Prevalence, treatment, and control of dyslipidemia in hypertensive subjects. Stacked bars are percentages of all hypertensive subjects. Asterisk indicates P<.05 and dagger indicates P<.001 (comparison between sexes within ethnic groups); double dagger indicates P<.01 and section mark indicates P<.001 (comparison between ethnic groups within sex groups).
Distribution of dyslipidemia subtypes. Subjects who remained dyslipidemic, despite treatment with lipid-regulating drugs, were pooled with those who were untreated, except among white women, in whom the dyslipidemia distributions differed significantly. The distributions also differed significantly between sexes in each ethnic group and between ethnic groups in each sex (see text for P values of ethnic-sex comparisons). HDL-C indicates high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; and TG, triglycerides.
O'Meara JG, Kardia SLR, Armon JJ, Brown CA, Boerwinkle E, Turner ST. Ethnic and Sex Differences in the Prevalence, Treatment, and Control of Dyslipidemia Among Hypertensive Adults in the GENOA Study. Arch Intern Med. 2004;164(12):1313-1318. doi:10.1001/archinte.164.12.1313
Copyright 2004 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2004
Two ethnically different, community-based samples of hypertensive adults were evaluated to determine the prevalence of dyslipidemia and how often dyslipidemia is drug-treated and controlled by such treatment.
We studied 1286 non-Hispanic black hypertensive subjects from Jackson and 1070 non-Hispanic white hypertensive subjects from Rochester who participated in the Genetic Epidemiology Network of Arteriopathy study. Subjects were categorized according to presence of coronary heart disease and risk factors for coronary heart disease.
Prevalence of dyslipidemia was significantly greater among whites than blacks (women, 64.7% vs 49.5%; and men, 78.4% vs 56.7%; P<.001 for both) and among men than women (P≤.02 in each ethnic group). Among dyslipidemic subjects, treatment with lipid-regulating drugs was significantly more common among whites than blacks (women, 25.4% vs 16.4%, P = .001; and men, 32.6% vs 12.8%; P<.001), and among whites, treatment was significantly more common among men than women (P = .03). With drug treatment, control of dyslipidemia varied from 33.9% (white men) to 51.9% (black men), but the differences among ethnic-sex groups were not statistically significant.
Dyslipidemia is highly prevalent in hypertensive adults. Fewer than one third of these adults are drug-treated, and fewer than half of those treated achieve recommended goals. Our findings suggest that an alarming 9 of 10 dyslipidemic hypertensive adults have untreated or undertreated dyslipidemia.
Dyslipidemia and hypertension are 2 of several modifiable risk factors for cardiovascular disease morbidity and mortality. Dyslipidemia is reported in 50% to 80% of hypertensive patients.1- 5 The co-occurrence of dyslipidemia and hypertension increases the risk of coronary heart disease (CHD) more than the sum of the risks associated with these component factors occurring alone.4,6 Medication directed at regulating lipids in hypertensive subjects has been shown to significantly lower the risk of stroke and myocardial infarction.7 Moreover, normalization of lipid levels in dyslipidemic patients with treated but uncontrolled hypertension may reduce the risk of CHD more than normalization of blood pressure.8
Despite evidence that the incidence and prevalence of CHD differ by ethnicity and sex, there are limited data regarding ethnic and sex differences in the prevalence of dyslipidemia among hypertensive subjects.3,5 Furthermore, to our knowledge, there are no studies characterizing the treatment and control of dyslipidemia among hypertensive subjects as a group or by ethnic and sex subgroups. Therefore, we examined 2 ethnically different samples of hypertensive adults (non-Hispanic blacks and whites) enrolled in the Genetic Epidemiology Network of Arteriopathy (GENOA).9 Using guidelines recommended by the National Cholesterol Education Program II (NCEP II),10 we sought to determine and compare among ethnic and sex groups the prevalence of dyslipidemia, how frequently lipid-regulating drugs (LRDs) were prescribed, and how frequently dyslipidemia was controlled among those treated with LRDs.
The samples consisted of 1286 non-Hispanic black hypertensive subjects (914 women and 372 men) recruited from Jackson and 1070 non-Hispanic white hypertensive subjects (584 women and 486 men) recruited from Rochester between 1995 and 1998 to participate in the GENOA study of the National Heart, Lung, and Blood Institute Family Blood Pressure Program.9 The GENOA participants in Jackson and Rochester were ascertained through sibships (ie, full biological offspring of the same mother and father) having 2 or more members diagnosed as having essential hypertension before age 60 years. All available members of the sibships were invited to participate in study visits; however, only participants classified as having hypertension were included in the analyses. Black subjects were from 683 separate sibships, of which 101 included 1 individual, 280 included 2 individuals, and 302 included 3 or more individuals. White subjects were from 570 separate sibships, of which 19 included 1 individual, 318 included 2 individuals, and 233 included 3 or more individuals. The study protocol was approved by the human studies committees of the University of Mississippi and Mayo Clinic and Foundation. All subjects signed a consent form.
Study visits were conducted between 7:00 and 9:00 AM. Trained examiners administered standardized questionnaires to elicit subjects' personal and family medical histories and use of prescription medications in the previous month. Subjects then underwent measurements of height (by wall stadiometer), weight (by electronic balance), waist and hip circumferences (by tape measure), and blood pressure (by random zero sphygmomanometer). Three blood pressure readings were obtained in each subject's right arm with an appropriate-sized cuff after the subject had been seated for at least 5 minutes. The means of the last 2 readings were used in the analyses. Blood samples were drawn (fasting) into ethylenediaminetetraacetic acid–containing Vacutainers (Becton Dickinson, Franklin Lakes, NJ) for determination of plasma concentrations of glucose, creatinine, total cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides in the Mayo General Clinical Research Center Immunochemical Core Laboratory, which is Clinical Laboratory Improvement Amendments–accredited and participates in the lipid standardization program of the Centers for Disease Control and Prevention, Atlanta, Ga.
Prescription medications were coded using a system based on the Medi-Span medical therapeutic classification,11 and LRDs were classified into 4 pharmacologically distinct categories: 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors (statins), bile acid–binding resins, fibric acid derivatives, and nicotinic acid. The diagnosis of hypertension was based on subjects' reporting a prior diagnosis of hypertension and current use of prescription antihypertensive medications, or systolic blood pressure of 140 mm Hg or higher or diastolic blood pressure of 90 mm Hg or higher measured at the study visit. The diagnosis of diabetes mellitus was based on subjects' reporting a prior diagnosis of diabetes mellitus or high blood sugar and current use of prescription antidiabetic medication, or fasting plasma glucose concentration of 126 mg/dL (7.0 mmol/L) or higher measured at the study visit. The diagnosis of CHD was based on subjects' reporting of a previous myocardial infarction, coronary artery bypass surgery, or angioplasty.12
Based on NCEP II guidelines, subjects were scored for the following risk factors for CHD: male sex and age 45 years or older or female sex and age 55 years or older, hypertension (which all subjects had), diabetes mellitus, current cigarette smoking, HDL cholesterol less than 35 mg/dL (<0.91 mmol/L), and family history of premature CHD in a first-degree male or female relative younger than 60 years. If HDL cholesterol was 60 mg/dL or greater (≥1.55 mmol/L), 1 risk factor was subtracted. Subjects were then classified into 3 categories of CHD risk: history of CHD, no history of CHD but at least 2 risk factors, or no history of CHD and fewer than 2 risk factors.
Dyslipidemia was considered present if 1 or more of the criteria in Table 1 were satisfied. Dyslipidemic subjects not treated with LRDs and those in whom dyslipidemia was treated but not controlled were further characterized as having elevated low-density lipoprotein (LDL) cholesterol (with or without low HDL cholesterol or elevated triglycerides), elevated triglycerides only (without elevated LDL or low HDL cholesterol), low HDL cholesterol and elevated triglycerides (without elevated LDL cholesterol), or low HDL cholesterol only (without elevated LDL cholesterol or triglycerides).
Data were summarized by means ± SDs for quantitative traits and by percentages for categorical traits. To assess differences between sexes within each ethnic group, and between ethnic groups within each sex, unpaired t tests were used for group means and χ2 tests of independence for percentages. Fisher exact test was used when cell sizes contained fewer than 5 subjects. Statistical analyses were performed with SAS version 8.0 (SAS Institute, Cary, NC). Test statistics were considered statistically significant at P<.05. Because the subjects were members of sibships (ie, not independent observations), generalized estimating equations were also used to evaluate the effect of familial correlations on estimates of the prevalence, treatment, and control of dyslipidemia. However, because these additional analyses did not alter the inferences, the results are not presented.
In both sexes, the mean age was significantly greater in black than white subjects (women, 59.8 ± 9.4 vs 57.8 ± 10.0 years; and men, 60.8 ± 9.5 vs 57.4 ± 10.1 years; P<.001 for both). Except for LDL cholesterol, mean values and percentages for the other descriptive characteristics differed significantly between sexes within 1 or both ethnic groups, or between ethnic groups within 1 or both sexes (Table 2).
More than 49% of the subjects met the study definition of having dyslipidemia (Figure 1 and Table 3). In both ethnic groups, the percentage of dyslipidemic subjects was significantly greater among men than women (blacks, 56.7% vs 49.5%; P = .02; and whites, 78.4% vs 64.7%; P<.001), and in each sex, dyslipidemia was significantly more common among whites than blacks (P<.001). The latter ethnic differences in dyslipidemia prevalence were observed in each CHD risk category (Table 3). In each ethnic-sex group, the prevalence of dyslipidemia differed significantly among CHD risk categories, with the highest prevalences among subjects reporting CHD.
Fewer than 33% of the dyslipidemic hypertensive subjects reported treatment with LRDs (Table 3), ie, fewer than 26% of all hypertensive subjects (Figure 1). In whites, the percentage of dyslipidemic subjects treated with LRDs was significantly greater among men than women (P = .03), and in both sexes, a greater percentage of white than black dyslipidemic subjects was treated with LRDs (women, 25.4% vs 16.4%; P = .001; and men, 32.5% vs 12.8%; P<.001). The difference in relative frequency of treatment in white vs black subjects was most pronounced among those with CHD (women, 69.0% vs 34.3%; P = .006; and men, 73.0% vs 18.9%; P<.001) (Table 3). In white women and men, the percentage of subjects treated with LRDs differed significantly among CHD risk categories (P<.001), with the highest percentages treated among those with CHD.
More than 94% of dyslipidemic subjects on LRD treatment were receiving monodrug therapy (range, 91.9% in white men to 98.6% in black women), with more than 83% of those on monotherapy receiving a statin (range, 83.6% in black women to 88.5% in black men) and 100% of those on combination regimens receiving a statin (data not shown). The low relative frequency of combination therapy precluded statistical comparisons of the monodrug therapy vs combination therapy percentages.
Fewer than 52% of the treated dyslipidemic hypertensive subjects achieved lipid levels consistent with NCEP II guidelines (Table 3), ie, fewer than 9% of all hypertensive subjects (Figure 1). Differences in these percentages among ethnic-sex groups were not statistically significant. In white women and men, the percentage of LRD-treated subjects achieving NCEP II goals differed significantly among CHD risk categories (P≤.01), with the highest control percentages among those without CHD or additional risk factors (Table 3).
In white women, the distribution of dyslipidemia subtypes differed significantly between those who reported no LRD therapy and those who remained dyslipidemic, despite drug treatment (P = .02). Because this was not the case in the other ethnic-sex groups (data not shown), drug-treated and untreated dyslipidemic subjects were pooled, whereas characteristics of treated and untreated white women are shown separately in Figure 2. Elevation of LDL cholesterol was the most prevalent form of dyslipidemia, except among LRD-treated white women, in whom isolated triglyceride elevation was the most prevalent dyslipidemia. The distribution of dyslipidemia subtypes differed significantly between sexes (P<.001 in each ethnic group) and between ethnic groups (P<.001 in each sex). Isolated elevation of triglycerides was more common among women than men and more common among whites than blacks, whereas isolated low HDL cholesterol was more common among men than women and more common among blacks than whites (Figure 2).
In addition to confirming that dyslipidemia is frequent among hypertensive adults, our observations suggest that small numbers of these adults in the community are receiving LRD therapy and approximately one half or fewer of those treated achieve control of dyslipidemia as defined by the NCEP II guidelines. Because 91.1% of dyslipidemic hypertensive subjects sampled had untreated or undertreated dyslipidemia, our findings suggest a serious deficiency in the management of dyslipidemia among individuals with hypertension in the community.
Our finding of a higher prevalence of dyslipidemia among white vs black subjects is consistent with previous comparisons between these ethnic groups in the US population overall.14,15 Our observation of a higher prevalence of dyslipidemia among men than women is also in general agreement with previous reports of sex differences in mean values for LDL cholesterol, HDL cholesterol, and triglycerides among hypertensive subjects.5,16- 18 The higher percentage of men than women with CHD (Table 2) most likely accounts for the observed sex differences in the prevalence of dyslipidemia (Figure 1), because within risk factor categories, the prevalence of dyslipidemia did not differ significantly between women and men (Table 3). Consistent with findings in another sample, subjects with preexisting CHD had the highest prevalence of dyslipidemia.5 This may be explained by lower cut points for LDL cholesterol and the fact that dyslipidemia itself contributes to CHD.
We are unaware of any previous studies of hypertensive subjects comparing LRD treatment between ethnic and sex groups. Our finding of more frequent treatment among white than black subjects is similar to the ethnic differences noted in the relative frequency of antihypertensive drug treatment among the GENOA cohorts.19 In addition, the relative frequency of LRD treatment and adherence to drug therapy are reported to be lower among black than white subjects.20- 25 These findings provide further evidence of suboptimal management of risk factors for cardiovascular disease among blacks vs whites. Among white subjects, treatment with LRDs was more common among men than women, similar to a previous investigation in patients with CHD.26 This difference may reflect sex bias in the application of therapies for cardiovascular conditions.27,28 Although white women had isolated triglyceride elevation more often than men—which may constitute a less compelling rationale for LRD treatment than elevated LDL cholesterol—this difference in dyslipidemia subtypes (Figure 2) did not account for the lower relative frequency of LRD treatment among white women (data not shown).
Control of dyslipidemia did not vary significantly across ethnic-sex groups; however, the relative frequency of control tended to be higher among LRD-treated blacks than whites. A similar trend was observed by Nieto and colleagues23 in samples from the US population that included 35% hypertensive subjects. Conversely, Pearson and colleagues29 described more frequent control of LDL cholesterol in drug-treated white than black subjects; however, their sample included only a small percentage of blacks (5%), and 45% of subjects were normotensive. Although Bramlet and colleagues30 reported that a higher percentage of LRD-treated men than women achieved NCEP II goals, their sample consisted of patients with CHD and included small percentages of blacks and women. Except in white women, dyslipidemia control was lowest in subjects with CHD. This may reflect the large reductions in LDL cholesterol (ie, >30%) often required in those with CHD31 and the infrequent use of combination LRD therapy. Our finding that only a small percentage of hypertensive subjects treated for dyslipidemia (<5%) was taking any combination of LRDs is consistent with previous observations in general population samples.29,32
Taking into account the sibship structure of our sample did not change the conclusions from our analyses. We found statistically significant evidence of sib-sib concordance in the presence of dyslipidemia among hypertensive siblings (Mantel-Haenszel odds ratio, 2.2; 95% confidence interval, 1.7-2.9 in blacks; and odds ratio, 1.7; 95% confidence interval, 1.2-2.4 in whites), but not in treatment or control with LRDs (data not shown). Inasmuch as the descriptive characteristics of GENOA subjects appear similar to those of other community-based samples of hypertensive adults, estimated prevalences of dyslipidemia and their ethnic and sex differences are likely to be generalizable. However, because hypertension treatment and control were more frequent in GENOA participants than in representative samples of the corresponding ethnic-sex subgroups of the US population,19 it is also reasonable to suggest that LRD treatment and, therefore, overall control of dyslipidemia among hypertensive adults in the community may be less than estimates from GENOA.
Given the high prevalence of dyslipidemia among hypertensive blacks and whites, the low relative frequencies of treatment with LRDs and control of dyslipidemia represent opportunities to improve the overall management of individuals with hypertension and thereby reduce morbidity and mortality from atherosclerotic complications. Additional studies are imperative to better understand reasons for underuse of LRD therapy and to develop more effective methods of evaluation, treatment, and control of dyslipidemia among individuals with hypertension.
Corresponding author: Stephen T. Turner, MD, Division of Hypertension, Department of Internal Medicine, Mayo Clinic and Foundation, 200 First St SW, Rochester, MN 55905 (e-mail: email@example.com).
Acccepted for publication August 19, 2003.
This work was supported by grants U01 HL54464, U01 HL54457, U01 HL54481, R01 HL53330, and R01 HL68737 from the US Public Health Service, Rockville, Md, and by funds from the Mayo Clinic and Foundation.
This study was presented as an abstract at the Fourth Scientific Forum on Quality of Care and Outcomes Research in Cardiovascular Disease and Stroke, American Heart Association; October 13, 2002; Washington, DC.
We appreciate assistance from the Mayo General Clinical Research Center and thank Tracy Fuller for assistance in the preparation of the manuscript.