|Home | About | Journals | Submit | Contact Us | Français|
Pre-hypertension is a major risk factor for hypertension. African Americans (blacks) have more prevalent and severe hypertension than whites, but it is unknown whether progression from pre-hypertension is accelerated in blacks. We examined this question in a prospective cohort study of 18,865 non-hypertensive persons (5,733 [30.4% black, 13,132 [69.6%]) white) 18–85 years old. Electronic health record data were obtained from 197 community-based outpatient clinics in the Southeast U.S. Days elapsing from study entry to hypertension diagnosis, mainly blood pressure [BP] ≥140 systolic and/or ≥90 mmHg diastolic on two consecutive visits established conversion time within a maximum observation period of 2550 days. Cox regression modeling was used to examine conversion to hypertension as a function of race, while controlling for age, sex, baseline systolic and diastolic BP, body mass index [BMI], diabetes mellitus and chronic kidney disease. The covariable adjusted median conversion time when 50% became hypertensive was 365 days earlier for blacks than whites (626 vs 991 days, p<0.001). Among covariables, baseline systolic BP 130–139 (Hazard Ratio 1.77, 95% Confidence Intervals [1.69–1.86]) and 120–129 mmHg (1.52 [1.44–1.60] as well as age ≥75 (1.40 [1.29–1.51] and 55–74 years (1.29 [1.23–1.35] were the strongest predictors of hypertension. Additional predictors included age 35–54 years, diastolic BP 80–89 mmHg, overweight and obesity, and diabetes mellitus (all p<0.001). Conversion from pre-hypertension to hypertension is accelerated in blacks, which suggests that effective interventions in pre-hypertension could reduce racial disparities in prevalent hypertension.
In 1939, Robinson and Brucer reported that individuals with blood pressures of 120–139/80–89 mmHg were more likely to become hypertensive and die earlier than people with normal blood pressures <120/<80.1 They labeled this blood pressure range as pre-hypertension. The Joint National Committee on High Blood Pressure adopted the term pre-hypertension for the first time in its Seventh Report (JNC 7) in May 2003. As in 1939, JNC 7 defined pre-hypertension by blood pressures 120–139/80–89 mmHg.2,3 Although pre-hypertension was controversial, several studies confirmed the 1939 report that pre-hypertension is a risk factor for hypertension and cardiovascular diseases (CVD).4 Among people with BP 130–139 systolic and/or 85–89 mmHg diastolic (Stage 2 pre-hypertension), the risk of developing hypertension is threefold that of normotensives with BP <120/<80.5-7 The risk of CVD morbidity and mortality is ~1.6–2.0 greater in Stage 2 pre-hypertensives than normotensives, even without progression to hypertension.7
In the U.S., African Americans (blacks) have a higher prevalence of hypertension and associated cardiovascular and renal complications than Caucasians (whites).8, 9 Moreover, even without progression to hypertension, pre-hypertensive blacks experience more CVD complications than whites.10 These observations suggest that effective interventions, which lower blood pressure in prehypertensives and retard the progression to hypertension, could potentially reduce black-white health disparities in prevalent hypertension.3, 11 In fact, The ISHIB Consensus Statement recommends comprehensive lifestyle intervention in blacks with blood pressure ≥115/≥75 mm Hg rather than ≥120/≥80 as in JNC 7.12
JNC 7 recommended therapeutic lifestyle change only for pre-hypertension in the absence of diabetes and clinical cardiovascular and renal disease.1 Clinical efficacy studies document that lifestyle changes, e.g., the Dietary Approaches to Stop Hypertension (DASH), lower blood pressure in non-hypertensive persons, especially blacks.13 However, adoption of healthy nutrition and physical activity patterns are limited in the U.S. population whose dietary patterns have become less DASH-like as body mass index and prevalent obesity have risen.14, 15
These observations suggest that adoption of lifestyle interventions is limited in the population and that a large proportion of individuals with Stage 2 pre-hypertension progress to hypertension within four years. Thus, safe and effective pharmacological options emerge as a logical complement to therapeutic lifestyle change for reducing prevalent hypertension.3 If blacks with pre-hypertension progress to hypertension at a faster rate than whites, then the rationale for studying pharmacological interventions in this high risk group would be strengthened. While several studies examined progression from pre-hypertension to hypertension,5, 7, 16 none addressed differences between black and white individuals. Our study addresses that important knowledge gap.
This study utilized the Outpatient Quality Improvement Network (OQUIN) Hypertension Initiative database.17 OQUIN captures data on patients from the Southeast U.S who are receiving healthcare at practices with electronic health record systems. Data collection is facilitated through a Business Associate Agreement with each clinic. The Agreement addresses the Health Insurance Portability and Accountability Act (HIPAA) regulations and includes a provision for use of de-identified data for research. The data file includes patient demographics (age, race, sex), visits and visit dates, vital signs, diagnoses and procedure codes, medications, and laboratory data. The source of race information retrieved from the electronic health record was not specified. This study was approved by the Medical University of South Carolina Institutional Review Board.
This study examined differences in conversion rates between non-Hispanic blacks (Blacks) and non-Hispanic whites (Whites) in a prospective observational cohort design. Patients were selected from among 1,720,242 patients seen at participating practices between January 1, 2003 and December 31, 2007. Subjects were eligible for this study if they were 18–85 years old and had ≥4 visits with a valid blood pressure (BP) measurement over a period ≥2 years.
Patients were excluded if they had initial BP values ≥140 systolic and/or ≥90 mmHg diastolic, a diagnosis of hypertension (ICD-9-CM 401–405), or prescription for antihypertensive medication. Persons with a history of active drug or alcohol abuse (ICD-9-CM 303-305), major psychiatric illness (ICD-9-CM 290-299), or malignancy (ICD-9-CM 140-208) were excluded. Persons with missing race, Hispanic ethnicity, or who did not meet inclusion criteria were excluded. Figure 1 depicts derivation of the study sample.
Recruitment was open-ended and began 1 January 2003 and closed 31 December 2007. Data collection for patients enrolled during this five-year period ended on 31 December 2009, which yielded a maximum observation period of 2,550 days.
The primary end-point was conversion from non-hypertensive, i.e., normal BP (<120/<80 mmHg) or pre-hypertension (120–139/80–89 mmHg), to hypertensive status. From the electronic health record, new onset hypertension was defined by: (i) systolic BP ≥140 and/or diastolic BP ≥90 mmHg on two consecutive visits (ii) new ICD9 (401–405) diagnosis of hypertension (iii) initiation of antihypertensive treatment identified from the medical record. The pre-hypertensive BP range was divided into Stage 1 pre-hypertension with BP 120–129 systolic and/or 80–84 mmHg diastolic and Stage 2 pre-hypertension with SBP 130–139 and/or DBP 85–89 mmHg.7 Based on the empirical joint distribution of systolic and diastolic BP values, nine non-hypertensive BP categories were defined including normal BP and 8 categories of pre-hypertension to determine their influence on progression to hypertension.
The number of days elapsing from study entry to hypertension conversion established the survival time (T). The main independent variable, race was obtained from demographic information in the electronic health record. For some patients living in South Carolina without race in the electronic health record, this information was obtained from the State Office of Research and Statistics Uniform Billing claims database. Race was categorized as black or white with exclusion of other race/ethnicity groups including Hispanic and unknown.
Body mass index (BMI, kg/m2) was calculated. Chronic kidney disease (CKD) was defined as estimated glomerular filtration rate [eGFR]) <60 ml/1.7 m2/min or history of CKD (ICD-9 codes 403, 404, 585). Diabetes mellitus was defined by ICD-9 code 250 and/or prescriptions for insulin or oral anti-diabetic medications.
Data analyses were conducted with SAS software package (SAS V9.03, Cary, NC).18 Descriptive statistics for group comparisons included t-tests for continuous variables and Chi-square for proportions. The relationship between the dependent variable (conversion to hypertension) and various risk factors or independent variables was examined as a function of time using survival analysis. Variables with bivariate association P-values ≤0.20 were included in the multivariable model. Multi-colinearity among covariates was evaluated by assessing deviations of regression coefficients and their standard errors in the fitted univariate and multivariate models,19 and none was detected. Covariates were entered simultaneously into the model. Age-adjusted Kaplan-Meier survival curves graphically presented the relationship of the cumulative proportion of conversions to hypertension among Blacks and Whites. A log-rank test was used to test the homogeneity of survival curves across racial strata.20 P-values <0.05 were considered significant.
Cox Proportional Hazards Regression was used to estimate the prognostic influence of race on conversion to hypertension, while simultaneously controlling for the confounding effect of covariates. This model estimates the instantaneous relative risk of conversion to hypertension averaged over the entire time of follow-up. The proportional hazard assumption was tested with the goodness-of-fit chi-square test, which compares the observed and expected survival probabilities, and by graphical means using the log-log Kaplan-Meier curves.21 The heterogeneity of the stratum-specific hazard ratios between systolic BP and conversion across the various stages of diastolic BP as proposed by Breslow-Day22 for analysis of cohort data was used.23 Adjusted HRs and 95% confidence intervals are reported.
In this study, 18,865 individuals (Figure 1) were followed through the end-point of hypertension or the end point of 2,550 days. Non-hypertensive subjects who met exclusion criteria except for race unknown or other than black or white (N=12,868) were younger, more likely to be men, overweight, and normotensive and to have CKD and diabetes, and less likely to have normal weight and blood pressures (data not shown, all p-values <0.01). A total of 12,045 patients (63.8%) progressed to clinical hypertension. Significant differences were identified between converters and non-converters in all covariables but sex (Table 1). The mean age of the cohort was 48.5 (SD ± 15.7) years, 43% were obese and 27.4% were diabetic. Converters were significantly older than non-converters.
Table 2 shows the baseline prevalence of pre-hypertension in this non-hypertensive cohort. Within this cohort, 28% were normotensive, 27.2% had Stage 1 pre-hypertension(120–129/80–84 mmHg), and 45.2% had Stage 2 pre-hypertension (130–139/85–89).7 Significant group differences were observed in several non-modifiable covariables with patients who were black, male, older, diabetic and with chronic kidney disease having a higher prevalence of Stage 2 pre-hypertension. Body mass index was the only readily modifiable covariable examined. There was a positive linear gradient of baseline stage 2 hypertension status with increasing body mass index, with the highest rate (52.4%) among Grade III obesity and the lowest (25.6%) among underweight individuals. All patients with chronic kidney disease and 72% with diabetes mellitus had stage 2 pre-hypertension at baseline.
Table 3 summarizes results of multivariable Cox Proportional Hazard Regression analyses and risk of conversion to hypertension over 7-years follow-up. Crude risk ratios were larger than adjusted ratios, which indicate confounding effects of covariables in the model. The overestimation eliminated by adjusting for other covariables ranged from 62.0% for CKD to 3.2% for sex. Blacks had greater risk of conversion to hypertension than whites (HR=1.35; 95% CI=1.30, 1.40) after adjusting for covariables. The strongest predictor of conversion was stage-2 systolic pre-hypertension followed by stage-1 systolic pre-hypertension. The risk of progressing to hypertension increased with advancing age. Patients who are ≥75 years old had 46% higher risk of conversion than those ages 18–34 years (P<0.001).
Persons with stage-2 and stage-1 diastolic pre-hypertension had significantly elevated risk of conversion independently of other co-variables. Unlike Stage–2 and Stage–1 systolic pre-hypertension, risk for conversion to hypertension did not differ between Stage–2 and Stage–1 diastolic pre-hypertension.
Compared to individuals with normal BMI (18.5–<25 kg/m2), risk of new onset hypertension rose with increasing BMI category (Table 3). Underweight was associated with a lower mean estimate of conversion to hypertension, although not statistically significant. Chronic kidney disease and diabetes mellitus increased risk of new onset hypertension.
Table 4 shows the empirical distribution of nine groups of non-hypertensive individuals and risk of each group for conversion to hypertension. The largest group is comprised of isolated systolic pre-hypertension with 20% Stage 1 and 16% Stage 2. Generally, the higher the stage of pre-hypertension, the greater was the risk of new onset hypertension. With normotensives (Group 0, reference), a biological gradient was noted with progressive covariate-adjusted increasing risk from Group 1 (normal systolic BP, Stage 1 diastolic pre-hypertension), to Group 8 (stage 2 systolic and diastolic pre-hypertension [linear trend of HR P<0.001]).
The nine stages of non-hypertension indicated a transadditive effect of systolic and diastolic blood pressure on risk for hypertension. For example, patients with stage 2 systolic and diastolic pre-hypertension (Group 8) had a HR for future hypertension of 2.42, which was greater than the sum of the separate components of isolated Stage 2 systolic and isolated Stage 2 diastolic prehypertension (HR 1.75 + 1.21, respectively, = 1.96) as shown in Table 3. The small difference between the crude and adjusted HRs, expressed as crude/adjusted, indicate that the degree of confounding by covariables was less than 4% for all nine grades of pre-hypertension. The differences in crude and adjusted hazard ratios using the 9 groups were smaller than ratios using only three groups, i.e., normal BP, Stage 1 and Stage 2 pre-hypertension (Table 3).
Figure 2 shows age-adjusted Kaplan-Meier probability curves of incident hypertension (percent remaining non-hypertensive) by race over 2,550 days. The probability of blacks and whites remaining hypertension-free separated during the first few months. The median conversion time [S(T)50] of blacks was 365 days earlier than whites (626 vs 991, p<0.001). Similarly, at the midpoint or 1,250 days, 65% of blacks had already converted to hypertension compared to 54% of the whites. By end or censoring point of the 2,550th day, about 70% of blacks and 56% of whites had converted.
The principal study finding is that black race, as recorded in the electronic health record, is associated with an accelerated risk of new onset hypertension. Our findings are based upon community practice-based electronic health record system data and included adjustments for several confounding clinical and demographic covariables. In addition to race, several other clinical characteristics were independently associated with new onset hypertension including age, baseline systolic and diastolic blood pressure, body mass index, chronic kidney disease, and diabetes mellitus. These findings are consonant with previous reports on risk factors in pre-hypertension for cardiovascular disease.10, 24-30 The dose-dependent risk of conversion noted with increasing BMI from underweight to Grade III obesity is consistent with studies linking obesity and adiposity to the pathophysiology of elevated blood pressure and hypertension.31-33
Another novel study finding was the transadditivity of systolic and diastolic BP when both values were examined simultaneously as predictors of new onset hypertension (Table 4). A biological gradient of risk was observed, which corresponded to the eight categories of pre-hypertension when systolic and diastolic BP are considered simultaneously rather than the three separate stages of systolic and diastolic pressure. Moreover, the difference between unadjusted and adjusted crude rates was reduced when nine rather than three categories of non-hypertensive systolic and diastolic blood pressure values were considered (Table 3). Thus, considering systolic and diastolic simultaneously substantially reduces the confounding by covariables that occurs when systolic or diastolic pressures are examined separately. While both systolic and diastolic pre-hypertension increase the risk of conversion to hypertension, the influence of systolic appears stronger. The stronger relationship of systolic than diastolic to cardiovascular events has also been reported, especially for people ≥50 years old.2, 34-36
Isolated stage 1 and 2 systolic pre-hypertension with normal diastolic (Grades 2 and 5) accounted for 36% of pre-hypertensive subjects in our cohort. Nearly 2/3 of adults with isolated systolic pre-hypertension were ≥55 years of age, which is consistent with the notion that age-related arterial stiffening,37, 38 possibly reflecting or coincident with age-related increase in salt-sensitivity of blood pressure,16 contributes to pre-hypertension and progression to hypertension.39
In this study, blacks had a 35% greater risk of conversion to hypertension than whites after adjustment for multiple covariables (Table 3). Moreover, the median, age-adjusted conversion time when 50% of the group became hypertensive occurred after only ~1.7 years of follow up in blacks as compared to ~2.7 years in whites (Figure 2). The median conversion time in our study, which did not have standardized BP measurements or visits frequencies, is comparable to the 2.2 years reported for the predominantly white Stage 2 pre-hypertensives in TROPHY, a rigorous clinical trial.3 The longer median conversion time for whites in our study than the overall TROPHY study is expected, since our report included Stage 1 pre-hypertensive and normotensive individuals. On the other hand, patients with diabetes mellitus at baseline were at greater risk for incident hypertension, and TROPHY excluded these patients.
As noted, JNC 7 recommended therapeutic lifestyle change only for most individuals with pre-hypertension. While lifestyle interventions, e.g., DASH, are efficacious in lowering blood pressure in non-hypertensive individuals,13 adoption in the population is limited.15 Even in a relatively intensive clinical efficacy study (TOHP), the relative reduction in progression to hypertension over four years was ~15% with an absolute reduction of 6–7%.40 The relative risk reduction for de novo hypertension over four years was virtually identical to TROPHY patients two years after angiotensin receptor blockade was discontinued.5 At the two year time point when angiotensin receptor blockers were discontinued, the active treatment group developed hypertension at approximately 1/3 the rate of the placebo group (relative risk 0.34, 95% confidence interval 0.25–0.44, p<0.001). Only 79 of 772 evaluable patients in TROPHY were black, which limits assessment of black-white differences in response to treatment.
This prospective cohort study relied on data in several different electronic health record systems obtained from a variety of outpatient clinical settings in the Southeast U.S. Race ascertainment was not standardized. Unlike the National Health and Nutrition Examination Surveys, blood pressure measurements were not standardized and likely included multiple suboptimal practices.8, 39 And unlike population studies, e.g. the Framingham Heart Study, time between visits was not standardized. Of note, the Framingham report on incident hypertension was based on a single follow-up visit four years after baseline assessment.5 Non-hypertensive patients excluded from the analysis for race unknown or other than black or white differed by age, sex, weight, blood pressure category and prevalence of diabetes and chronic kidney disease. For these and other reasons, our sample and findings may not represent the broader population of non-hypertensive individuals. On the other hand, the data reflect findings in contemporary medical practice among a diverse group of clinical sites in the Southeast U.S. Despite limitations, black race was associated with an accelerated transition from non-hypertensive to hypertensive status with 50% transitioning to hypertension in only 1.7 years vs. 2.7 years in whites. The finding of racial differences in progression from hypertension to pre-hypertension does not necessarily implicate biological differences but can include sociocultural and economic factors.41, 42
Our study findings are potentially relevant to national health goals for reducing prevalent hypertension and improving health equity Healthy People 2000 and 2010 included a goal to reduce the prevalence of hypertension in the U.S. population from 29% to 16%.8, 43, 44 Despite JNC 7 efforts to facilitate a reduction of incident hypertension by resurrecting the 1939 term ‘pre-hypertension’,1 the lifestyle only focus in pre-hypertension has limited effectiveness in the population.13,14 Perhaps in recognizing this reality, Healthy People 2020 set a more realistic goal of reducing prevalent hypertension to 26.9% for all race groups. Even this dilutional goal is unlikely to be obtained among African Americans, who had an age-adjusted hypertension prevalence of 37.9% in 2007–2008,8 without a safe, widely adopted, and effective pharmacological intervention for slowing the rapid progression from pre-hypertension to hypertension. Given the excess of prevalent hypertension and related complications in blacks,9 their accelerated transition from pre-hypertension to hypertension and greater risk of CVD complications even in the absence of progression,10 this is an opportune time for clinical effectiveness trials addressing these critical issues.
Public health efforts over the past 20 years, focused mainly on therapeutic lifestyle change, have not succeeded in reducing prevalent hypertension or racial differences in prevalent hypertension. Most hypertension progresses from the pre-hypertensive stage, but it was not known if blacks with pre-hypertension progress to hypertension faster than whites. Our study indicates that progression to hypertension is significantly faster in blacks than whites with pre-hypertension. These findings raise the possibility that effective interventions in pre-hypertension could reduce incident hypertension and associated racial differences with longer-term potential to reduce the magnitude of prevalent hypertension and its impact on health disparities.
SOURCES OF FUNDING
This work was supported by the State of South Carolina, United States Army W81XWH-10-2-0057, NIH Clinical Translational Science Award 1UL1RR029882, NIH HL091841, NIH DK067615, the American Society of Hypertension, and the South Carolina Department of Health and Environmental Control.
Anbesaw Selassie: None
C. Shaun Wagner: None
Marilyn L Laken: None
M. LaFrance Ferguson: None
Keith C. Ferdinand: Speakers' Bureau at AstraZeneca (<$10,000), Novartis (<$10,000), Forest (<$10,000), and Daiichi-Sankyo (<$10,000); honoraria from AstraZeneca (>$10,000), Novartis (>$10,000), Forest (>$10,000); Consultant/Advisory Board AstraZeneca (<$10,000), Novartis (<$10,000), and Forest (<$10,000).
Brent M. Egan. Grant support: Daiichi-Sankyo (>$50,000), Novartis (>$50,000), Takeda (>$50,000); Lecturer with honoraria on CME-accredited programs: American Society of Hypertension Carolinas-Georgia-Florida Chapter (>$10,000), International Society of Hypertension in Blacks (<$10,000); Consultant: NicOx (<$10,000).