Suboptimal Outcome of Management of Metabolic Cardiovascular Risk Factors in Hispanic Patients With Essential Hypertension
Abstract We studied outcome of management of metabolic cardiovascular risk factors in 155 randomly chosen Hispanic hypertensive patients (mean age, 63±1 years; 79% female) screened for dyslipidemia. Hypertriglyceridemia (n=12) or high risk–adjusted low-density lipoprotein cholesterol (LDL-C) (n=89) was found in 65%. Triglycerides did not change (6.16±0.58 to 7.44±2.34 mmol/L; P=NS) over 2.2±0.5 years. Only 58 patients with high LDL-C were treated, and 8 had no follow-up lipid tests. In the other 50, LDL-C decreased by 10±3% (P<.001) over 2.8±0.2 years but attained goal in only 12. In a subset of 24 patients with extended follow-up (3.8±0.2 years), there was an initial marked decline in LDL-C, followed by a rebound to baseline levels. In 29 of 54 patients with normal LDL-C, lipid testing was markedly overused compared with recommendations. Obesity (n=94, 61%) did not improve in those with repeated data (+0.6±0.8 kg; P=NS; n=40) over 2.7±0.3 years. Forty-four of 63 patients with type II diabetes had repeated measurement of glycosylated hemoglobin, with no change (10.5±0.5% to 11.2±0.5%; P=NS) over 2.2±0.3 years. Ten-year risk of coronary events (Framingham cohort parametric regression) calculated for 61 patients with known untreated blood pressures (169±3/98±1 mm Hg) was 21.0±1.7%, with a skewed distribution reaching high values (66%) and attributable in large part (72%) to modifiable risk factors. With the use of the model, hypothetical correction of metabolic risk factors was more powerful than reduction of systolic blood pressure to 140 mm Hg (Δrisk, −8.3±0.9% versus −4.2±0.5%; P<.0001) and predicted a reduction in risk 25-fold larger than that actually sustained by the patient sample (−0.3±0.6%; P=NS). In summary, Hispanics with essential hypertension exhibit a multifactorial risk profile, with the potential (as yet unrealized) for large reduction in cardiovascular morbidity. Our studies make it apparent that in this population, adequate treatment of concomitant metabolic illnesses is equally as important as or even more important than management of blood pressure.
Prevalence of hypertension in Hispanic Americans has been reported to be less than that of non-Hispanic whites,1 or intermediate between the rates of non-Hispanic whites and African Americans,2 and varying between subgroups of Hispanics.3 Despite this, hypertension carries a high morbidity in this minority, as exemplified by the incidence of end-stage renal disease of Mexican Americans, which is 2.5 times that of hypertensive non-Hispanic whites.4 While a multitude of possible socioeconomic characteristics, including language barrier, educational level, and insurance status, may contribute to this discrepancy,5 it is also known that Hispanics have a high prevalence of concomitant cardiovascular risk factors such as dyslipidemia, obesity, and diabetes.6 7 This multiple risk factor profile would be expected to contribute to the development of atherosclerotic disease and hence to the cardiovascular morbidity of Hispanic hypertensive patients.
The socioeconomic characteristics mentioned above might conceivably affect recognition, awareness, and treatment of the concomitant metabolic risk factors of Hispanic subjects with essential hypertension. We set out to explore this possibility by studying, in retrospective fashion, the management and outcome of these metabolic risk factors in a population of Hispanic patients cared for at the medical clinic of a teaching hospital.
We investigated traditional clinical markers of outcome, such as serum lipids in all patients, HbA1C in those with concomitant diabetes, and body weight in those with obesity. In addition, we applied the parametric regression model derived from the Framingham cohort8 to calculate the initial and final actual risks for coronary events in the sample, as well as the potential relative risk reductions that could be achieved by concentrating therapeutic efforts on treating blood pressure versus treating the other risk factors present in the sample.
This study was a retrospective review of data in a population sample consisting of Hispanic subjects with essential hypertension who had had at least one screening lipid profile (see below) over the previous 2 years. It took place at the general medical clinic of the Mount Sinai School of Medicine, New York, NY, a clinic supervised by full-time faculty of the Department of Medicine and predominantly staffed by residents in training. The medical records of this clinic had a uniform design, including standardized forms for listing diagnoses and medications and flowcharts for recording time-dependent clinical or laboratory variables. Data entry by house officers was supervised by full-time faculty.
The sample was randomly drawn from a list of 6601 chart numbers, obtained from the billing department, representing the active patients of the clinic. This instrument, with numbers that followed the original registration date of the patient at the Mount Sinai Hospital, was used to avoid the biases that may be introduced by drawing a patient sample from an alphabetical list.
Rough estimates of the prevalence of hypertension (60%) and the percentage of Hispanics (75%) in the patient population were available from previous surveys in the clinic. With these data, it was calculated that approximately 330 patients with lipid profiles would have to be identified to find 150 Hispanic hypertensives, the desired sample size. Therefore, every 20th chart number from the list was entered into the laboratory computer, in search of a lipid profile (directly measured total cholesterol, triglycerides, and HDL-C and calculated LDL-C). This computer stores test results of individual patients in an active (readable) file for 2 years. Lipid profiles were easily retrieved by use of a specific query, which avoided display and review of the multiple other data present for each patient. If results of at least one lipid profile were not found for the chart number entered, the following and preceding numbers in the list were tried sequentially until a lipid profile was identified.
Once 330 numbers were selected, these patients’ charts were reviewed for ethnic/racial information and for a diagnosis of hypertension. A patient was considered to be of Hispanic ethnic background if (1) he/she was identified as such, either on the form containing registration information or on the initial clinical note by the medical provider, or (2) if his/her surname was listed in the US census list of Hispanic surnames. The diagnosis of hypertension was either obtained from the chart list of chronic health problems or assumed to be present if three consecutive progress notes disclosed blood pressures >140/90 mm Hg or the medication records indicated administration of antihypertensive agents.
The charts of the 155 subjects with an available lipid profile, Hispanic ethnicity, and essential hypertension were reviewed in detail. Recorded demographic characteristics included age, sex, height, and weight. Concomitant cardiovascular risk factors included diabetes (if recorded on the list of chronic health problems, or if the patient was receiving treatment with insulin or oral hypoglycemic agents or had serum glucose >10.0 mmol/L), obesity (BMI ≥25.8 kg/m2 for women or ≥26.4 kg/m2 for men, or when recorded on the list of chronic health problems or progress notes if BMI was not verifiable because of missing height or weight data), smoking (considered to be present in current smokers and also in those who quit within the prior year), and family history of premature CHD (ie, definite MI or sudden death before age 55 in a first-degree relative).
Concomitant atherosclerotic diseases considered as risk factors for CHD in the 1987 report of the NCEP9 were also recorded. They included a personal history of cerebrovascular disease (listed in the chart as cerebrovascular accident, transient ischemic attack, or hemiplegia) and evidence for peripheral vascular disease (history of intermittent claudication or revascularization surgery, examination with absent or weak lower-extremity pulses or a bruit, and, less commonly, results of an angiogram).
The presence of established CHD was assessed by thorough documentation of a past MI in progress notes, unequivocal evidence for MI or ischemia on an ECG, focal left ventricular wall motion abnormalities reported in a nuclear cardiac imaging study or an echocardiogram, or unequivocal angina by history, supported by the finding of ischemic changes on an ECG or stress test.
Review of charts many times disclosed lipid profiles performed before the 2-year period initially screened with the computer search. All values, from the first one available, were recorded for each patient through the end of 1994. For the same period, data were collected on body weight in all subjects and HbA1C in diabetic patients. Evidence for secondary causes of dyslipidemia, such as hypothyroidism or nephrotic syndrome, was also noted.
Diet was considered an enacted treatment when there was clear indication in the progress notes that dietary advice had been provided or when referrals to a registered nurse or nutritionist had been made for such purpose. Information about cholesterol-lowering drugs was obtained from medication lists or notes.
Classification of dyslipidemias, goals of therapy, and adequacy of screening and follow-up were based on the criteria in the 1987 report of the NCEP.9 These criteria and not more recently published guidelines10 were used because the clinical activities monitored by this study were conducted while the former were in effect. It was the practice of the clinic to screen patients with fasting lipid profiles not random isolated total cholesterol values. Therefore, NCEP criteria for LDL-C, not total cholesterol, were used for comparison with data in our patients. In brief, LDL-C was considered normal if <3.36 mmol/L (130 mg/dL), in which case follow-up was deemed adequate and cost-effective if serum lipids were remeasured in 5 years. LDL was considered borderline if ≥3.36 mmol/L (130 mg/dL) but <4.14 mmol/L (160 mg/dL). For those without established CHD or two risk factors (see above), follow-up was deemed adequate if dietary advice was given and serum lipids were remeasured in 1 year. For those with established CHD and/or two risk factors, an adequate approach had to include evaluation for secondary causes of hyperlipidemia, defining a goal LDL-C (see below), and starting therapy. LDL-C was considered definitely abnormal if ≥4.14 mmol/L (160 mg/dL), regardless of the presence or absence of other risk factors, in which case defining a goal LDL-C value and initiating treatment were considered the standard of therapy. In judging outcome, 1987 NCEP criteria were also used: Goal LDL-C for both the borderline and high LDL-C groups was <4.14 mmol/L for patients with no CHD and less than two risk factors or <3.36 mmol/L otherwise. Analyses of LDL-C data did not include patients with triglyceride levels >4.52 mmol/L because calculated LDL-C is inaccurate in these circumstances. These patients were classified as hypertriglyceridemic, and results of the treatment of this metabolic problem were analyzed separately.
Calculations of risk for cardiovascular morbidity were made by use of the parametric regression model derived from the Framingham cohort.8 This model is of the form P=1−exp(−eu), where P is the probability of developing a CHD event and u=[log(time in years)−μ]/ς. In turn, μ and ς are sex-dependent and age-related functions of a multivariate regression of cardiovascular risk factors, including the continuous variables blood pressure (systolic or diastolic), cholesterol/HDL-C ratio, and age and the categorical variables smoking, diabetes, and LVH by ECG. The model was applied to a subset of 61 subjects whose blood pressures while not on antihypertensive therapy were known to the investigators. These patients’ ECGs were scored for LVH by Estes’ criteria; LVH was deemed definite if the score was ≥5.
The groups of variables used in the model differed depending on the specific purpose of the calculations.
First, we estimated the actual baseline risk of our population, using the values for untreated blood pressures and metabolic risk factors, current smoking status, and presence/absence of LVH by ECG.
Second, we calculated the relative contributions of all risk factors to total risk in the sample, by rerunning the model with one risk factor replaced by its “normal” or “corrected” value at a time. For example, the risk attributable to hyperlipidemia was calculated as the difference between the actual risk and that calculated by assigning a value of 4.44 (eg, 5.17/1.16 [200/45]) to the cholesterol/HDL-C ratio of all patients.
Third, we assessed the relative power of blood pressure reduction versus treatment of metabolic risk factors by subtracting from actual baseline risk (1) the risk after hypothetical reduction of blood pressure to 140 mm Hg systolic or 90 mm Hg diastolic, with all other risk factors remaining unmodified, and (2) the risk after hypothetical reduction of cholesterol/HDL-C ratio to 4.44, cessation of smoking, and achievement of long-term control of diabetes without change in baseline blood pressure. For the effect of controlling diabetes we used two different assumptions: (1) an HbA1C <7% becoming diabetes=“no” in the model, as if control of diabetes would remove 100% of its attributable risk, and (2) an HbA1C <7% removing 41% of the risk attributable to diabetes only. This second assumption is derived from the Diabetes Control and Complications Trial,11 in which the subjects receiving intensive therapy sustained a 41% reduction in macrovascular disease events by reducing HbA1C from 9% to approximately 7%. The two subtractions above were used as estimates of risk reduction as a result of controlling hypertension rather than controlling metabolic and behavioral risk factors.
Finally, we calculated the actual reduction in risk achieved by treatment of metabolic risk factors in our patients using their data from the beginning and end of the study.
Data are presented as mean±SEM. The significance of the changes in all variables, from the beginning to the end of the study, was assessed with paired Student’s t tests. These tests and correlation analyses were carried out with the use of a statistical package (SAS Institute). A value of P<.05 was used to reject the null hypothesis.
One hundred fifty-five Hispanic patients with essential hypertension, screened for the presence of dyslipidemia, were identified. This sample was characterized by a predominance of women (79%) and by an age range of 23 to 88 years, with a mean of 63±1 years. The high prevalence of other risk factors (diabetes mellitus, obesity, smoking, and family history of premature cardiovascular disease) and those of concomitant peripheral or cerebrovascular disease and established CHD are shown in Table 1⇓. In the 109 patients without CHD, the average number of cardiovascular risk factors (including hypertension) was three, with a range from one to six. Hypertension was the sole cardiovascular risk factor in only 8 patients. Therefore, the goal LDL-C in the entire population, except for these 8 subjects, was 3.36 mmol/L (130 mg/dL).
In 47 of the 94 patients with obesity, this diagnosis was obtained from statements in the charts (see “Methods”) but could not be verified by calculation of BMI because of missing weight or height data. In the remaining 47, BMI was 31.8±0.7 kg/m2 in women (n=40) and 31.5±1.0 in men, exceeding the conservative cutoffs for diagnosis (see “Methods”) by 23% and 19%, respectively. All 63 patients with diabetes mellitus had the type II, or adult-onset, form of the disease.
Thirteen patients (8.4%) were receiving replacement therapy for hypothyroidism. Two of them had normal lipid profiles, 1 exhibited hypertriglyceridemia, and the other 10 had abnormally high LDL-C levels. All these patients were euthyroid (ie, had normal levels of thyroid-stimulating hormone) over the period of study. Therefore, their dyslipidemia was considered primary (not secondary to hypothyroidism), and their lipid data were analyzed jointly with those of all other patients. Four patients (2.6%) had proteinuria in the nephrotic range (ie, >3 g/d). In 2 of them, 1 with primary and the other with diabetic glomerular disease, serum lipids were normal. The other 2 patients had diabetic nephropathy and high LDL-C levels. Their data were pooled with those of other patients for analysis, since their secondary dyslipidemia had an irreversible cause and required the same management as the primary form.
Lipid abnormalities were found in 101 patients, that is, 65% of the total screened population. Table 2⇓ shows that 12 patients had hypertriglyceridemia (5.96±0.50 mmol/L). Eight of them had diabetes, and 2 were obese. Only 10 patients had follow-up lipid data, which showed that triglycerides did not change significantly over 2.2±0.5 years (6.16±0.58 mmol/L to 7.44±2.34). Actually, an increase of 16±29% occurred despite diet (n=8) and pharmacological therapy (gemfibrozil in 6, lovastatin in 2). Changes in triglycerides correlated with changes in HbA1C in 7 diabetic patients (r=.76, P<.05) over the same period.
Eighty-nine patients (57%) had LDL-Cs above their risk-factor adjusted goal values (ie, 4.14 mmol/L in 4 and 3.36 in all others). More than one third of them (n=31) were not treated during the period of study. The remaining 58 patients received diet alone (n=19 [33%]) or in combination with cholestyramine (n=17 [29%]), gemfibrozil (n=12 [21%]), or 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (n=10 [17%]). Eight of these treated patients had no repeat lipid testing. Table 2⇑ shows that in 50 patients with follow-up lipids, a significant (10±3%) reduction of LDL-C (4.50±0.10 to 4.03±0.13 mmol/L) took place over 2.8±0.2 years. However, only 12 (24%) reached or exceeded goal LDL-C at the end of study. Fig 1⇓ shows that at 2.1±0.2 years of follow-up, LDL-C in all 50 patients with repeated lipid analysis exhibited a marked decrease (from 4.50±0.10 to 3.44±0.13 mmol/L; Δ=−1.09±0.16; P<.0001), but in 24 patients who had additional follow-up to 3.8±0.2 years this improvement was not sustained, exhibiting a rebound to 4.24±0.18 mmol/L, which was not significantly different from baseline.
Fifty-four patients (35%) had desirable risk factor−adjusted LDL-C levels. According to the different recommendations for those with normal (<3.36 mmol/L, n=50) and borderline (3.36<LDL<4.14 mmol/L, n=4) LDL-C values, the group should have had 14 repeated lipid tests over the 2.7±0.2 years of follow-up. Sixty-four were actually carried out. In 22 patients, the required 4 tests were appropriately obtained. In 3 patients who would have needed 5 lipid profiles, only 1 was done. Finally, in 29 subjects who needed only 5 tests, 59 were carried out, ie, an excess of 54.
Results of management of diabetes and obesity are also shown in Table 2⇑. Fifty-nine of the 63 diabetic patients had an initial measurement of HbA1C, with a mean of 10.5±0.4%. In 44 patients with repeated data, this parameter was not substantially modified (10.5±0.5% to 11.2±0.5%) over the course of 2.2±0.3 years. Similar to observations with LDL-C, 15 patients with multiple HbA1C measurements exhibited an initial significant response (12.2±0.7% to 9.2±0.7%; Δ, −3.0±0.9; P<.004) over 1.3±0.3 years, which was not sustained, rebounding to 12.1±1.0% at 2.5±0.4 years of follow-up (data not shown). Changes in HbA1C did not correlate with simultaneous changes in weight of diabetic patients. Finally, of the 94 obese patients, 54 had no follow-up weight data. The remaining 40 sustained a nonsignificant 0.4±0.9% weight gain (78±2 to 79±2 kg) over 2.7±0.3 years. Again, in 9 patients with multiple data, a significant initial weight loss (77±3 to 73±4 kg; Δ, −4±1 kg; P<.002) over 1.9±0.5 years disappeared at 4.1±0.4 years of follow-up (final weight, 79±4 kg).
The risk of developing a coronary event over 10 years, in 61 patients with known untreated blood pressures (169±3/98±1 mm Hg) and ECGs (LVH, n=7), was 21.0±1.7% or 20.8±1.8% (calculated with systolic or diastolic values, respectively). The frequency distribution of this coronary risk was skewed to the right, reaching values as high as 66% in some patients. Despite the fact that our population was relatively elderly, most of its attributable risk (72% of the total) was due to the six modifiable risk factors that participate in the model. With their relative prevalences in our population, we calculated that diabetes and hyperlipidemia (cholesterol/HDL-C ratio) each contributed more than a quarter of total modifiable risk. Hypertension followed, contributing either the same magnitude or somewhat less than a fifth, depending on the use of systolic or diastolic values in the calculations. Finally, LVH contributed one eighth and smoking one sixteenth of modifiable risk; these lesser magnitudes reflected their lower prevalences in the population.
Hypothetical correction of hypertension (see “Methods”), without modification of other risk factors, diminished overall risk from 21.0±1.7% to 16.8±1.5% (with systolic pressure reduced to 140 mm Hg) or from 20.8±1.8% to 18.8±1.7% (with diastolic pressure reduced to 90 mm Hg). In contrast, hypothetical correction of diabetes, dyslipidemia, and smoking (with unmodified baseline blood pressures) reduced the former risk values to 12.7±1.2% or 12.3±1.3%. Fig 2⇓ shows the data for calculations with systolic blood pressure, indicating that the effect of correcting metabolic and behavioral risk factors (−8.3±0.9%) was significantly greater than that of correcting hypertension (−4.2±0.5%). With the use of the assumption that HbA1C <7% confers only 41% reduction in the risk attributable to diabetes (see “Methods”), the respective values were −6.3±0.7% versus −4.2±0.5% (P<.02). Calculations performed with diastolic blood pressures produced analogous results (−8.5±0.9% versus −2.0±0.2%; P<.0001).
Finally, the model was used to calculate the actual risk reduction sustained by this group of patients as a consequence of treatment of their metabolic risk factors (without the use of changes in blood pressure). This was −0.3±0.6% ( P=NS) for calculations with systolic blood pressure and −0.2±0.6 (P=NS) for calculations with diastolic values. These values, 25-fold smaller than the maximal theoretical risk reduction calculated above, provide a quantitative assessment of the outcome of management of metabolic risk factors in Hispanic patients with essential hypertension.
Ischemic heart disease12 and stroke13 are still the leading and third causes of death of Hispanic Americans, respectively. Also, incidence of hypertensive end-stage renal disease in this group is 2.5 times higher than in non-Hispanic whites.4 It is therefore apparent that Hispanics have not fully benefited from the overall decline of cardiovascular mortality that has occurred in the United States over the last two decades.14 15 Excess hypertensive target organ damage of Hispanics resembles that of African Americans, which is paradoxical because the prevalence of hypertension in Hispanics is much lower, falling in the range of1 or lower than7 that of non-Hispanic whites.
Most explanations for the disparity between the low prevalence and adverse outcomes of hypertension in Hispanics have invoked socioeconomic and cultural disadvantages that affect their access to health care. These include cultural isolation and language barrier, lack of public awareness, misperception of availability of resources, lack of actual resources (poverty), low educational level, and unemployment or uninsurability.16 Others have speculated that these same conditions predispose Hispanics to concomitant cardiovascular risk factors such as obesity or excess alcohol intake.17 Although as a group Hispanics have the lowest awareness rates for hypertension,1 this is not sufficient to account for their prevalence of hypertensive complications.
It is conceivable that associated risk factors for cardiovascular disease are major contributors to morbidity in Hispanics. Mexican Americans in the Southwest have high prevalences of diabetes, obesity, and insulin resistance, a metabolic pattern that may reflect admixture with a gene pool from American Indians.18 Dyslipidemias, which develop at an early age19 and are characterized by hypertriglyceridemia and low HDL-C levels,6 are also common in Hispanics.
Much less is known about Hispanics from New York, who are predominantly of Caribbean origin (Puerto Rico and the Dominican Republic) as were our patients. They exhibit less knowledge about cardiovascular risk factors and higher prevalence of sedentary lifestyle compared with whites from New York.20 Also, the women exhibit higher rates of obesity,20 whereas Puerto Rican men have the highest rates of current smoking when compared with other groups of Hispanics.21 We hypothesized that our Caribbean Hispanic patients also have a multifactorial risk profile and investigated whether this is managed adequately to decrease their cardiovascular morbidity. The Fifth Report of the Joint National Committee22 recommends assessing absolute risk by means of numerical estimates derived from epidemiological studies. We used the parametric regression model derived from the Framingham cohort8 because Framingham models have been previously shown to be good predictors of cardiovascular outcomes in populations dissimilar to that from which they were derived.23
We found high rates of metabolic risk factors and established atherosclerotic complications in our patients. However, we cannot speculate about prevalence of these events, because our random sample was drawn from subjects already seeking medical care not the general population. Also, included patients had already been screened for hyperlipidemia, which probably selected a sicker subset. Diabetes and obesity, however, were not criteria for inclusion in the study. Therefore, their respective rates of 41% and 61% may be indicators of high prevalence in the population. A striking observation in our study was that the most common lipid abnormality (57%) was high risk-adjusted levels of LDL-C.
Because of the multiplicity and large prevalence of metabolic abnormalities, it is not surprising that our calculations disclosed high overall risk and also a larger (twofold) risk reduction potential for treatment of metabolic factors than for management of blood pressure. In view of this, understanding the causes for the dismal outcome of treatment of metabolic abnormalities in our patients acquires the utmost importance.
We can only speculate about these causes. There may be factors inherent to the population, including known unhealthy purchase patterns of processed foods17 and salt24 in Hispanics. Our observations regarding rebound of initially improved LDL-C, weight, and HbA1C levels bespeak the recognized difficulty in sustaining long-term behavioral modification. Rebound of cholesterol levels25 and high rates of discontinuation of antihyperlipidemic drugs in nonresearch settings26 have both been reported.
Factors pertaining to the medical care setting may also play a role. It has been argued that the scanty number of Hispanic health providers in the United States may be detrimental to culture-sensitive delivery of medical care to Hispanic patients.16 Independent of ethnicity, physicians may be prone to neglect therapy when medical conditions have an extremely poor prognosis, eg, obesity.27 It is noteworthy that in addition to low success rates in treatment, our clinic was deficient in obtaining the data (ie, body weight, HbA1C, and serum lipids) required for diagnosis and follow-up of metabolic problems. Underscreening for high cholesterol may affect as much as two thirds of the US population,28 and suboptimal follow-up of HbA1C has been shown in a large population of elderly diabetics.29
In conclusion, our data suggest that to realize the potential for reduction of cardiovascular morbidity in Hispanic hypertensives, a major educational effort must be undertaken to improve management of their metabolic risk factors. This could be modeled on the programs that have improved awareness, treatment, and control of hypertension. Such an effort may be particularly important in medical settings manned by house officers, in which attempted educational maneuvers to increase compliance with NCEP recommendations have, to date, shown only marginal benefits.30
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|CHD||=||coronary heart disease|
|HDL-C||=||high-density lipoprotein cholesterol|
|LDL-C||=||low-density lipoprotein cholesterol|
|LVH||=||left ventricular hypertrophy|
|NCEP||=||National Cholesterol Education Program|
Reprint requests to Cheryl L. Laffer, MD, PhD, 4.156 John Sealy Hospital, University of Texas Medical Branch at Galveston, 301 University Blvd, Galveston, TX 77555-0572.
- Received June 17, 1995.
- Revision received September 16, 1995.
- Accepted October 6, 1995.
Burt VL, Whelton P, Roccella EJ, Brown C, Cutler JA, Higgins M, Horan MJ, Labarthe D. Prevalence of hypertension in the US adult population: results from the Third National Health and Nutrition Examination Survey, 1988-1991. Hypertension. 1995;25:305-313.
Pugh JA, Stern MP, Haffner SM, Eifler CW, Zapata M. Excess incidence of treatment of end-stage renal disease in Mexican Americans. Am J Epidemiol. 1988;127:135-144.
Haffner SM, Mitchell BD, Stern MP, Hazuda HP, Patterson JK. Decreased prevalence of hypertension in Mexican-Americans. Hypertension. 1990;16:225-232.
Anderson KM, Wilson PWF, Odell PM, Kannel WB. An updated coronary risk profile: a statement for health professsionals. Circulation. 1991;83:356-362.
National Cholesterol Education Program. Highlights of the Report of the Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults. Bethesda, Md: National Institutes of Health; 1987. Publication NIH 88-2926.
Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Summary of the Second Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel II). JAMA. 1993;269:3015-3023.
Yatsu FM. Strokes in Asians and Pacific-Islanders, Hispanics, and Native Americans. Circulation. 1991;83:1471-1472.
Francis CK. Hypertension and cardiac disease in minorities. Am J Med. 1990;88:3S-8S.
Caralis PV. Coronary artery disease in Hispanic Americans: how does ethnic background affect risk factors and mortality rates? Postgrad Med. 1992;91:179-193.
Samet JM, Coultas DB, Howard CA, Skipper BJ, Hanis CL. Diabetes, gallbladder disease, obesity, and hypertension among Hispanics in New Mexico. Am J Epidemiol. 1988;128:1302-1311.
Shea S, Stein AD, Basch CE, Lantigua R, Maylahn C, Strogatz DS, Novick L. Independent associations of educational attainment and ethnicity with behavioral risk factors for cardiovascular disease. Am J Epidemiol. 1991;134:567-582.
Kerr GR, Amante P, Decker M, Callen PW. Ethnic patterns of salt purchase in Houston, Texas. Am J Epidemiol. 1982;115:906-916.
National Diet-Heart Study Research Group. The National Diet-Heart Study Final Report. Circulation. 1968;37(suppl I):I-1-I-428.