Ethnic differences in blood pressure control in men

Ethnic disparities in cardiovascular and renal diseases are significant for African Americans, particularly men.^1-7 Hypertension is not only a more prevalent but also a more powerful risk factor for these diseases in African Americans than in whites.^8,9 African Americans also have lower control rates of hypertension than whites.¹⁰ Recent National Health and Nutrition Examination Survey reports show that only 26.5% of African American men with hypertension have reached the blood pressure control goal of less than 140/90 mm Hg compared with 36.5% of white men with hypertension. Among African American and white women with hypertension, however, blood pressure rates are similar, at 29.4% versus 30.5%.¹⁰

Health care system factors, such as access to health care and insurance, as well as cultural, social, and biological factors, may be the cause of the ethnic differences in blood pressure control, although the reasons are still unknown. Previous studies, however, have suggested that improved access to medications and care do not result in improved blood pressure control.^11,12 Because the US Department of Veterans Affairs (VA) health system provides health care and medications to veterans of all ethnic backgrounds without discrimination, we compared blood pressure control and visit frequency (a marker of access) between African American and white men with hypertension in VA and non-VA health systems to further assess this relationship.

Methods

Blood pressure treatment and control were compared in African American (VA, n = 4,379; non-VA, n = 2,754) and white (VA, n = 7,987; non-VA, n = 4,980) hypertensive men. The data source was the Hypertension Initiative proj­ect.¹³ Patient demographic characteristics, including age, sex, ethnicity, height, weight, dates of medical visits, and blood pressure, were documented at each visit from January 2001 through December 2003 and were reported as mean ± standard error of the mean (SEM) for comparative purposes and as mean ± standard deviation for descriptive purposes.

Chi-square tests were used to ex­amine ethnic differences in treatment and control of hypertension. Using the Cochran-Mantel-Haenszel method, African American and white patients were further evaluated, controlling for sex and age. The blood pressure control rate was compared between VA and non-VA patients using multivariate logistic regression, adjusting for comorbid conditions, such as cardiovascular disease, congestive heart failure, renal disease, diabetes, and lipid disorders; Rural/Urban Code Assign­ment (RUCA; a classification of rural/urban residency); body mass index (BMI); ethnicity; and age.

Descriptive variables by ethnic groups.

Table 1.

African Americans

Whites

P value

All sites

n

7,133

12,967

Age, y

60.4 ± 0.01

63.2 ± 0.01

< .001

SBP, mm Hg

139.9 ± 0.2

136.3 ± 0.2

.001

DBP, mm Hg

81.0 ± 0.2

77.0 ± 0.1

< .001

BMI, kg/m2

28.8 ± 0.05

29.3 ± 0.03

<.001

No. of RF

2.64 ± 0.01

2.81 ± 0.01

.001

< 140/90 mm Hg (%)

47.3

55.1

<.001

< 150/95 mm Hg (%)

68.1

76.7

<.001

VA sites

n

4,379

7,987

Age, y

62.3 ± 0.2

66.1 ± 0.1

<.001

SBP, mm Hg

140.2 ± 0.3

137.3 ± 0.2

<.001

DBP, mm Hg

79.9 ± 0.2

75.7 ± 0.1

<.001

BMI, kg/m2

29.4 ± 0.1

29.8 ± 0.1

<.001

No. of RF

2.86 ± 0.01

2.99 ± 0.01

<.001

< 140/90 mm Hg (%)

49.4

55.6

<.001

< 150/95 mm Hg (%)

69.6

76.2

<.001

Non-VA sites

n

2,754

4,980

Age, y

57.4 ± 0.3

58.4 ± 0.2

.016

SBP, mm Hg

139.5 ± 0.4

134.7 ± 0.2

<.001

DBP, mm Hg

82.7 ± 0.2

79.2 ± 0.2

<.001

BMI, kg/m2

27.9 ± 0.1

28.5 ± 0.1

<.001

No. of RF

2.29 ± 0.01

2.5 ± 0.01

<.001

< 140/90 mm Hg (%)

44.0

54.2

<.001

< 150/95 mm Hg (%)

65.7

77.5

<.001

Values are mean ± SEM. SBP = systolic blood pressure; DBP = diastolic blood pressure; BMI = body mass index; No. of RF = number of risk factors (hypertension, diabetes mellitus, hyperlipidemia). (Tables 1, 2 and Figure reprinted, with permission, from Arch Intern Med. 2005;165[9]:1041-1047. Copyright© 2005, American Medical Association. All rights reserved.)

Results

There were 12,366 participants from VA sites, of whom 64.6% were white (mean age, 64.8 ± 0.1 years) and 7,734 participants from non-VA sites, of whom 64.4% were white (mean age, 58.1 ± 0.2 years). The BMI for VA subjects was 29.6 ± 0.01 kg/m² and 28.3 ± 0.07 kg/m² for non-VA subjects (P < .001). The mean systolic blood pressure for VA subjects was 138.3 ± 0.2 mm Hg compared with 136.4 ± 0.2 mm Hg for non-VA subjects (P < .001). VA subjects had a mean diastolic blood pressure of 77.2 ± 0.1 mm Hg compared with 80.4 ± 0.1 mm Hg for non-VA subjects (P < .001). Among VA subjects, 53.4% had a blood pressure below 140/90 mm Hg and 73.8% had a blood pressure below 150/95 mm Hg, compared with 50% and 73.3%, respectively, for non-VA patients. As shown in Table 1, whites were older than African Americans in both groups (P < .05). The Figure shows that whites also had lower blood pressure (P < .001) and had blood pressure controlled to less than 140/90 mm Hg more often on their last visit (P < .01). Blood pressure was within 10/5 mm Hg of goal in about 20% of African American and white patients at both sites.

In both ethnic groups, baseline systolic blood pressure was higher and diastolic blood pressure was lower in VA patients compared with non-VA patients. VA patients were also ap&shy;proximately 6 years older than non-VA patients. Blood pressure control to less than 140/90 mm Hg was comparable between white hypertensives in VA (55.6%) and non-VA (54.2%) settings. In contrast, blood pressure control was better among African American hypertensive patients in VA (49.4%) compared with non-VA (44.0%) settings (P < .01), even after controlling for age, comorbid conditions, and RUCA. The difference in the percentage of white and African American hypertensive men with blood pressure controlled to less than 140/90 mm Hg was smaller in VA clinics than in non-VA clinics but re&shy;mained significant (Figure).

African American men with hy&shy;pertension received a comparable number of prescriptions for blood pressure medications at VA sites (1.76 ± 0.02 versus 1.73 ± 0.01) and received more prescriptions at non-VA sites than whites (2.36 ± 0.03 versus 2.08 ± 0.02; P < .001). Whites at VA sites had fewer visits in the previous year than African Americans (4.12 ± 0.03 versus 4.53 ± 0.05; P < .001) and more visits at non-VA sites (3.67 ± 0.02 versus 2.80 ± 0.02; P < .001). At VA sites, whites re&shy;ceived fewer angiotensin re&shy;ceptor antagonists and a similar num&shy;ber of a,b- blocking agents, a-adrenergic blocking agents, angio&shy;tensin-converting enzyme (ACE) inhibitors, cal&shy;cium channel blocking agents, and diuretics than African Americans. At non-VA sites, whites received fewer prescriptions than African Ameri&shy;cans for angiotensin receptor blockers, a,b-blocking agents, a-adrenergic block&shy;ing agents, ACE inhibitors, calcium channel blocking agents, and di&shy;uretics. More b blocking agents were prescribed to whites than to African Americans at both sites.

Between VA and non-VA sites, age and ethnicity were major determinants of differences in blood pressure control as shown on multivariate logistic regression analysis. Older age, higher BMI, and history of nephropathy were associated with lower blood pressure control rates. Patients with heart failure and cardiovascular disease had higher blood pressure control rates.

Multivariate logistic regression model of blood pressure control (< 140/90 mm Hg) for African Americans

.

Table 2.

Variable

Reference category

Odds of adequate control (95% CI)

VA status: non-VA

VA

0.839 (0.742—0.949)

CV risk factors^*

NA

1.106 (1.039—1.178)

Age

NA

1.002 (0.998—1.006)

RUCA U

RUCA IR

1.498 (1.128—1.989)

RUCA LR

RUCA IR

1.331 (0.945—1.875)

RUCA SR

RUCA IR

1.820 (1.353—2.448)

BMI

0.995 (0.984—1.006)

Nephropathy: yes

No

0.957 (0.836—1.096)

CVD: yes

No

1.069 (0.958—1.194)

CHF: yes

No

1.194 (1.010—1.412)

CV = cardiovascular; NA = not applicable; RUCA = Rural/Urban Code Assignment; U = urban; IR = isolated rural; LR = large rural; SR = small rural; CVD = cardiovascular diseases, including stroke, transient ischemic attack, angina, myocardial infarction, cerebrovascular disease, coronary heart disease, peripheral arterial disease;

CHF = congestive heart failure.

^*CV risk factors, including hypertension, hyperlipidemia, and diabetes mellitus, were scored as 1, 2, or 3.

Patients at VA and non-VA sites had similar socioeconomic status as measured by RUCA. At both sites, whites had higher education and in&shy;come status than African Americans. After controlling for RUCA, nephro&shy;p&shy;athy, heart failure, cardiovascular disease, diabetes, lipid disorders, BMI, and age, multivariate logistic regression showed that blood pressure control remained lower for hypertensive African Americans at non-VA clinics than at VA clinics (odds ratio [OR], 0.839; 95% confidence interval [CI], 0.742—0.949; Table 2). In other words, non-VA African American patients were 16.1% (95% CI, 5.1—25.8) less likely to have their blood pressure controlled to less than 140/90 mm Hg. There were no differences between VA and non-VA white pa&shy;tients using the same analysis.

Discussion

Hypertension is an important modif&shy;iable risk factor for cardiovascular and renal diseases. Proper treatment of hypertension reduces cardiovascular and renal events in all ethnic groups and reduces disparities in cardiovascular and total mortality be&shy;tween African Americans and whites.^14,15 The benefits of better blood pressure control for improving health and reducing disparities have not been fully realized because hypertension control rates remain low, particularly among ethnic minorities.^9,10

Our study shows that African American men with hypertension had better blood pressure control in VA than in non-VA settings. As shown in the Figure, panel A, 44.0% of African American non-VA men and 49.4% of VA African American men reached the blood pressure goal of less than 140/90 mm Hg. At both sites, however, blood pressure control among white men was similar. The ethnic disparity between white and African American men was greater at the non-VA clinics than at VA clinics. This indicates that the sites of care affect blood pressure control more in African American men than in white men.

Differences in RUCA, comorbid conditions, BMI, the number of antihypertensive medications, median in&shy;come, education level, and age did not account for the higher blood pressure control rates among African Ameri&shy;can patients in non-VA sites, as shown in the multivariate analysis. The multivariate analysis actually indicated that the disparity in blood pressure control between African American men receiving care at VA compared with non-VA sites would have been even greater if confounding variables had been equally distributed. African American men with hypertension re&shy;ceiving care at the non-VA sites were 16.1% (95% CI, 5.1—25.8) less likely to have a blood pressure below 140/90 mm Hg at their last visit than their counterparts at the VA. African Americans also had more visits per year than whites in the VA sites and fewer visits than whites at non-VA sites. It is possible that higher visit frequency contributes to better blood pressure control in African American men seen at VA sites.

These findings suggest that the VA system offers advantages in controlling hypertension, especially for Afri&shy;can Americans. Our data indicate that better access to health care, assessed by visit frequency, and better access to medications may possibly contribute to the success in controlling blood pressure among African American VA patients.

One interesting finding was that about 20% of men with hypertension at both VA and non-VA sites had blood pressures less than 150/95 mm Hg but greater than or equal to 140/90 mm Hg, suggesting that modest improvements in blood pressure control in this group could result in approximately 65% to 70% improvements overall in blood pressure control rates. This finding un&shy;derscores the need to intensify therapy to achieve the goal of less than 140/90 mm Hg.

One limitation of our analysis was that the data originated mainly from medical practices in South Carolina, and the findings may not be typical of health care practices nationally. Al&shy;though information was collected on the number and class of antihypertensive medication prescribed, data on prescription fill and refill rates was not obtained. This is an important limitation because data on actual fill and refill rates would provide insight into adherence to therapy. In addition, data on patients from non-VA sites were obtained from several different electronic record systems and from report cards completed at visits, whereas a common electronic medical record system was used to obtain data on men from VA clinics.

Conclusions

Blood pressure was more often controlled to goal in African Ameri&shy;can men at VA than at non-VA clinics. White patients, however, had a similar blood pressure control rate at the 2 sites. The disparity in blood pressure control rates between the two ethnic groups was approximately 40% less at VA clinics than at non-VA clinics. Differences in access to care and medications may have contributed to the success of patients in the VA settings. The results of our study indicate that ethnic disparities in control of blood pressure and in cardiovascular events may be diminished by ensuring access to health care and antihypertensive medications.