Prehypertension, Racial Prevalence and Association with Risk Factors: Analysis of The REasons for Geographic And Racial Differences in Stroke (REGARDS) Study

Abstract

Background

There are few available data on the epidemiology of prehypertension

Objective

To determine racial, clinical, and demographic differences in the prevalence of prehypertension and its cross-sectional association with vascular risk factors.

Methods

Cross-sectional analysis of 5553 prehypertensives, 20351 hypertensive’s, and 4246 non hypertensive participants (age ≥45), from a population-based national cohort study (REGARDS total population 30239, of whom 30150 had adequate blood pressure measurements) enrolled from January 2003-October 2007 with over-sampling from the southeastern Stroke Belt, and black individuals. Baseline data were collected using a combination of telephone interview and in-home evaluation. Prehypertension was defined according to JNC 7 guidelines.

Results

The prevalence of pre-hypertension was associated with age and black race (62.9% in blacks compared to 54.1% in whites). A higher prevalence of pre-hypertension was observed in obese individuals, self-reported heart disease; and, those with elevated hsCRP, diabetes, and microalbuminuria compared to those without these factors. Heavy alcohol consumption in white participants was associated with increased odds of pre-hypertension (OR = 1.32) but was even greater in black participants (OR=2.27).

Conclusion

The prevalence of prehypertension increased by age and African-American race. In addition, a higher prevalence of pre-hypertension was observed with elevated hsCRP, diabetes, microalbuminuria, and those with heavy alcohol consumption compared to those without these factors.

Introduction

The Seventh Report of the Joint national Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) defined prehypertension (preHTN) as a blood pressure of 120-139 mmHg systolic and/or 80-89 mmHg diastolic, which in the past was called transient hypertension, borderline hypertension or high normal blood pressure. Although the terminology has changed, there is agreement that preHTN is a precursor of hypertension and that there is the potential for its association with an excess morbidity and death from cardiovascular disease (CVD).¹ One study suggested that if preHTN were eliminated, that potentially 47% of all heart attacks might be prevented.² There is also controversy concerning the potential cardiovascular risk associated with preHTN. This controversy is mostly centered on the wide variation in cardiovascular risk among this group; and, is dependant upon co-existing risk factors as has been suggested by Gu et al³ who found, after adjustment for risk factors, no association of preHTN with CVD mortality. Data from the 1999 and 2000 National Health and Nutrition Examination Survey (NHANES III) estimated that the prevalence of preHTN among adults in the United States was approximately 31 percent. ⁴ The prevalence of preHTN in NHANES III was higher among men than women (39 and 23 percent, respectively). In WHI, prevalence by race-ethnic groups were provided and ranged from 32-40%. WHI also reported an increased risk of myocardial infarction (MI), stroke, CVD death and other outcomes in the prehypertensive groups over 7.7 years of follow-up. ⁵ The Trial of Preventing Hypertension (TROPHY) provided evidence to support that the treatment of preHTN in reducing the development of hypertension. ⁶

Few data describe potential racial differences in the prevalence of preHTN, or racial differences in its association with risk factors and other clinical variables. Therefore, we addressed these questions in the Reasons for Geographic And Racial Differences in Stroke (REGARDS) cohort, by reporting the prevalence and risk factor profiles of participants with the diagnosis of preHTN. These analyses extend previous research by capitalizing on the large, geographically dispersed, race- and gender-balanced sample of the REGARDS cohort.

METHODS

Study Population

REGARDS is a national cohort of community dwelling individuals age 45 years or older, recruited with approximately equal representation of white and black, and male and female participants (n=30239). Twenty-one percent (20% was the original goal) of the sample was randomly selected from the “buckle” of the Stroke Belt (coastal plain region of North Carolina, South Carolina, and Georgia), 35% (30% goal) from the Stroke Belt states (remainder of North Carolina, South Carolina, and Georgia plus Alabama, Mississippi, Tennessee, Arkansas, and Louisiana), and the remaining 44% (50% was the goal) from the other 40 contiguous states. Individuals were identified from commercially available lists of residents, and recruited using an initial mailing followed by telephone contact. Defined according to standards recommended by Morton et al., the telephone response rate was 33% and the cooperation rate was 49% (similar to other epidemiological studies, the Multi-Ethnic Study of Atherosclerosis, MESA, for example with a 39.8% participation rate among those contacted and to whom the study was explained.). ⁷ In this cross-sectional analysis, 5553 prehypertensives, 20351 hypertensive’s, and 4246 non hypertensive participants were analyzed. (Figure 1)

Figure 1.

The Exclusionary Cascade representing the analysis population

Trained interviewers obtained demographic information and medical history using computer-assisted telephone interview (CATI). Consent was obtained verbally by telephone and subsequently in writing during a follow-up in-home visit 2-4 weeks after the telephone interview. Height, weight and blood pressure measurements were obtained from the in-person home visit. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were defined as the average of two measurements taken by a trained technician using a standard protocol and regularly tested aneroid sphygmomanometer, measured in the fasting state (except that there was no prohibition of coffee) after the participant was seated for 5 minutes. As reported in JNC 7⁸ “At least two measurements should be made and the average recorded,” and this was the protocol used by REGARDS. Blood pressure quality control was monitored by central examination of digit preference and retraining of technicians took place as necessary. Participants were followed by telephone at six-month intervals for cognitive assessment and surveillance of medical events including potential stroke and myocardial infarction events. The study methods were reviewed and approved by all involved Institutional Review Boards. Additional methodological details are provided elsewhere. ⁹

Study recruitment ended in October of 2007 with a final cohort size of 30239. There was a total of 5553 with prehypertension, 522 of whom had a DBP of 80-89 but with a SBP of <120 mmHg; 2893 with DBP < 80 but with SBP of 120-139 mmHg; and, 2138 with DBP of 80-89 and SBP 120-139 mmHg.

Statistical Analysis

As part of our primary hypothesis, we sought to examine race as an a priori effect modifier. Categorical covariates of interest were race, geographic location (stroke belt, stoke buckle, or other), urban/rural location, income, education, sex, smoking, physical activity, alcohol use, prior history of CVD (self-reported, MI, vascular disease, or vascular intervention or EKG evidence of MI), self-reported stroke/TIA, and diabetes (defined as fasting glucose level >126 ml/dL, non-fasting glucose >200 ml/dL, or self-reported medication use for glucose control). Alcohol use was categorized as number of drinks per day (none, moderate defined as 1-7 drinks/week for women and 1-14 for men, heavy defined as >7 drinks/day for women and >14 for men). Physical activity level was defined by response to the question “How many times per week do you engage in intense physical activity, enough to work up a sweat?” categorized as none, 1-3 times, or 4 or more times per week. Body mass index (BMI) was determined as a function of measured height and weight. Urban/rural status was defined by residence on the basis of the percentage of the census track residing inside of urban areas/clusters; the status was rural if ≤25% urban, mixed between 25-75% urban, and urban if ≥75% urban. Smoking was categorized as never, current or past. Pearson’s chi-squared was used to obtain a p value to compare frequency distributions of pre-hypertension. We examined high sensitivity C reactive protein (hsCRP in mg/L), BMI, age, estimated GFR, and urinary albumin-creatinine ratio (ACR) as both continuous and categorical covariates. Microalbuminuria was defined as ACR greater than 17 mg/G for men and 25 mg/G for women. ¹⁰ We used cut points of 1.0 and 3.0 to create a three category variable for hsCRP and of 3.0 to create the dichotomous hsCRP variable.

Logistic regression (PROC LOGISTIC in SAS 9.1 Cary, NC) was used to calculate odds ratios for multivariable models. We first assessed the interaction between race and region and their association with prehypertension. This interaction p value was 0.0015. Therefore, we presented all results stratified by race. The covariates described above were tested as potential confounders in all models. We used backwards elimination to determine significant correlates of preHTN in the models.

Results

The total REGARDS cohort consists of 30239 individuals, of whom 30150 had blood pressure measures. Individuals with elevated SBP (6314) or DBP (602) or individuals using antihypertensive medication, both observed (12,818) and self-reported (617), were excluded. This left an analysis population of 9799 (Figure 1), of whom 5553 (51%,-17% of the 30150) individuals were prehypertensive (Table 1).

Table 1. Prevalence of pre-hypertension by demographic and socioeconomic characteristics.

	n	Blacks (n=2817)  % of black subpopulation with  pre-hypertension	n	Whites (n=6982)  % of white subpopulation with  pre-hypertension
Total with pre-hypertension	1773	62.9	3780	54.1
Classification of pre-hypertension
SBP<120 mmHg and DBP≥80 mmHg	162		360
DBP < 80 mmHg and SBP≥120 mmHg	836	2057
SBP≥120 mmHg and DBP≥80 mmHg	775		1363
Gender
Male	986	58.9	1790	47.4
Female	787	69.1	1990	62.0
Region
Stroke Belt	520	58.5	1328	55.6
Stroke Buckle	309	63.3	780	51.3
Non Belt	944	65.8	1672	54.4
Age group
45–54	383	55.7	572	43.3
55–64	773	63.9	1610	52.4
65–74	459	66.9	1116	60.2
75+	158	69.6	482	65.8
Rural/urban (missing 3)
<=25% Urban	186	62.2	932	54.2
Mixed 25-75% Urban	88	61.9	491	64.0
Urban	1499	63.2	2357	54.1
Income
$20K	347	65.2	363	58.6
$20K–$34K	450	61.9	755	58.4
$35K–74K	536	62.3	1308	55.7
$75+	228	61.2	897	47.8
Refused to report	212	65.8	457	54.1
Years of education (missing 6)
<High school	253	67.8	220	65.5
High school	468	63.8	851	51.0
Some college	504	62.2	1031	56.0
College+	548	61.2	1678	50.7

Prehypertension defined as systolic blood pressure between 120 and 139 mmHg or diastolic blood pressure between 80 and 89 mmHg. Hypertensives were excluded from analysis.

Percentages refer to racial specific prevalences of pre-hypertension within each stratum of covariates. Example: 58.9% of black males have pre-hypertension compared to 47.4% of white males.

The prevalence (Table 1) of preHTN was greater with age in both black and white participants. The prevalence of preHTN differed by region and ranged between 51% for whites in the stroke buckle to 66% for black individuals living in the stroke belt (p=0.02). The prevalence of preHTN was higher in black participants across all age and gender strata, 58.9% of black males had preHTN while 47.4% of white males had preHTN (p<.001). We used multivariable logistic regression to further examine the racial disparity in the odds of preHTN (Figure 2). Except for low income, obesity, and diabetes, blacks had a greater odds of preHTN compared to whites in regards to gender, region of the country, age, smoking status, and years of education.

Figure 2.

A Forest plot of the odds ratios of prehypertension comparing black to white participants in the REGARD study. Models adjust for age, gender, income, and education.

We used multivariable logistic regression to describe the association of risk factors for preHTN stratified by race (Table 2). Participants who reported heavy drinking were more likely to have prehypertension than non drinkers. The magnitude of effect was greater in black participants than white participants (OR=2.27 and 1.32 respectively). In addition, moderate alcohol consumption in white participants was associated with a reduction in the odds of pre-hypertension (OR = 0.95) and was 1.08 in black participants (NS). In white participants, behaviors such as smoking and exercise were not associated with pre-hypertension. In contrast, for black participants we observed a marginally significant (p= 0.02) unexpected association where higher levels of exercise were associated with a higher odds of pre-hypertension, a finding that we feel is likely a type 1 error (i.e., spurious finding perhaps associated with multiple testing), or potentially inaccurate self-reporting of exercise frequency or a result of untested interactions. White participants with higher BMIs were at slightly higher odds of having preHTN (OR=1.71 vs 1.55 in black participants, p<0.001). Finally, while there was a significant (p < 0.05) positive association of pre-HTN with CVD and CVD risk factors (elevated hsCRP, stroke, heart disease, diabetes, cognitive dysfunction, microalbuminuria, and chronic kidney disease), there were no apparent racial differences in their cross-sectional association with pre-HTN.

Table 2. Association of pre-hypertension with demographic, socioeconomic and lifestyle factors stratified by race.

	Whites				Blacks
	Model 1w		Model 2w		Model1b		Model 2b
	OR	p†	OR	p†	OR	p†	OR	p†
Region (Non belt is reference)
Belt	1.05	(0.93, 1.18)	1.06	(0.95, 1.19)	0.78	(0.65, 0.95)	0.77	(0.65, 0.94)
Buckle	0.89	(0.78, 1.02)	0.91	(0.79, 1.03)	0.91	(0.72, 1.14)	0.90	(0.72, 1.13)
Age (10 yrs)	1.43	(1.35, 1.53)	1.48	(1.40, 1.57)	1.30	(1.18, 1.44)	1.32	(1.21, 1.45)
Male	1.75	(1.56, 1.93)	1.73	(1.56, 1.92)	1.78	(1.49, 2.13)	1.86	(1.56, 2.21)
BMI (5 kg/m2)	1.74	(1.56, 1.93)	1.73	(1.65, 1.85)	1.57	(1.45, 1.71)	1.58	(1.45, 1.71)
Income
$20K	1.26	(1.01, 1.56)			0.95	(0.69, 1.31)
$20K–$34K	1.24	(1.04, 1.46)			1.05	(0.78, 1.42)
$35K–74K	1.16	(1.02, 1.33)			1.01	(0.77, 1.33)
Refused	1.19	(0.98, 1.42)			1.17	(0.82, 1.67)
Years of education (missing 6)
<High school	1.65	(1.26, 2.15)	1.67	(1.30, 2.15)	1.09	(0.81, 1.46)
High school	1.22	(1.06, 1.41)	1.27	(1.11, 1.45)	1.05	(0.83, 1.33)
Some college	1.16	(1.02, 1.32)	1.21	(1.06, 1.36)	1.01	(0.81, 1.26)
Alcohol Consumption
Moderate (0-7 women, 0-14 men)	0.96	(0.86, 1.08)	0.95	(0.85, 1.06)	1.07	(0.88, 1.30)	1.08	(0.91, 1.30)
Heavy (7+ women 14+ men)	1.35	(1.06,1.72)	1.32	(1.04,1.66)	2.20	(1.25, 3.80)	2.27	(1.32, 3.94)
Exercise (4+ times per week is reference)
None	0.94	(0.82, 1.07)			0.74	(0.60, 0.92)	0.74	(0.61, 0.92)
1 to 3 times	1.04	(0.93, 1.17)			0.88	(0.72, 1.07)	0.87	(0.72, 1.06)
Smoke (Never is reference)
Current	0.94	(0.80, 1.10)			1.09	(0.86, 1.37)
Past	1.06	(0.95, 1.19)			1.21	(0.99, 1.46)
Pre-existing diabetes	0.87	(0.72, 1.05)			1.02	(0.80, 1.30)
eGFR<60	1.26	(1.02, 1.55)			1.11	(0.81, 1.53)

^†Wald Chi square

Model 1w is the full model and controls for region of residence, age, gender, BMI as a continuous variable, income, education, exercise, smoking status, and alcohol consumption

Model 2w is the most parsimonious model and controls for region of residence, age, gender, BMI as a continuous variable, education, and alcohol consumption. Income, exercise, and smoking status were not statistically significant covariates.

Model 1b is the full model and controls for region of residence, age, gender, BMI as a continuous variable, income, education, exercise, smoking status, and alcohol consumption.

Model 2b is the most parsimonious model and controls for region of residence, age, gender, BMI as a continuous variable, education, and alcohol consumption. Income, exercise, smoking status, and education were not statistically significant covariates.

DISCUSSION

There has been controversy surrounding the classification of BP termed preHTN, mostly centered upon the wide variation in risk among this group. While accumulating literature suggests that preHTN is associated with adverse clinical outcomes, the epidemiology of preHTN has been less well described. An analysis of a large cohort of men and women from the REGARDS Study helps to fill this gap in knowledge. Of 9799 non-hypertensive individuals, 51% (17% of the entire REGARDS cohort) were defined as demonstrating preHTN (including 522 of whom had DBP>80 but SBP<120; 2893 had DBP<80 but SBP≥120; and, 2138 with a DBP of 80-89 and a SBP of 120-139 mmHg). The prevalence of pre-hypertension has been reported to be about 30% of the adult US population.¹ The overall prevalence found in our study was 17%, however, excluding subjects with hypertension it was 51%. We observed that the prevalence of pre-hypertension was higher by age and black race (62.9% compared to 54.1% in whites), and a higher prevalence of pre-hypertension was observed in obese individuals, self-reported heart disease; and, those with elevated hsCRP, diabetes, and microalbuminuria compared to those without these factors. Heavy alcohol consumption in white participants was associated with an increased odds of pre-hypertension (OR = 1.32) but was even greater in black participants (OR=2.27 p<.03). Some of these observations are in concert with prior reported studies. For example, in a retrospective analysis from the San Antonio Heart Study, it was found that at baseline, prehypertension was associated with male gender, higher BMI, and measures of impaired glucose intolerance. In terms of race/ethnic variables, NHANES III found no difference in the prevalence between Non-Hispanic whites, Non-Hispanic blacks, Mexican Americans or other (all had a prevalence of about 31%), compared to our analytic sample where the African American prevalence of preHTN was higher than whites.

The lifetime risk of developing hypertension approaches 90%, however, the etiology of hypertension remains elusive for the vast majority.¹¹ Howard et al reported that the prevalence of hypertension was related to the duration and age at exposure to the stroke belt, however, neither they nor we analyzed the prevalence of prehypertension by those variables.¹² The complications of long standing hypertension are well known and relate to arterial disease with clinical manifestations in the heart, brain and kidney. However, the paradigm of elevated blood pressure resulting in vasculopathy continues to be challenged. Rather than hypertension resulting in altered vascular structure and function, it appears that changes in vascular integrity (structure) precede, and may be causal, in the development of elevated blood pressure with resultant hypertension that ultimately leads to clinical events. ¹³ Alterations in cardiovascular structure and function that have been shown to precede the finding of elevated blood pressure include the occurrence of left ventricular hypertrophy in children and young adults of hypertensive parents, ¹⁴ although one cannot rule out the role of blood pressure in causing cardiovascular remodeling in studies where only resting, occasional measurements of blood pressure are made; diastolic filling abnormalities in normotensive individuals predisposed to hypertension; ¹⁵ endothelial dysfunction as a precursor to the finding of hypertension; ¹⁶ and increased arterial stiffness in normotensive subjects predisposed to develop hypertension. ¹⁷ Recently it has been reported that in confirmed prehypertensive subjects, intimal-medial thickness is increased in the common carotid artery when compared to subjects who remain normotensive. ¹⁸ It was also found in an analysis of subjects from The Bogalusa Heart Study that preHTN prevalence was higher among men than women and among black compared to white participants; and, that compared to normotensives, those with preHTN had higher left ventricular (LV) mass index, LV diameter, and carotid artery intima-media thickness. ¹⁹ If changes in vascular integrity precede the development of hypertension, preHTN could fit well into the natural history of this evolution. In contrast, Tomiyama et al pointed out that baseline preHTN BP was not an adequate predictor of the development of hypertension, although as they hypothesized that the rather short-term follow-up was a possible explanation for the weak association. ²⁰

Limitations

The REGARDS study is subject to several limitations. Some risk factors were based on self report (although this is common to many epidemiologic studies), and individuals without telephones were necessarily excluded from selection into the study population. These excluded individuals may be of lower socioeconomic status and, therefore, may have different BP and risk factor profiles than those included in this analysis. Finally, participant BP was measured on a single occasion albeit by standard protocol.

In conclusion, we have characterized the cross-sectional associations of prehypertension with a number of lifestyle, geographic, and demographic factors in order to better understand the epidemiology of this potential precursor of clinical cardiovascular disease. PreHTN is common (51% of the non-hypertensive group), and is more likely to be present in those at risk for or with manifest vascular disease. We found that a higher prevalence of pre-hypertension was observed with elevated hsCRP, diabetes, heavy alcohol consumption, and microalbuminuria compared to those without these factors. Longitudinal studies are necessary to determine if BP in the prehypertensive range alone (i.e. in the absence of other risk factors) is a sufficient risk predictor for future CVD.