertension may mediate the association of serum urate with stroke risk. If further confirmed, the findings of the study will encourage clinicians to monitor hypertension severity while assessing the risk of stroke in individuals with high serum urate levels. Additionally, targeting this high-risk population and implementing early measures of hypertension prevention may help to reduce the risk for stroke events among patients with hyperuricemia.
Summary
High serum urate levels are a moderate risk factor for ischemic stroke, and the substantial attenuation of this association by apparent treatment resistant hypertension and number of hypertensive medication classes suggests that severe hypertension may be a mediator.
PERSPECTIVES
In conclusion, our novel findings include evidence that hypertension severity and apparent treatment resistant hypertension may mediate the effect of hyperuricemia on stroke. The effect of hyperuricemia on stroke was not modified by age, gender or race in the REGARDS population. This research points to the joint roles of hyperuricemia and hypertension severity and treatment resistance with increasing ischemic stroke risk.
ACKNOWLEDGEMENTS:
• The authors thank the other investigators, the staff, and the participants of the REGARDS study for their valuable contributions. A full list of participating REGARDS investigators and institutions can be found at www.regardsstudy.org
• Part of this findings were presented in an oral presentation at International Stroke Conference 2019.54
SOURCE(s) of FUNDING:
• This research project is supported by cooperative agreement U01 NS041588 co-funded by the National Institute of Neurological Disorders and Stroke (NINDS) and the National Institute on Aging (NIA), National Institutes of Health, Department of Health and Human Service. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NINDS or the NIA. Representatives of the NINDS were involved in the review of the manuscript but were not directly involved in the collection, management, analysis or interpretation of the data.
• KGS reports personal fees from Abbvie, grants and personal fees from Amgen, grants and personal fees from Ironwood/AstraZeneca, personal fees from Bayer, personal fees from Gilead, grants and personal fees from Horizon, personal fees from Kowa, personal fees from Radius, personal fees from Roche/Genentech, grants and personal fees from SOBI, grants and personal fees from Takeda, personal fees from Teijin, outside the submitted work.
• JAS has received consultant fees from Crealta/Horizon, Fidia, UBM LLC, Medscape, WebMD, the National Institutes of Health and the American College of Rheumatology. JAS owns stock options in Amarin pharmaceuticals and Viking therapeutics. JAS is a member of the executive of OMERACT, an organization that develops outcome measures in rheumatology and receives arms-length funding from 36 companies. JAS is a member of the Veterans Affairs Rheumatology Field Advisory Committee. JAS is the editor and the Director of the UAB Cochrane Musculoskeletal Group Satellite Center on Network Meta-analysis. JAS previously served as a member of the following committees: member, the American College of Rheumatology’s (ACR) Annual Meeting Planning Committee (AMPC) and Quality of Care Committees, the Chair of the ACR Meet-the-Professor, Workshop and Study Group Subcommittee and the co-chair of the ACR Criteria and Response Criteria subcommittee.
DISCLOSURES:
• MRI, SJ, GH, LDC, NAL, SLB, AG, JRC, EJR, NSC: None;
• RJR acknowledges support from the Arthritis National Research Foundation.
• EBL received research funding from Amgen, has served Amgen advisory boards, and is a consultant for a Novartis funded research project;
• MC reports grants from NIH, during the conduct of the study;
• MLF reports grants from NIH, during the conduct of the study;
REFERENCES
- 1.Zhu Y, Pandya BJ, Choi HK. Prevalence of gout and hyperuricemia in the us general population: The national health and nutrition examination survey 2007–2008. Arthritis Rheum. 2011;63:3136–3141 [DOI] [PubMed] [Google Scholar]
- 2.Borghi C, Rosei EA, Bardin T, Dawson J, Dominiczak A, Kielstein JT, Manolis AJ, Perez-Ruiz F, Mancia G. Serum uric acid and the risk of cardiovascular and renal disease. J Hypertens. 2015;33:1729–1741; discussion 1741 [DOI] [PubMed] [Google Scholar]
- 3.Abeles AM. Hyperuricemia, gout, and cardiovascular disease: An update. Curr Rheumatol Rep. 2015;17:13. [DOI] [PubMed] [Google Scholar]
- 4.Culleton BF, Larson MG, Kannel WB, Levy DJAoim. Serum uric acid and risk for cardiovascular disease and death: The framingham heart study. 1999;131:7–13 [DOI] [PubMed] [Google Scholar]
- 5.Feig DI, Kang DH, Johnson RJ. Uric acid and cardiovascular risk. N Engl J Med. 2008;359:1811–1821 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kanbay M, Segal M, Afsar B, Kang D-H, Rodriguez-Iturbe B, Johnson RJJH. The role of uric acid in the pathogenesis of human cardiovascular disease. 2013;99:759–766 [DOI] [PubMed] [Google Scholar]
- 7.Kim SY, Guevara JP, Kim KM, Choi HK, Heitjan DF, Albert DA. Hyperuricemia and risk of stroke: A systematic review and meta-analysis. Arthritis Rheum. 2009;61:885–892 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Li M, Hou W, Zhang X, Hu L, Tang Z. Hyperuricemia and risk of stroke: A systematic review and meta-analysis of prospective studies. Atherosclerosis. 2014;232:265–270 [DOI] [PubMed] [Google Scholar]
- 9.Hozawa A, Folsom AR, Ibrahim H, Nieto FJ, Rosamond WD, Shahar E. Serum uric acid and risk of ischemic stroke: The aric study. Atherosclerosis. 2006;187:401–407 [DOI] [PubMed] [Google Scholar]
- 10.Gaffo AL, Jacobs DR Jr., Sijtsma F, Lewis CE, Mikuls TR, Saag KG. Serum urate association with hypertension in young adults: Analysis from the coronary artery risk development in young adults cohort. Ann Rheum Dis. 2013;72:1321–1327 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Perlstein TS, Gumieniak O, Williams GH, Sparrow D, Vokonas PS, Gaziano M, Weiss ST, Litonjua AA. Uric acid and the development of hypertension: The normative aging study. Hypertension. 2006;48:1031–1036 [DOI] [PubMed] [Google Scholar]
- 12.Corry DB, Eslami P, Yamamoto K, Nyby MD, Makino H, Tuck ML. Uric acid stimulates vascular smooth muscle cell proliferation and oxidative stress via the vascular renin-angiotensin system. J Hypertens. 2008;26:269–275 [DOI] [PubMed] [Google Scholar]
- 13.Kang DH, Park SK, Lee IK, Johnson RJ. Uric acid-induced c-reactive protein expression: Implication on cell proliferation and nitric oxide production of human vascular cells. J Am Soc Nephrol. 2005;16:3553–3562 [DOI] [PubMed] [Google Scholar]
- 14.Feig DI, Soletsky B, Johnson RJ. Effect of allopurinol on blood pressure of adolescents with newly diagnosed essential hypertension: A randomized trial. JAMA. 2008;300:924–932 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gaffo AL, Jacobs DR Jr., Lewis CE, Mikuls TR, Saag KG. Association between being african-american, serum urate levels and the risk of developing hyperuricemia: Findings from the coronary artery risk development in young adults cohort. Arthritis Res Ther. 2012;14:R4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Zhong C, Zhong X, Xu T, Xu T, Zhang Y. Sex-specific relationship between serum uric acid and risk of stroke: A dose-response meta-analysis of prospective studies. J Am Heart Assoc. 2017;6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Huang J, Hu D, Wang Y, Zhang D, Qu Y. Dose-response relationship of serum uric acid levels with risk of stroke mortality. Atherosclerosis. 2014;234:1–3 [DOI] [PubMed] [Google Scholar]
- 18.Anton FM, Garcia Puig J, Ramos T, Gonzalez P, Ordas J. Sex differences in uric acid metabolism in adults: Evidence for a lack of influence of estradiol-17 beta (e2) on the renal handling of urate. Metabolism. 1986;35:343–348 [DOI] [PubMed] [Google Scholar]
- 19.Maynard JW, McAdams-DeMarco MA, Law A, Kao L, Gelber AC, Coresh J, Baer AN. Racial differences in gout incidence in a population-based cohort: Atherosclerosis risk in communities study. Am J Epidemiol. 2014;179:576–583 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.National institute on alcohol abuse and alcoholism. Helping patients who drink too much, a clinician’s guide. 2005;2012 [Google Scholar]
- 21.Prentice RL. A case-cohort design for epidemiologic cohort studies and disease prevention trials. Biometrika. 1986;73:1–11 [Google Scholar]
- 22.Zakai NA, Judd SE, Alexander K, McClure LA, Kissela BM, Howard G, Cushman M. Abo blood type and stroke risk: The reasons for geographic and racial differences in stroke study. J Thromb Haemost. 2014;12:564–570 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Cushman M, Judd SE, Howard VJ, Kissela B, Gutierrez OM, Jenny NS, Ahmed A, Thacker EL, Zakai NA. N-terminal pro-b-type natriuretic peptide and stroke risk: The reasons for geographic and racial differences in stroke cohort. Stroke. 2014;45:1646–1650 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Olson NC, Cushman M, Judd SE, Kissela BM, Safford MM, Howard G, Zakai NA. Associations of coagulation factors ix and xi levels with incident coronary heart disease and ischemic stroke: The regards study. J Thromb Haemost. 2017;15:1086–1094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Panwar B, Jenny NS, Howard VJ, Wadley VG, Muntner P, Kissela BM, Judd SE, Gutierrez OM. Fibroblast growth factor 23 and risk of incident stroke in community-living adults. Stroke. 2015;46:322–328 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kim SY, Guevara JP, Kim KM, Choi HK, Heitjan DF, Albert DA. Hyperuricemia and coronary heart disease: A systematic review and meta-analysis. Arthritis Care Res (Hoboken). 2010;62:170–180 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kim SY, Guevara JP, Kim KM, Choi HK, Heitjan DF, Albert DA. Hyperuricemia and risk of stroke: A systematic review and meta‐analysis. Arthritis Care & Research. 2009;61:885–892 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Khanna D, Fitzgerald JD, Khanna PP, Bae S, Singh MK, Neogi T, Pillinger MH, Merill J, Lee S, Prakash S, Kaldas M, Gogia M, Perez-Ruiz F, Taylor W, Liote F, Choi H, Singh JA, Dalbeth N, Kaplan S, Niyyar V, Jones D, Yarows SA, Roessler B, Kerr G, King C, Levy G, Furst DE, Edwards NL, Mandell B, Schumacher HR, Robbins M, Wenger N, Terkeltaub R, American College of R. 2012 american college of rheumatology guidelines for management of gout. Part 1: Systematic nonpharmacologic and pharmacologic therapeutic approaches to hyperuricemia. Arthritis Care Res (Hoboken). 2012;64:1431–1446 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Richette P, Doherty M, Pascual E, Barskova V, Becce F, Castaneda-Sanabria J, Coyfish M, Guillo S, Jansen TL, Janssens H, Liote F, Mallen C, Nuki G, Perez-Ruiz F, Pimentao J, Punzi L, Pywell T, So A, Tausche AK, Uhlig T, Zavada J, Zhang W, Tubach F, Bardin T. 2016 updated eular evidence-based recommendations for the management of gout. Ann Rheum Dis. 2017;76:29–42 [DOI] [PubMed] [Google Scholar]
- 30.Stroke−-1989. Recommendations on stroke prevention, diagnosis, and therapy. Report of the who task force on stroke and other cerebrovascular disorders. Stroke. 1989;20:1407–1431 [DOI] [PubMed] [Google Scholar]
- 31.Adams HP Jr., Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh EE 3rd. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. Toast. Trial of org 10172 in acute stroke treatment. Stroke. 1993;24:35–41 [DOI] [PubMed] [Google Scholar]
- 32.Carey RM, Calhoun DA, Bakris GL, Brook RD, Daugherty SL, Dennison-Himmelfarb CR, Egan BM, Flack JM, Gidding SS, Judd E. Resistant hypertension: Detection, evaluation, and management: A scientific statement from the american heart association. Hypertension. 2018;72:e53–e90 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Egan BM, Zhao Y, Axon RN, Brzezinski WA, Ferdinand KC. Uncontrolled and apparent treatment resistant hypertension in the united states, 1988 to 2008. Circulation. 2011;124:1046–1058 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Irvin MR, Booth III JN, Shimbo D, Lackland DT, Oparil S, Howard G, Safford MM, Muntner P, Calhoun DA. Apparent treatment-resistant hypertension and risk for stroke, coronary heart disease, and all-cause mortality. Journal of the American Society of Hypertension. 2014;8:405–413 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Howard VJ. Reasons underlying racial differences in stroke incidence and mortality. Stroke. 2013;44:S126–128 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Staessen J The determinants and prognostic significance of serum uric acid in elderly patients of the european working party on high blood pressure in the elderly trial. Am J Med. 1991;90:50S–54S [PubMed] [Google Scholar]
- 37.Judd CM, Kenny DA. Process analysis: Estimating mediation in treatment evaluations. Evaluation review. 1981;5:602–619 [Google Scholar]
- 38.Lee H, Herbert RD, McAuley JH. Mediation analysis. JAMA. 2019;321:697–698 [DOI] [PubMed] [Google Scholar]
- 39.VanderWeele T Explanation in causal inference: Methods for mediation and interaction. Oxford University Press; 2015. [Google Scholar]
- 40.VanderWeele TJ. Mediation analysis: A practitioner’s guide. Annu Rev Public Health. 2016;37:17–32 [DOI] [PubMed] [Google Scholar]
- 41.MacKinnon DP, Dwyer JH . Estimating mediated effects in prevention studies. Evaluation review. 1993;17:144–158 [Google Scholar]
- 42.MacKinnon DP, Fairchild AJ, Fritz MS. Mediation analysis. Annu Rev Psychol. 2007;58:593–614 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Jiang Z, VanderWeele TJ. When is the difference method conservative for assessing mediation? Am J Epidemiol. 2015;182:105–108 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Feig DI. Hyperuricemia and hypertension. Adv Chronic Kidney Dis. 2012;19:377–385 [DOI] [PubMed] [Google Scholar]
- 45.Grayson PC, Kim SY, LaValley M, Choi HK. Hyperuricemia and incident hypertension: A systematic review and meta-analysis. Arthritis Care Res (Hoboken). 2011;63:102–110 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Juraschek SP, Kovell LC, Miller ER, Gelber AC. Dose-response association of uncontrolled blood pressure and cardiovascular disease risk factors with hyperuricemia and gout. PloS one. 2013;8:e56546. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Muntner P, Davis BR, Cushman WC, Bangalore S, Calhoun DA, Pressel SL, Black HR, Kostis JB, Probstfield JL, Whelton PK, Rahman M, Group ACR. Treatment-resistant hypertension and the incidence of cardiovascular disease and end-stage renal disease: Results from the antihypertensive and lipid-lowering treatment to prevent heart attack trial (allhat). Hypertension. 2014;64:1012–1021 [DOI] [PubMed] [Google Scholar]
- 48.Abramson JL, Jurkovitz CT, Vaccarino V, Weintraub WS, McClellan W. Chronic kidney disease, anemia, and incident stroke in a middle-aged, community-based population: The aric study. Kidney Int. 2003;64:610–615 [DOI] [PubMed] [Google Scholar]
- 49.Weiner DE, Tighiouart H, Elsayed EF, Griffith JL, Salem DN, Levey AS. Uric acid and incident kidney disease in the community. Journal of the American Society of Nephrology. 2008;19:1204–1211 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Nakagawa T, Mazzali M, Kang D-H, Kanellis J, Watanabe S, Sanchez-Lozada LG, Rodriguez-Iturbe B, Herrera-Acosta J, Johnson RJ. Hyperuricemia causes glomerular hypertrophy in the rat. American journal of nephrology. 2003;23:2–7 [DOI] [PubMed] [Google Scholar]
- 51.Jordan DM, Choi HK, Verbanck M, Topless R, Won H-H, Nadkarni G, Merriman TR, Do R. No causal effects of serum urate levels on the risk of chronic kidney disease: A mendelian randomization study. PLoS medicine. 2019;16:e1002725. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Vaziri N, Freel RW, Hatch M. Effect of chronic experimental renal insufficiency on urate metabolism. Journal of the American Society of Nephrology. 1995;6:1313–1317 [DOI] [PubMed] [Google Scholar]
- 53.Hutcheon JA, Chiolero A, Hanley JA. Random measurement error and regression dilution bias. Bmj. 2010;340:c2289. [DOI] [PubMed] [Google Scholar]
- 54.Chaudhary NS, Gaffo AL, Cushman M, Colantonio LD, Levitan EB, Judd SE, Singh J, Limdi NA, Bridges SL, Howard G, Reynolds RJ, Irvin MR. Abstract 188: Association between uric acid and stroke in the regards case-cohort study. Stroke. 2019;50:A188–A188 [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.