Factors Associated With the Prevalence of Hypertension in the Southeastern United States

Abstract

Background

Lifestyle and socioeconomic status have been implicated in the prevalence of hypertension; thus, we evaluated factors associated with hypertension in a cohort of blacks and whites with similar socioeconomic status characteristics.

Methods and Results

We evaluated the prevalence and factors associated with self-reported hypertension (SR-HTN) and ascertained hypertension (A-HTN) among 69 211 participants in the Southern Community Cohort Study. Multivariable logistic regression models were used to estimate the odds ratios (ORs) and 95% confidence intervals (CIs) for factors associated with hypertension. The prevalence of SR-HTN was 57% overall. Body mass index was associated with SR-HTN in all race-sex groups, with the OR rising to 4.03 (95% CI, 3.74–4.33) for morbidly obese participants (body mass index, >40 kg/m²). Blacks were more likely to have SR-HTN than whites (OR, 1.84; 95% CI, 1.75–1.93), and the association with black race was more pronounced among women (OR, 2.08; 95% CI, 1.95–2.21) than men (OR, 1.47; 95% CI, 1.36–1.60). Similar findings were noted in the analysis of A-HTN. Among those with SR-HTN and A-HTN who reported use of an antihypertensive agent, 94% were on at least one of the major classes of antihypertensive agents, but only 44% were on ≥2 classes and only 29% were on a diuretic. The odds of both uncontrolled hypertension (SR-HTN and A-HTN) and unreported hypertension (no SR-HTN and A-HTN) were twice as high among blacks as whites (OR, 2.13; 95% CI, 1.68–2.69; and OR, 1.99; 95% CI, 1.59–2.48, respectively).

Conclusions

Despite socioeconomic status similarities, we observed suboptimal use of antihypertensives in this cohort and racial differences in the prevalence of uncontrolled and unreported hypertension, which merit further investigation.

Hypertension is a well-established risk factor for all-cause and cardiovascular disease (CVD) mortality, accounting for an estimated 7.5 million deaths per year or 13.5% of total annual deaths worldwide.^1,2 The public health significance of hypertension provided the impetus for the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7).³ Moreover, in the report of the recent landmark Global Burden of Disease Study, hypertension now tops the list of risk factors for death and disability worldwide,⁴ further highlighting the need to better characterize the risk factors for hypertension, specifically unreported and uncontrolled hypertension. Substantial evidence indicates that blacks have a higher prevalence of hypertension than whites^5–7 and that severe hypertension and hypertension complicated by target organ damage are more common in blacks and lead to higher rates of CVD and adverse events.⁸ Black and white hypertensive patients have also been noted to have differential response to therapeutic strategies, comprising dietary interventions⁹ and antihypertensive drugs,^10–12 including modulators of the renin-aldosterone-angiotensin system.^13–17

Recent efforts to determine reasons for racial disparities in the prevalence of hypertension and related adverse sequelae have included assessments of the impact of socioeconomic status (SES), including social environment.^18–25 In this context, a unique opportunity to enhance our understanding is presented by the Southern Community Cohort Study (SCCS), a prospective epidemiological cohort study examining racial differences in cancer and other chronic diseases such as CVD, diabetes mellitus, and hypertension in a population of blacks and whites with similar and well-documented SES indexes (eg, income, education, employment, and health insurance status).^26–28 We have evaluated factors associated with the prevalence of hypertension among 69 211 blacks and whites with self-reported data and in a subset of 16 932 individuals in whom the presence of hypertension was ascertained using actual blood pressure (BP) measurements. This is the largest evaluation of factors associated with prevalent self-reported hypertension (SR-HTN) and ascertained hypertension (A-HTN) in a population of blacks and whites residing in the southeastern United States, a region known for its high CVD risk factor prevalence and attendant morbidity and mortality.

Methods

Study Population

The SCCS is an ongoing epidemiological study that enrolled adult participants (age, 40–79 years) resident within 12 states in southeastern United States from March 2002 until September 2009. This population-based study is a prospective, combined cross-sectional and longitudinal observational study. The cross-sectional analyses, which pertain to baseline findings, include the present study. Approximately 86% of participants were enrolled through urban and rural community health centers, which provide primary health care for medically underserved populations, and the remaining 14% were enrolled through mail-based general population sampling.²⁸ The present report focuses on 69 211 black and white community health center participants, which ensured similarities in socioeconomic characteristics and access to health care at cohort entry regardless of race. The reason for the restriction to black and white participants was that the sample size for other racial groups was too small for stable statistical analysis. The educational and income levels of the participants are low compared with other established cohorts, with less than a high school education in one third of community health center–enrolled participants and an annual household income of <$15 000 in 61% of participants. Details of the rationale, design, and population of the SCCS have been described.²⁸ SCCS participants provided written informed consent, and the Institutional Review boards of Vanderbilt University Medical Center and Meharry Medical College approved all SCCS protocols.

Data Collection

All enrollees completed a baseline, computer-assisted personal interview (available at www.southerncommunitystudy.org) that ascertained information about demographic, socioeconomic, and anthropometric characteristics; personal and family medical histories; diet; exercise; tobacco and alcohol use; and medication use. The reliability of the SCCS questionnaire has been verified through a series of validation studies.²⁸ Participants who enrolled in 2004 or later and responded affirmatively to the question of whether a doctor ever told them they have high BP were asked additional follow-up questions on the use (and names) of prescription medications for BP control. During community health center enrollment, BP measurements were obtained for a subset of 16 932 black and white participants.

The indexes of SES were total household income in the previous year (<$15 000, $15 000–$24 999, $25 000–$49 999, $50 000–$99 999, and ≥$100 000), highest level of education completed (less than 12th grade, high school/vocational school, some college/junior college, completed college, and graduate school), and health insurance status (none, any private insurance, Medicaid only, Medicare only, military only, and other). We calculated body mass index (BMI) using participants’ self-reported weight and height at baseline. A BMI of <18.5 kg/m² defines underweight, 25 to 29.9 kg/m² defines overweight, 30 to 39.9 kg/m² defines obesity, and >40 kg/m² defines morbid obesity. Smoking was categorized as never, former (smoked 100+ cigarettes in their lifetime but do not currently smoke), and current <10, 10 to 19, and ≥20 cigarettes/d. Alcohol consumption was classified as heavy (>3 drinks/d), moderate (≤3 drinks/d), or none during the past year. Finally, information was obtained on self-reported family history of heart disease and personal history of myocardial infarction or coronary bypass surgery, stroke, depression, diabetes mellitus, and high cholesterol.

Statistical Analysis

We evaluated the distribution of baseline characteristics of participants by race and sex in the whole population (Table 1), and then among those with A-HTN but no SR-HTN (Table 2), and finally among those with A-HTN and SR-HTN (Table 3). Crude frequency distributions of categorical variables were compared using χ² tests, whereas t tests or ANOVA was used for continuous variables. The presence of A-HTN was defined as systolic BP (SBP) >140 mm Hg or diastolic BP (DBP) >90 mm Hg. Accordingly, uncontrolled hypertension was defined as participants with SR-HTN in whom A-HTN was present; similarly, those without SR-HTN in whom A-HTN was present were classified as unreported hypertension. All patients on antihypertensive medications were automatically classified as having a history of hypertension by nature of the fact that they had to have SR-HTN to be queried about medication use. Thus, only those with SR-HTN served as the denominator for delineating the prevalence of uncontrolled hypertension.

Table 1.

Self-reported Hypertension in the Southern Community Cohort Study: Baseline Characteristics and Prevalence Among Participants

Characteristic	Overall, % (n=69 211)	Blacks, %		Whites, %
		Men (n=20 900)	Women (n=29 257)	Men (n=6647)	Women (n=12 407)
Hypertension	57	51	64	51	52
Age at interview, y
40–49	44	39	51	40	39
50–59	64	59	73	57	56
60–69	75	74	84	65	67
70–79	78	77	85	70	71
Sex
Women	61
Men	51
Race
Black	59
White	52
Marital status
Married	58	59	63	57	51
Single	49	42	57	43	45
Divorced	57	51	65	48	51
Widowed	72	62	78	58	66
Education
Less than high school	63	56	71	55	59
High/vocational school	55	47	63	49	53
Some/junior (2-y) college	54	48	60	51	48
College	52	49	57	50	43
Graduate school	49	54	57	48* P<0.05	35
Annual household income, $
<15 000	59	52	67	52	56
15 000–24 999	55	48	62	50	52
25 000–49 999	54	51	59	52	47
50 000–99 999	47	48	54	48	39
≥100 000	41	49*	48	41† P> 0.05	31
Health insurance
None	50	43	60	44	49
Any private insurance	54	51	60	51	46
Medicaid only	61	57	65	57	56
Medicare only	73	71	80	66	65
Military only	51	48	55	56	54
Other	71	67	79	61	64
BMI, kg/m2
Underweight, <18.5	36	41	37	29	29
Normal, 18.5–24.9	38	36	46	32	33
Overweight, 25–29.9	53	52	59	50	47
Obese, 30–39.9	68	67	70	68	62
Morbidly obese, >40	77	79	78	77	74
Smoking
Never	63	57	67	56	56
Former	66	66	71	64	58
Current <10 cpd	50	44	57	48	46
Current 10–19 cpd	46	42	55	41	44
Current ≥20 cpd	48	46	57	43	46
Alcohol drinking
Heavy (>3 drinks/d)	44	42	48	45	43
Moderate (≤3 drinks/d)	52	48	59	46	46
None	65	62	70	58	57
Family history of heart disease
No	54	48	61	46	46
Yes	62	57	69	57	57
MI/CABG	55	49	63	47	50
No	83	84	89	78	78
Yes
Diabetes mellitus
No	50	44	58	45	46
Yes	82	81	84	79	77
Stroke
No	55	49	63	49	51
Yes	81	84	85	77	72
High cholesterol
No	48	42	56	39	41
Yes	76	78	81	73	67
Depression
No	55	49	62	48	49
Yes	63	62	71	58	56
Among those with hypertension, n (%) or mean (SD)
Age at diagnosis of hypertension, mean (SD), y	41.1 (12)	41.0 (12)	39.7 (12)	43.0 (13)	44.1 (12)
Duration of hypertension, mean (SD). y	12.5 (11)	11.5 (11)	13.6 (12)	11.3 (11)	11.5 (11)
Antihypertensive medications‡
No	3604 (18)	1218 (24)	1115 (13)	551 (24)	720 (17)
Yes	16 603 (82)	3776 (76)	7489 (87)	1747 (76)	3591 (83)

BMI indicates body mass index; CABG, coronary artery bypass surgery; cpd, cigarettes per day; and MI, myocardial infarction.

All P values for comparison of percentage distributions of persons with versus without hypertension, overall and within race subcategories, were <0.001 unless otherwise noted.

^*P<0.05.

^†P>0.05.

^‡The use of antihypertensive medications was ascertained only for those who enrolled in the cohort in 2004 or later, so percentages are calculated on the basis of that subgroup.

Table 2.

Ascertained Hypertension in the Southern Community Cohort Study: Baseline Characteristics and Prevalence Among 5109 Participants Without Self-reported Hypertension

	Overall, %	Blacks, %		Whites, %
		Men (n=1057)	Women (n=2214)	Men (n=474)	Women (n=1364)
Hypertension	26	31	28	27	17
Age at interview, y
40–49	23	27	25	26	13
50–59	30	35	33	27	22
60–69	29	35	34	29	21
70–79	33	38	36	24*	29
Sex
Female	24
Male	29
Race
Black	29
White	20
Marital status
Married	26	33	28	25	18
Single	25	25	27	27	17
Divorced	26	34	30	26	16
Widowed	26	28	30*	42*	17*
Education
Less than 12th grade	27	31	29	35	17
High school/vocational school	26	30	29	23	18
Some/junior (2-y) college	26	31	29	24	19
College	23	37	24	26	8
Graduate school	23*	22*	26*	24*	19*
Annual household income, $
<15 000	26	30	29	28	17
15 000–24 999	25	29	27	29	16
25 000–49 999	26	37	25	16	20
50 000–99 999	25	32	36	25	11
≥100 000	30*	38*	50*	44*	8*
Health insurance
None	26	30	30	26	18
Any private insurance	27	35	27	29	17
Medicaid only	22	32	24	26	12
Medicare only	28	36	31	28	17
Military only	28	33	56	21	15
Other	24*	17*	29*	26*	20*
BMI, kg/m2
Underweight, <18.5	16	14	27	18	9
Normal, 18.5–24.9	20	26	21	19	12
Overweight, 25–29.9	25	31	26	27	17
Obese, 30–39.9	28	37	29	33	21
Morbidly obese, >40	37	44	41	43	23
Smoking
Never	27	36	28	21	19
Former	28	31	27	34	23
Current <10 cpd	24	22	30	12	15
Current 10–19 cpd	23	33	27	19	10
Current ≥20 cpd	23	29*	30*	28*	15
Alcohol drinking
Heavy (>3 drinks/d)	31	34	33	28	16
Moderate (≤3 drinks/d)	26	29	28	28	17
None	25*	31*	28*	25*	17*
Family history of heart disease
No	25	30	27	28	16
Yes	26*	32*	31*	26*	18
MI/CABG
No	26	31	29	27	17
Yes	23*	29*	20*	23*	18*
Stroke
No	26	31	28	27	17
Yes	25*	25*	29*	28*	20*
Depression
No	27	31	28	26	19
Yes	22	27*	29*	28*	16*
Diabetes mellitus
No	26	31	28	27	17
Yes	26*	31*	29*	25*	15*
High cholesterol
No	27	32	29	26	17
Yes	23*	27*	26*	27*	18*

BMI indicates body mass index; CABG, coronary artery bypass surgery; cpd, cigarettes per day; and MI, myocardial infarction.

All P values for comparison of percentage distributions of persons with versus without hypertension, within race-sex subcategories, were <0.05 unless otherwise noted.

^*P≥0.05.

Table 3.

Ascertained Hypertension in the Southern Community Cohort Study: Baseline Characteristics and Prevalence Among 6183 Participants With Self-reported Hypertension^*

	Overall, %	Blacks, %		Whites, %
		Men (n=1370)	Women (n=2864)	Men (n=640)	Women (n=1309)
Hypertension	54	61	56	47	44
Use of antihypertensive medications†	87	85	89	83	84
Age at interview, y
40–49	54	60	57	47	42
50–59	55	62	57	44	45
60–69	51	59	53	51	43
70–79	54‡	56‡	56‡	53‡	49‡
Sex
Women	53
Men	56
Race
Black	58
White	45
Marital status
Married	52	59	56	49	41
Single	56	62	56	41	45
Divorced	54	61	58	47	44
Widowed	54‡	61‡	55‡	40‡	51‡
Education
Less than 12th grade	55	61	56	49	45
High school/vocational school	53	62	56	47	43
Some/junior (2-y) college	54	57	58	45	46
College	54	64	61	42	37
Graduate school	42‡	33‡	44‡	36‡	45‡
Annual household income, $
<15 000	54	60	57	51	44
15 000–24 999	53	61	55	41	46
25 000–49 999	57	66	61	47	45
50 000–99 999	43	48	53	39	33
≥100 000	47	42‡	71‡	43‡	38‡
Health insurance
None	58	66	61	50	48
Any private insurance	54	63	58	41	39
Medicaid only	50	57	52	47	40
Medicare only	49	49	53	47	40
Military only	55	58	43	56	56
Other	49	56	50	35‡	45‡
BMI, kg/m2
Underweight, <18.5	45	38	63	0	35
Normal, 18.5–24.9	51	59	51	56	35
Overweight, 25–29.9	52	61	53	48	41
Obese, 30–39.9	55	61	57	47	49
Morbidly obese, >40	55	65‡	60‡	38‡	43
Smoking
Never	56	62	58	50	45
Former	51	59	53	45	42
Current <10 cpd	55	59	54	64	47
Current 10–19 cpd	56	62	59	44	47
Current ≥20 cpd	50	61‡	58‡	45‡	43‡
Alcohol drinking
Heavy (>3 drinks/d)	57	65	52	46	47
Moderate (≤3 drinks/d)	56	65	59	45	46
None	52	54	56	48‡	43‡
Family history of heart disease
No	55	60	56	46	44
Yes	52‡	63‡	56‡	47‡	43‡
MI/CABG
No	54	61	56	49	44
Yes	49	56‡	54‡	40‡	46‡
Stroke
No	54	61	57	46	44
Yes	53‡	59‡	54‡	51‡	45‡
Depression
No	56	61	57	50	48
Yes	48	60‡	54‡	42‡	41
Diabetes mellitus
No	55	64	56	48	46
Yes	52	55	56‡	45‡	40‡
High cholesterol
No	57	63	58	53	47
Yes	51	58‡	55‡	42	42‡

BMI indicates body mass index; CABG, coronary artery bypass surgery; cpd, cigarettes per day; and MI, myocardial infarction.

All P values for comparison of percentage distributions of persons with versus without hypertension, within race-sex subcategories, were <0.05 unless otherwise noted.

^*This analysis was restricted to those who enrolled in the cohort in 2004 or later and had information on use of antihypertensive medications.

^†Among those with ascertained hypertension.

^‡P≥0.05.

In the analysis of SR-HTN and A-HTN, multivariable logistic regression models were used to estimate the adjusted prevalence odds ratios (ORs) and 95% confidence intervals (CIs) for factors associated with hypertension, overall and by race-sex. In addition, 3 fundamentally important questions were addressed using SR-HTN and A-HTN variables: (1) the epidemiological question about factors associated with the overall prevalence of hypertension in the SCCS population (number with SR-HTN or A-HTN/total number), (2) the health services question about factors associated with the probability of diagnosis among those with hypertension (number with SR-HTN/ number with SR-HTN or A-HTN), and (3) the clinical effectiveness question about determinants of the probability of control among those with diagnosed hypertension (number with SR-HTN but not A-HTN/number with SR-HTN). These sequences of nested probabilities were similarly modeled using multivariable logistic regression. Furthermore, through multivariable linear regression models, we evaluated the relationships between associated factors and SBP and DBP among participants with measured BP and analyzed pulse pressure (PP) and mean arterial pressure (MAP) to explore whether differences in arterial stiffness may account, in part, for racial differences in the prevalence of hypertension. PP was calculated as SBP minus DBP; MAP was calculated as follows: (DBP-1/3[PP]).

In all analyses, the candidate covariates were identified a priori and included race and sex (when appropriate), age, income, education, health insurance status, marital status, BMI, alcohol intake, cigarette smoking, caffeine intake (derived from responses to the SCCS dietary food frequency questionnaire), physical activity, and medical history (personal and family). A 2-tailed value of P<0.05 was accepted as the threshold for statistical significance. Analyses were conducted using SAS software, version 9.3 (SAS Institute, Inc, Cary, NC).

Results

Table 1 presents baseline characteristics and the distribution of SR-HTN among the 69 211 SCCS participants included in this analysis. Overall, 72% (n=50 157) were blacks and 28% (n=19 054) were whites, and ≈60% were women. At the baseline interview, 39 363 participants reported a previous diagnosis of hypertension, yielding an overall (crude) prevalence of 57%. The prevalence of SR-HTN was significantly higher among blacks than whites (59% versus 52%; P<0.001). The observed racial difference was, however, driven by the higher prevalence among black women (64%) compared with white women (52%), whereas black and white men had similar prevalences (51%). About 82% of those with SR-HTN reported the use of antihypertensive medications, with virtually identical proportions among black and white men (76%), whereas black women were somewhat more likely to use antihypertensive medications than white women (87% versus 83%; P<0.001). Overall, a diagnosis of hypertension was reported more frequently among participants who were older, had higher BMI, were not current smokers, had lower levels of education and income, had a family history of heart disease, and had a personal history of myocardial infarction/coronary bypass surgery, diabetes mellitus, high cholesterol, and depression. These findings were generally consistent across the 4 race and sex groups, with the prevalence of SR-HTN highest among black women across nearly all strata examined.

The distribution of A-HTN according to baseline characteristics among participants with and without SR-HTN is presented in Tables 2 and 3. Overall, there was a higher prevalence of A-HTN among blacks compared with whites. Among participants with measured BP but without SR-HTN, the prevalence of A-HTN in blacks and whites was 29% and 20%, respectively (P<0.001). Women accounted for most of the excess among blacks, with a prevalence of 28% for black women versus 17% for white women (P<0.001), as opposed to 31% for black men and 27% for white men (P=0.049). Among participants with measured BP and with SR-HTN, the prevalence of A-HTN was 58% and 45% in blacks and whites, respectively (P<0.001). The excess prevalence among blacks persisted regardless of sex: 61% versus 47% (P<0.001) for black versus white men and 56% versus 44% (P<0.001) for black versus white women. Table 4 provides details of the distribution of classes of antihypertensive medication among participants with A-HTN and SR-HTN. Notably, among those with A-HTN and SR-HTN, >94% were on at least 1 of the major classes of antihypertensive agents, whereas only 44% were on at least 2 classes and only 29% were on a diuretic agent. Compared with whites, blacks were less likely to be treated with angiotensin-converting enzyme inhibitors or β-blockers but more likely to be treated with calcium channel blockers or α₂-agonists (P<0.01 for all comparisons).

Table 4.

Ascertained Hypertension in the Southern Community Cohort Study: Distribution of Classes of Antihypertensive Medications Among 2880 Participants With Self-reported Hypertension Who Acknowledged the Use of Antihypertensive Medications^*

	Overall (n=2880)	Blacks, n		Whites, n
		Men (n=708)	Women (n=1443)	Men (n=248)	Women (n=481)
ACE inhibitors†	926	196	453	106	171
ACE receptor blockers‡	423	92	227	30	74
Calcium channel blockers§	873	196	500	51	126
β-Blockers†	487	94	215	52	126
Any diuretic‡	835	164	468	62	141
Types of diuretic
Thiazide diuretics	645	124	359	49	113
Loop diuretics	192	41	106	14	31
Potassium-sparing diuretics:	9	3	4	1	1
aldosterone antagonists
Potassium-sparing diuretics:	74	17	45	1	11
sodium channel blockers
α1-Blockers‡	29	8	12	4	5
α2-Agonists§	175	40	102	12	21
Vasodilators‡	4	0	4	0	0
Other‡	153	44	75	8	26

ACE indicates angiotensin-converting enzyme.

^*Data on type of antihypertensive medication used were missing for ≈18% of participants who reported medication use. More than 94% were on at least 1 of the major classes of agents (ACE, ACE receptor blockers, calcium channel blockers, β-blockers, and diuretic), 44% were on ≥2 classes, and 10% were on ≥3 classes.

^†Compared with whites, blacks were less likely to be on these classes of antihypertensive medications (P<0.01).

^‡Compared with whites, blacks were more likely to be on these classes of antihypertensive medications (P<0.01).

^§Black vs white comparisons were not significant (P>0.05).

The adjusted model for SR-HTN in the overall cohort (Figure 1) demonstrates an increased OR for SR-HTN among blacks (OR, 1.84; 95% CI, 1.75–1.93), which was more pronounced in women (OR, 2.08; 95% CI, 1.95–2.21) than men (OR, 1.47; 95% CI, 1.36–1.60), and there was evidence of race-sex interaction (P<0.0001). Education and income were associated inversely with SR-HTN, with ORs of 0.62 (95% CI, 0.55–0.70) and 0.62 (95% CI, 0.51–0.75) among those with the highest levels of education (graduate school) and income (>$100 000), respectively. Other factors consistently positively associated with SR-HTN were age, BMI, history of depression, diabetes mellitus, high total cholesterol, and a family history of heart disease (Figure 1). The patterns of association were consistent across all race-sex groups (Figure IA–ID in the online-only Data Supplement). Of note, the association between morbid obesity and SR-HTN was most pronounced among white men (OR, 4.57; 95% CI, 3.43–6.10; Figure 1C) and white women (OR, 4.64; 95% CI, 3.97–5.43; Figure 1D).

Figure 1.

Self-reported hypertension in the Southern Community Cohort Study. Odds ratios for the associated factors in the overall population. BMI indicates body mass index; CI, confidence interval; and cpd, cigarettes per day.

In the adjusted model for A-HTN among subjects with SR-HTN, blacks had increased OR for hypertension, that is, uncontrolled hypertension (OR, 2.11; 95% CI, 1.67–2.67; Figure 2A). Similarly, among those with A-HTN but without SR-HTN, blacks had increased OR for hypertension, that is, unreported hypertension (OR, 1.99; 95% CI, 1.59–2.48; Figure 2B). Morbid obesity was more strongly associated with unreported hypertension (OR, 5.76; 95% CI, 3.98–8.32; Figure 2B) than uncontrolled hypertension (OR, 1.90; 95% CI, 1.34–2.71; Figure 2A). Compared with women, men had increased OR for unreported hypertension (OR, 1.60; 95% CI, 1.30–1.96; Figure 2B). Elderly participants (70–79 years compared with 40–49 years of age) had an increased OR for uncontrolled hypertension (OR, 1.89; 95% CI, 1.15–3.12), whereas increased ORs for unreported hypertension were nearly twice as high among those at or >50 versus <50 years of age.

Figure 2.

Ascertained hypertension in the Southern Community Cohort Study. Odds ratios for the associated factors among participants (A) with self-reported hypertension and (B) without self-reported hypertension. BMI indicates body mass index; CI, confidence interval; and cpd, cigarettes per day.

Consistent with the findings reported above, advanced age (elderly), black race, male sex, obesity, diabetes mellitus, and high cholesterol were strongly associated with increased odds for the presence of hypertension in the adjusted model for the overall prevalence of hypertension (Figure 3A). Similarly, the same factors except sex were strongly associated with increased odds for diagnosis among those with hypertension (Figure 3B). On the contrary, male sex (OR, 0.79; 95% CI, 0.72–0.87) and black race (OR, 0.6; 95% CI, 0.54–0.67) were both strongly associated with decreased odds for the control of BP among those with diagnosed hypertension (Figure 3C); other factors inversely associated with control of BP were overweight or obesity, age (≥60 years), and heavy alcohol use (Figure 3C).

Figure 3.

A, Overall prevalence of hypertension in the Southern Community Cohort Study (SCCS). Odds ratios for the associated factors. B, Probability of diagnosis among those with hypertension in the SCCS. Odds ratios for the associated factors. C, Probability of blood pressure control among those with diagnosed hypertension in the SCCS. Odds ratios for the associated factors. BMI indicates body mass index; CI, confidence interval; and cpd, cigarettes per day.

The linear regression models for SBP and DBP for participants with and without SR-HTN (Figure 4A–4D) demonstrate that the average SBP was higher and the average DBP was lower with increasing age, reaching a 6.8-mm Hg (95% CI, 4.12–9.46) increase in SBP and a 5.9-mm Hg (95% CI, −7.49 to −4.30) decrease in DBP among participants with ages 70 to 79 years versus 40 to 49 years of age with SR-HTN (Figure 3A and 3B). These findings were similar for participants without SR-HTN. Higher BMI was associated with higher average SBP and DBP measurements; however, the strength of association was strongest among participants without SR-HTN (Figure 4A and 4B). There was a positive association between male sex and SBP and DBP, although it was statistically significant only among participants with and without SR-HTN (Figures 3A, 3B, 4A, and 4B). In all adjusted analyses, the association between alcohol or smoking and hypertension was generally weak and inconsistent. Of note, physical activity was not associated with hypertension in this cohort.

Figure 4.

Relationship between baseline covariates and (A) systolic and (B) diastolic blood pressures in the Southern Community Cohort Study (SCCS) participants with self-reported hypertension (SR-HTN). Relationship between baseline covariates and (C) systolic and (D) diastolic blood pressures in the SCCS participants without SR-HTN.

In analyses of MAP and PP among participants with SR-HTN and measured BP (data not shown), black race was associated with a statistically significant 2.33-mm Hg increase in PP (95% CI, 1.27–3.39; P<0.001) in the adjusted linear model for PP. Relative to the reference age group (40–49 years), there was a significant association between PP and increasing age: for ages 50 to 59 years, 60 to 69 years, and 70 to 79 years, there was a 3.96-mm Hg (95% CI, 2.99–4.94; P<0.0001), a 7.89-mm Hg (95% CI, 6.62–9.16; P<0.0001), and a 12.69-mm Hg (95% CI, 10.63–14.75; P<0.0001) increase in PP, respectively. Of note, sex was not a predictor of PP among those with SR-HTN. Black race was associated with a 4.2-mm Hg increase in MAP (95% CI, 3.33–5.14; P<0.0001). In contrast to the pattern observed for PP, male sex was associated with 1.87-mm Hg increase (95% CI, 1.06–2.68; P<0.0001) in MAP, but the association with age was weak and inconsistent.

Among those with uncontrolled hypertension, male sex was associated with a 2.34-mm Hg decrease in PP (95% CI, −3.73 to −0.94; P=0.001) and black race was associated with a 2.04-mm Hg increase in PP (95% CI, 0.388–3.69; P=0.015). Also, there was strong association between age and PP; compared with those aged 40 to 49 years, there was a 5.17-mm Hg (95% CI, 3.74–6.61; P<0.0001), a 10.47-mm Hg (95% CI, 8.57–12.37; P<0.0001), and a 16.16-mm Hg increase (95% CI, 13.10–19.22; P<0.0001) in PP for those aged 50 to 59, 60 years, to 69 and 70 to 79 years of age, respectively. Male sex was associated with a 1.78-mm Hg (95% CI, 0.87–2.69; P=0.0001) increase in MAP, and black race associated with a 2.78-mm Hg (95% CI, 1.70–3.85; P<0.0001) increase in MAP.

Among those with unreported hypertension, male sex was associated with a PP decrease of 3.18 mm Hg (95% CI, −5.47 to −0.90; P=0.006), whereas black race was not associated with PP (P=0.87). Again, there was a strong association between age and PP, up to a 19.95-mm Hg increase (95% CI, 13.44–26.46; P<0.0001) for ages 70 to 79 years compared with 40 to 49 years. Male sex (P=0.18) and black race (P=0.74) were not associated with MAP in this group.

Discussion

In the SCCS cohort, race, age, and BMI were strongly associated with SR-HTN even after adjusting for potential confounders. In addition, education and income showed modest inverse associations with SR-HTN, whereas associations with smoking or alcohol were weak and inconsistent. Furthermore, among participants with SR-HTN, the high prevalence of A-HTN (58% of blacks versus 45% of whites; P<0.001) implies high prevalence of uncontrolled hypertension and is particularly concerning. Notably, in this subgroup, although >94% reported that they were on at least one of the major classes of antihypertensive agents, only 44% were on ≥2 classes and only 29% reported a diuretic, a recommended first-line agent. Collectively, the pattern of medication use in this cohort suggests suboptimal implementation of JNC 7 guidelines.³ Equally concerning is the high prevalence and racial disparity of A-HTN (29% of blacks versus 20% of whites; P<0.001) among participants without SR-HTN, consistent with unreported hypertension. Furthermore, it is interesting to note that black race, male sex, obesity, and heavy alcohol use were strongly associated with decreased odds for BP control in this cohort. These findings have important implications for future research and healthcare delivery related to the treatment of hypertension.

The fundamental mechanism for the well-established relationship between age and hypertension centers on changes in arterial integrity and function leading to less distensible vessels and an increase in PP, which result in a gradual rise in SBP through adult life but a slight decline in DBP after middle age.²⁹ Thus, the noted associations in our cohort for both blacks and whites mirror established understanding. Similarly, the relationship between BMI and hypertension is well known,²⁹ and likewise the underlying mechanisms, which include endothelial dysfunction and arteriosclerosis,³⁰ increased sympathetic activity,^31–33 activation of the renin-aldosterone- angiotensin system,^34–36 and elevated stroke work because of BP-independent left ventricular remodeling—increase in wall thickness, mass, and cavity size.³⁷

US morbidity and mortality statistics demonstrate that, compared with whites and other races, blacks have greater CVD and associated mortality. Arguably, hypertension is the most prominent beacon for CVD and disproportionately affects blacks.^6,38,39 However, unlike BMI and age, the underlying mechanisms for the higher risk of hypertension in blacks are not entirely clear. The bulk of epidemiological studies in this arena have focused on SES, health insurance coverage, and access to care; it is generally established that SES, along with a parallel disproportionate burden of CVD risk factors, constitutes reasons for health disparities. Although we found inconsistent associations between SES indexes and A-HTN, others have noted inverse links between SES and hypertension,¹⁹ as we did for SR-HTN. However, in this cohort of blacks and whites with nearly identical SES characteristics,^26–28 a higher risk of hypertension persisted in blacks. Furthermore, contrary to conventional wisdom, the association between BMI and hypertension was more pronounced in whites than in blacks. Thus, our findings raise additional questions and call for refinement or re-evaluation of the existing paradigm(s).

Overall, in the analysis of patients with SR-HTN in whom BP data were available, there was a strong association between age and PP that increased with advancing age and thus was consistent with the conventional understanding that PP increases with age because of increasing arterial stiffness. The association between age and PP was further accentuated in the subset of participants with uncontrolled or unreported hypertension. Of note, female sex was consistently associated with higher PP among participants with uncontrolled and unreported hypertension. This association was not apparent in the analysis of PP among those with SR-HTN. In summary, the results support the conclusions that black race and female sex were associated with a higher PP, which collectively suggest that greater arterial stiffness is a plausible underpinning for the race-sex disparities in the prevalence of hypertension. However, this finding does not account for the decreased odds of BP control associated with black race, male sex, obesity, and heavy alcohol use, all of which warrant further investigation.

Despite the results of this study and the unique characteristics of SCCS participants who are of similar and generally low SES regardless of race, SES cannot be ruled out as a contributor to racial disparities associated with hypertension. There are several potentially important social, economic, and environmental determinants that were not evaluated in this study. Most important, the interaction of SES and biology is likely relevant in precipitating biomolecular characteristics that could influence hypertension differentially by race. Credence for this suggestion stems from the fact that molecular markers, for example, plasma levels of fibrinogen, homocysteine, and intercellular adhesion molecule-1, are significantly increased in blacks compared with whites even after adjustment for traditional risk factors.^40,41 Furthermore, blacks and whites with identical SES may still have significant differences in family history of hypertension, dietary salt intake, consumption of fruits and vegetables, and other biosocial determinants of hypertension; compounding these may be important differences in salt sensitivity, which modulates the incidence (and prevalence) of hypertension and hypertensive target-organ damage and overall CV mortality.⁴² In addition, recent evidence suggests that past SES factors—prenatal and childhood stress, neighborhood poverty and crime exposure, parental health and SES, and poor intergenerational economic mobility—may play much greater roles in the pathogenesis of racial health disparities than present SES conditions.⁴³ The association between adult-onset disease and fetal and early-life exposures, the proposition of the Barker hypothesis, has been documented in many populations.^44–51 Therefore, future research that seeks to assess the role of SES in hypertension should extend the evaluation beyond current SES characteristics to include family background and neighborhood conditions during childhood.

Limitations and Strengths

Some inconsistent associations may stem, in part, from random error because of misclassification and inaccuracies of self-reported data, which may have attenuated the power to observe potentially subtle associations. Another limitation is that this was a cross-sectional evaluation of cohort data, thus precluding any inferences on the temporal relationship between examined covariates and hypertension; this limitation may have also contributed to the inconsistencies in the association between hypertension and lifestyle factors, which are subject to change after a diagnosis of hypertension. As the cohort is followed up over time, future analyses will prospectively examine hypertension as a predictor of subsequent mortality, stroke, CVD, and other health outcomes. We emphasize that the similar and generally low income and education levels of SCCS participants regardless of race greatly minimized confounding by SES, generally an important mediator of racial differences in health care and burden of disease.

Despite the above limitations, this study has many strengths, including its large sample size, extensive data on lifestyle and SES indexes, and the internal validation from comparative analyses of A-HTN using measured BP. Furthermore, the observed relationships between age and BP are consistent with conventional understanding of pathophysiology and lend credence to the overall data and results. Finally, this study is the largest evaluation of the factors associated with hypertension in a cohort of blacks and whites of similar, predominantly low contemporaneous SES in the southeastern United States.

Conclusions

In this evaluation of hypertension in a large cohort of blacks and whites, race, age, and BMI were strongly associated with the risk of hypertension in all models before and after adjustment for potential confounders. However, the mechanisms underlying the higher risk of hypertension in blacks were not explained by differences in lifestyle, education, or income. The reasons for the suboptimal use of evidence-based treatment guidelines in this population need to be explored. In this context, analysis is currently underway to determine the characteristics and factors associated with the use of antihypertensive medications in the SCCS cohort. In terms of future research directions, we suggest that additional insight into the biomolecular milieu, prenatal and early-life exposures, historic SES conditions, and their interplay in patients with hypertension may improve our understanding of health disparities.

WHAT IS KNOWN

• Hypertension is an established risk factor for cardiovascular disease and is the leading risk factor for death and disability worldwide.

• The prevalence of cardiovascular disease risk factor is high in the Southeastern United States.

WHAT THE STUDY ADDS

• The prevalence of unreported and uncontrolled hypertension across 12 states in the southeastern United States is high.

• After controlling for socioeconomic status, black race, age, and body mass index are strongly associated with hypertension, including unreported and uncontrolled hypertension.

• The pattern of medication use suggests suboptimal implementation of evidence-based hypertension treatment guidelines.