| |
High Uric Acid Levels Signal Hypertension in Blacks
|
| |
| |
By Neil Osterweil, Senior Associate Editor, MedPageToday.com
Reviewed by Zalman S. Agus, MD; Emeritus Professor at the University of
Pennsylvania School of Medicine.
November 06, 2006
WINSTON-SALEM, N.C., Nov. 6 -- High serum levels of uric acid are strongly associated with risk for hypertension, particularly among blacks, according to researchers here.
* Explain to patients who ask that black men and women with high uric acid levels tend to have higher rates of hypertension than white men and women, and that high blood pressure in blacks is harder to control. The reasons for the racial-ethnic differences are unclear, however.
They found the association in a biracial cohort of middle-age men and women who took part in the Atherosclerosis Risk in Communities study. Black men had about a 30% greater risk for hypertension if they had higher serum uric acid levels compared with white men, and black women had about a 15% greater risk than white women for hypertension and hyperuricemia.
"The novel angle of our study is that the association between uric acid and hypertension is much stronger in blacks, a group that disproportionately suffers from kidney disease, stroke and other complications of hypertension," reported Philip B. Mellen, M.D., of Wake-Forest here, and colleagues, in the online edition of Hypertension.
The findings suggest that a therapeutic strategy of lowering serum uric acid, now being explored in clinical trials, has the potential to reduce hypertension and the attendant risk of cardiovascular and cerebrovascular complications, said Dr. Mellen and colleagues.
Although African Americans have higher rates of hypertension and associated complications than whites, blacks have been under-represented in studies exploring whether the link between hyperuricemia and blood pressure, the authors noted.
In their study, 9,104 participants who were free of hypertension at baseline were evaluated for hypertension at three-year intervals over four examinations during the study.
The authors created adjusted Cox proportional hazards models to assess the risk of incident hypertension or progression of blood pressure category for each standard-deviation higher baseline serum uric acid.
The mean age of participants at baseline was 53.3 (range 45-64). The mean systolic blood pressure was 113.8 + 12.2 mm Hg, and the mean diastolic pressure was 70.2 + 8.6 mm Hg. The mean serum uric acid level was 5.7 + 1.4 mg/dL.
The investigators found that among the group as a whole, higher levels of uric acid were associated with greater risk for hypertension after adjustment for age, baseline blood pressure, body mass index, renal function, diabetes, and smoking. Each standard deviation higher was associated with a hazard ratio of 1.10 (95% confidence interval, 1.04-1.15).
In subgroup analyses, the hazard ratio for each higher standard deviation of serum uric acid in black men was 1.32 (95% CI, 1.14 to 1.54), and in black women was 1.16 (95% CI, 1.03 to 1.31). In contrast, there was virtually no association between serum uric acid in either white men (hazard ratio 1.01, 95% CI 0.94 to 1.09) or in white women (hazard ratio 1.04, 95% CI 0.96 to 1.11).
The authors saw a similar pattern when they modeled blood pressure progression, with an overall hazard ratio of 1.10 (95% CI, 1.05 to 1.14). For black men, the hazard ratio was 1.26 (95% CI, 1.11 to 1.42), and for black women it was 1.18 (95% CI, 1.06 to 1.31). In contrast, for white men the hazard ratio was 1.05 (0.99 to 1.11), and for white women it was 1.05 (1.00 to 1.12).
"The positive association between serum uric acid and incident hypertension was stronger in blacks and most pronounced among black men," the authors wrote. "This observation is significant considering the fact that blacks suffer disproportionately from hypertension and its associated target-organ damage. Hypertension control rates are lower among blacks, despite the fact that hypertension awareness and treatment is equal or higher in blacks compared with non-Hispanic whites."
The investigators acknowledged that although they adjusted their analyses for potential confounding factors, it's possible that the higher body mass indices, diabetes prevalence, and baseline blood pressure among blacks compared with whites could have also have been confounding factors.
Another limitation to the study pointed out by the authors was one imposed by the study exclusion criteria. "The exclusion of individuals who were hypertensive at baseline selected out individuals with hypertension onset at a younger age, and, thus, our findings may not be reflective of earlier-onset hypertension," they wrote.
"This is of particular concern given the fact that the majority of blacks (56%) in Atherosclerosis Risk in Communities were hypertensive at baseline, which could introduce selection bias," they added. "Similarly, the prevalence of baseline renal insufficiency, an exclusion criterion, was higher in blacks."
"Serum Uric Acid Predicts Incident Hypertension in a Biethnic Cohort. The Atherosclerosis Risk in Communities Study"
Philip B. Mellen*; Anthony J. Bleyer; Thomas P. Erlinger; Gregory W. Evans; F. Javier Nieto; Lynne E. Wagenknecht; Marion R. Wofford; and David M. Herrington
From the Sections of General Medicine (P.B.M.), Nephrology (A.J.B.), and Cardiology (D.M.H.), Department of Internal Medicine, and the Departments of Biostatistics (G.W.E.) and Epidemiology (L.E.W.), Division of Public Health Sciences, Wake Forest University, Winston-Salem, NC; the Department of Internal Medicine (T.P.E.), University of Texas Medical Branch, Austin; the Department of Population Health Sciences (F.J.N.), University of Wisconsin, Madison; and the Department of Medicine, Division of General Medicine and Hypertension (M.R.W.), University of Mississippi Medical Center, Jackson.
Author Discussion:
"..serum uric acid (SUA) was positively associated with incident hypertension over 9 years of follow-up, and this relationship was stronger in blacks. More research is warranted concerning the physiological and clinical consequences of hyperuricemia, especially in blacks..... In this biethnic cohort of middle-aged men and women, each SD increment of baseline SUA was associated with a 7% greater risk of incident hypertension over 9 years of followup, independent of age, baseline blood pressure, body mass, diabetes status, smoking status, and renal function.....
.... The positive association between SUA and incident hypertension was stronger in blacks and most pronounced among black men. This observation is significant considering the fact that blacks suffer disproportionately from hypertension and its associated target-organ damage.13,14 Hypertension control
rates are lower among blacks, despite the fact that hypertension awareness and treatment is equal or higher in blacks compared with non-Hispanic whites...
.... the prognostic significance of SUA with respect to both hypertension and its cardiovascular end points may vary by ethnicity....
... The relationship between SUA and incident hypertension was less pronounced in whites, but this is not inconsistent with earlier reports....
.... Although our analyses adjusted for important potential confounders, there remains the possibility that the observed relationship was secondary to residual confounding. The black participants had higher body mass indices, diabetes
prevalence, and baseline blood pressure, and it is possible that models accounting for these differences failed to capture important associated confounders. Similarly, adjustment for estimated GFR may fail to fully account for underlying preclinical renal impairment, which would be reflected by baseline SUA elevations and could precede elevated blood pressure. Of note, SUA has been strongly correlated with urinary microalbumin, a marker of early vascular damage and renal dysfunction, in prehypertensive subjects.34 In the absence
of baseline measures of urinary albumin, we are unable to assess the potential mediating or confounding role of preclinical proteinuria in the associations we observed...."
Abstract--Serum uric acid has been positively associated with incident hypertension, but previous studies have had limited ability to explore this relationship across sex and ethnic strata.
We sought to evaluate this association in a biethnic cohort of middle-aged men and women. Participants in the Atherosclerosis Risk in Communities (ARIC) study who were free of hypertension at baseline (N=9104) were evaluated for hypertension at 3-year intervals over 4 examinations.
Adjusted Cox proportional hazards models evaluated risk of incident hypertension or progression of blood category for each SD higher baseline serum uric acid.
At baseline, the mean age was 53.3 years (range: 45 to 64 years), with a mean (SD) systolic blood pressure of 113.8 (12.2) mm Hg, mean diastolic blood pressure of 70.2 (8.6) mm Hg, and mean serum uric acid of 5.7 (1.4).
Higher serum uric acid was associated with greater risk of hypertension in the overall cohort (hazard ratio for each SD of higher uric acid [95% CI]: 1.10 [1.04 to 1.15]) and in subgroup analyses (black men: 1.32 [1.14 to 1.54]; black women: 1.16 [1.03 to 1.31]; white men: 1.01 [0.94 to 1.09]; white women: 1.04 [0.96 to 1.11]), after adjustment for age, baseline blood pressure, body mass index, renal function, diabetes, and smoking.
The pattern was similar when modeling blood pressure progression (overall: 1.10 [1.05 to 1.14]; black men: 1.26 [1.11 to 1.42]; black women: 1.18 [1.06 to 1.31]; white men: 1.05 [0.99 to 1.11]; white women: 1.05 [1.00 to 1.12]).
In conclusion, serum uric acid was positively associated with incident hypertension over 9 years of follow-up, and this relationship was stronger in blacks than in whites. More research is warranted concerning the physiological and clinical consequences of hyperuricemia, especially in blacks.
Introduction
The positive association between serum uric acid (SUA) and hypertension was observed over a century ago,1 but in the past it has been unclear whether hyperuricemia played a causal role in arterial hypertension or if it was merely a
marker of an underlying pathophysiological process. Although elevated SUA levels have been predictive of hypertension in longitudinal studies,2-9 the relationship between uric acid and blood pressure is confounded by numerous factors, including age, diabetes, obesity, alcohol use, and sodium intake or volume status. These factors have rendered it difficult to establish a causal role for SUA in the development of hypertension on the basis of epidemiological investigations.
Recent findings in animal models have helped elucidate possible mechanisms whereby SUA may lead to hypertension. In animal models, experimental hyperuricemia causes hypertension through multiple mechanisms, including decreased NO synthesis, upregulated renin production, renal tubular injury,10 and an afferent arteriolopathy.11 Moreover, a causal role for hyperuricemia has been demonstrated in rat models for the metabolic syndrome.12 These results from basic science investigations have spurred a renewed interest in discerning a causal role for elevated SUA in primary hypertension.
One limitation of the epidemiological studies of SUA and hypertension has been an underrepresentation of blacks in prospective studies. Although blacks have a higher prevalence of hypertension13 and experience more associated target
organ damage than do whites,14,15 data on the relationship between SUA and hypertension in this group is limited. One prospective study in a large, biethnic cohort of young adults suggested that the relationship was stronger in blacks.16 Our goal was to evaluate the longitudinal relationship between SUA and incident hypertension in a large, biethnic cohort of middle-aged adults.
Study Population
The Atherosclerosis Risk in Communities (ARIC) study is a prospective
epidemiological cohort study of clinical and subclinical cardiovascular risk in 4 US communities (Washington County, Md, Minneapolis, Minn, Jackson, Miss, and Forsyth County, NC). The cohort is composed of 15 792 participants, ages 45 to 64 years at baseline, selected by probability sampling at the 4 field centers. Of
note, the Jackson field center recruited blacks exclusively, whereas the sample from Forsyth County was biethnic, and samples from the other centers were predominantly white. Response rates were 65% in Washington County, 67% in Minneapolis, 46% in Jackson, and 45% (blacks) and 71% (whites) in Forsyth County. The protocol was approved by the institutional review board of each field center, and all of the participants signed written informed consent. The baseline
examination (visit 1) took place in 1987-1989, with 3 follow-up exams at about 3-year intervals thereafter (visit 2: 1990-1992; visit 3: 1993-1995; visit 4: 1996-1998).
Participants were excluded from the current analyses for the following reasons: baseline hypertension (systolic blood pressure [SBP] >?=140 mm Hg, diastolic blood pressure [DBP] >/=90 mm Hg, or current use of antihypertensive medications for any reason, N=5495; blacks: N=2374 [56%]; whites: N=3121 [20%]), prevalent coronary heart disease (by history or ECG criteria, N=766; blacks: N 171 [4%]; whites: N=594 [5%]), history of stroke (N 286; blacks: N=91 [2%]; whites: N=195 [2%]), stage 4 or 5 chronic kidney disease (estimated glomerular filtration rate [GFR] <30 mL/min/ 1.73 m2,17 N=247; blacks: N 169 [4%]; whites: N=23 [ 1%]), current use of uricosuric medications (N=78; blacks: N=20 [0.5%]; whites: N=58 [0.5%]), self-identified ethnicity other than black or white (N=48), or missing values for important predictor variables (uric acid, N 204; SBP and DBP, N 68; body mass index [BMI], N=78). A total of 9104 participants were included in the primary analyses (black N=1687; white N=7417). Secondary analyses evaluating the baseline association between SUA and blood pressure included additional individuals not taking antihypertensive medications at baseline (N=10 430).
Results
At the baseline examination, the mean age of the cohort was 53.3 years (range: 44 to 65 years), with a mean (SD) SBP of 113.8 (12.2) mm Hg and DBP of 70.2 (8.6) mm Hg. There were significant differences among the race-sex subgroups
for all of the covariates evaluated (Table 1). When comparing covariates by uric acid levels (above or below the median) within race-sex subgroups, BMI, SBP, and DBP were consistently higher in those with higher uric acid, and estimated
GFR was consistently lower (Supplementary Tables I to IV available online at http://hyper.ahajournals.org). In crosssectional analyses of individuals not taking antihypertensive medications at baseline (N 10 430), SUA was positively associated with both SBP and DBP in the overall sample and in all of the race- gender subgroups except black women (Figure I, available online). This pattern was attenuated but essentially unchanged when the analyses were restricted to
individuals with SBP 140 mm Hg and DBP 90 mm Hg (data not shown).
In unadjusted analyses, SUA above the median was associated with incident hypertension in all of the race-sex subgroups (Figure 1). However, in multivariable modeling, the association between SUA and incident hypertension was largely attenuated by adjusting for baseline SBP and DBP (Model 2) in white men and women (Table 2). In models adjusting for baseline age, clinical center, SBP, DBP, BMI, diabetes status, smoking status, and estimated GFR, each SD
higher baseline SUA was associated with an increased hazard of incident hypertension for the overall cohort (hazard ratio: 1.07 [95% CI: 1.02 to 1.13]), black men (hazard ratio: 1.32 [95% CI: 1.14 to 1.54]), and black women (hazard ratio: 1.16 [95% CI: 1.03 to 1.31]). This trend was not significant for white men (hazard ratio: 1.01 [95% CI: 0.94 to 1.09]) or white women (hazard ratio: 1.04 [95% CI: 0.96 to 1.11]). These findings did not change after accounting for alcohol intake, sodium intake, and leisure time physical activity (model 5). The patterns were similar in analyses using different SUA thresholds (race-sex-specific median, fourth versus first quartile, or 7 mg/dL), modeling a 1-mg/dL
increment of baseline SUA, using follow-up measures of SUA, or restricted to individuals with estimated GFR 60 mL/min/1.73 m2 (Table V, available online). Additional models that included other medications (hormone replacement therapy, salicylates, and nonsteroidal anti-inflammatory medications), dietary factors (animal protein intake and vitamin C intake), alternative measures of renal function (serum creatinine or GFR estimated by Cockcroft-Gault), or alternative measures of obesity (waist circumference) did not alter these findings (data not shown).
Similarly, each SD greater than baseline SUA was associated with a significantly increased risk for blood pressure progression (increase in JNC 7 category 1) in the overall sample (hazard ratio: 1.12 [95% CI: 1.07 to 1.19]; P 0.0001). The
ethnic difference in event rates was less pronounced with this outcome (black men: 51%; black women: 55%; white men: 49%; white women: 50%), but the differences in the relationship between SUA and blood pressure among race-sex
subgroups were similar (Figure 2).
Finally, sensitivity analyses evaluated incident hypertension or blood pressure progression using baseline SBP and DBP estimated from repeated blood pressure measures over follow-up. In these models, greater SUA remained predictive of the primary and secondary outcomes, although statistical significance of this relationship was attenuated (Table 3). The pattern among race-sex groups was similar.
TABLE 1. Characteristics of ARIC Participants Free of Hypertension at Baseline Exam (N 9104)
Due to limited space here, the last column showing p-values was deleted but they ALL were <0.0001. Diabetews was more common among Black men & women. Smoking was more common among Black men. Uric acid was 5.7 mg/dL for the total cohort; 6.5 for Black men; 5.3 for Black women; 6.5 for whire men; anf 5.0 for White women.
Discussion
In this biethnic cohort of middle-aged men and women, each SD increment of baseline SUA was associated with a 7% greater risk of incident hypertension over 9 years of followup, independent of age, baseline blood pressure, body mass, diabetes status, smoking status, and renal function. This study extends findings from previous studies on the longitudinal association between SUA and hypertension to middle-aged blacks, because the ARIC cohort had nearly twice as many cases of incident hypertension among blacks (N 626) as any previous prospective study.2,8,16
The positive association between SUA and incident hypertension was stronger in blacks and most pronounced among black men. This observation is significant considering the fact that blacks suffer disproportionately from hypertension and
its associated target-organ damage.13,14 Hypertension control rates are lower among blacks, despite the fact that hypertension awareness and treatment is equal or higher in blacks compared with non-Hispanic whites.13,27 Our results are consistent with findings from the Coronary Artery Risk Development in (Young) Adults study (CARDIA), in which the association between SUA and incident hypertension was stronger in blacks than in whites.16 This is also consistent with earlier reports from the ARIC cohort with respect to SUA and cardiovascular events. A previous study found an association between SUA and incident heart disease in black women but not in other women.28 Similarly, SUA in the highest quartile was associated with a 79% increased risk of ischemic stroke in blacks compared with a 27% increased risk among whites, although this interaction was not statistically significant.29 Thus, the prognostic significance of SUA with respect to both hypertension and its cardiovascular end points may vary by ethnicity.
The association we observed was weaker in black females than in black men. This could be because of the overall high rate of incident hypertension and a diminution of a true association because of competing risks for hypertension.
Black women had the highest mean baseline BMI of the groups examined, and models accounting for BMI resulted in the greatest attenuation of the association. This is consistent with an earlier study that found that the association between
SUA and incident hypertension was diluted in obese individuals. 5 The relationship between SUA and incident hypertension was less pronounced in whites, but this is not inconsistent with earlier reports. As Sundstrom et al noted,9 the observed strength of the association between SUA and hypertension has been less pronounced in older cohorts, and the ARIC study is the oldest cohort in which this association has been explored. One potential explanation, consistent with the model proposed by Johnson et al,30 is that the pathogenic role for uric acid is greater early in the natural history of hypertension. In the ARIC study, 35% of the cohort was hypertensive at baseline, and this study did not assess the predictive role of SUA for those individuals. This study was unable to assess the relative strength of the association across subgroups as they approach middle age. However, taken in conjunction with the earlier CARDIA findings, the strength of the association in blacks compared with whites seems to be consistent in different age groups.
Exploring ethnic variation in disease pathogenesis is challenging, because self-identified race is a complex construct reflecting biological, psychosocial, economic, and environmental factors.31 Thus, we are reluctant to ascribe the differences we observed to underlying "genetic" factors, because self-identified race is a poor proxy for genetic ancestry,32 and other factors correlated to this construct may be equally or more important in disease pathogenesis,33 albeit challenging to measure. Nonetheless, our study complements other investigations that have not been able to prospectively address the question of SUA and incident hypertension in a large number of blacks.
In addition to the biethnic cohort, this study has several significant strengths. ARIC is the largest epidemiological cohort in which the longitudinal relationship between SUA and hypertension has been addressed with fully adjusted models. As noted previously, this middle-aged cohort complements earlier investigations in younger populations, potentially reflecting a different stage in the pathogenesis of hypertension. Furthermore, the serial examinations at which to detect hypertension offer greater power to evaluate hypertension incidence over time.
Although our analyses adjusted for important potential confounders, there remains the possibility that the observed relationship was secondary to residual confounding. The black participants had higher body mass indices, diabetes
prevalence, and baseline blood pressure, and it is possible that models accounting for these differences failed to capture important associated confounders. Similarly, adjustment for estimated GFR may fail to fully account for underlying preclinical renal impairment, which would be reflected by baseline SUA elevations and could precede elevated blood pressure. Of note, SUA has been strongly correlated with urinary microalbumin, a marker of early vascular damage and renal dysfunction, in prehypertensive subjects.34 In the absence
of baseline measures of urinary albumin, we are unable to assess the potential mediating or confounding role of preclinical proteinuria in the associations we observed.
There are several limitations to this study. The exclusion of individuals who were hypertensive at baseline selected out individuals with hypertension onset at a younger age, and, thus, our findings may not be reflective of earlier-onset hypertension.
This is of particular concern given the fact that the majority of blacks (56%) in ARIC were hypertensive at baseline, which could introduce selection bias. Similarly, the prevalence of baseline renal insufficiency, an exclusion criterion, was higher in blacks. Thus, the differences among ethnic groups should be
viewed with caution. However, as noted above, a similar pattern was observed in the younger CARDIA cohort. Another potential limitation is misclassification of hypertensive individuals as normotensive at baseline. However, the pattern of our results was consistent when modeling the alternative outcome of blood pressure progression. Furthermore, variability in measures of blood pressure, SUA, or their covariance is a source of measurement error that could bias results.35 Although the data required to fully account for this were not available, the observed measurement variability was more pronounced for blood pressure,19 and our findings were consistent in sensitivity analyses designed to
minimize baseline measurement error in SBP and DBP.
An additional consideration is the fact that most of the black participants came from 1 field center (Jackson: 86%), limiting our ability to distinguish the effect of ethnicity from field center effects. The exclusion of the field center covariate
from these analyses did not alter our findings, suggesting that field center did not completely account for the observed heterogenous associations across race-sex subgroups.
Perspectives
SUA was positively associated with incident hypertension over 9 years of follow-up, and this relationship was stronger in blacks. More research is warranted concerning the physiological and clinical consequences of hyperuricemia, especially in blacks.
Figure 2. Hazard ratio for incident blood pressure progression for each SD increase in baseline serum uric acid. Blood pressure progression increase of 1 JNC 7 category; Cox models adjusted for age, field center, SBP, DBP, BMI, diabetes status, smoking status, and estimated GFR (overall model also adjusted for race and sex). Note that x axis scale is log2. AAM indicates black men; AAW, black women; WM, white men; WW, white women.
|
|
| |
| |
|
|
|
