Hypertension and other chronic diseases are becoming increasingly important health problems for many Native American people, including the Navajo. A community-based survey that included three standardized measurements of blood pressures, was conducted during 1991-92 on the Navajo Reservation. Among the 780 adults examined, the overall age-standardized prevalence of hypertension, defined as an elevated systolic (
140 mm Hg) or diastolic (90 mm Hg) blood pressure, or possession of prescription antihypertensive medications, was 19% (24% among men and 15% among women). The prevalence of hypertension increased with age and relative weight, and among men, was associated with diabetes mellitus. Among women, hypertension was associated with a central distribution of body fat, cigarette smoking, self-reported diabetes mellitus and impaired glucose tolerance. Although only 50% of the persons found to have elevated blood pressure at the examination reported they had been previously told that they had hypertension, persons who had been previously diagnosed with hypertension had a slightly higher rate (~60%) of blood pressure control than that seen in the general U.S. population. On the basis of these results, the prevalence of hypertension among the Navajo appears to have substantially increased since the 1930s. Improved prevention and management of hypertension, especially for overweight and diabetic individuals, may reduce morbidity and mortality from cardiovascular and renal disease.
KEY WORDS:
hypertension ·
blood pressure ·
American Indians ·
diabetes ·
body weight
Chronic diseases, including diabetes mellitus, heart disease and cancer, have become increasingly important health problems for the Navajo people over the last 50 years (Coulehan et al. 1986
, Fulmer and Roberts 1963, Sugarman et al. 1990). Although considerable attention has been directed to the problems of diabetes and its complications among American Indians, comparatively little information is available concerning the prevalence and effect of hypertension (Alfred 1970, Clifford et al. 1963, DeStefano et al. 1979, Kunitz and Levy 1986, Salsbury 1937, Sievers 1977, Sugarman 1990, Welty and Coulehan 1993, Young 1991). Hypertension is an important risk factor for coronary heart disease, cerebrovascular disease and renal failure, and the risk for these diseases can be reduced by lowering blood pressure levels (Kaplan 1990). Furthermore, for patients with established diabetic nephropathy, blood pressure control is important in preserving kidney function (Klahr et al. 1994).
The prevalence of hypertension in the Navajo Nation was low before the 1960s (Fulmer and Roberts 1963, Salsbury 1937), but on the basis of a study in a single service unit (DeStefano et al. 1979), rose substantially by the 1970s. In this report, we examine the prevalence of hypertension by using information gathered during the first population-based survey of the health status of the Navajo.
MATERIALS AND METHODS 
Sample.
The survey design and methods of the Navajo Health and Nutrition Survey are described in detail elsewhere in this issue (White et al. 1997b). Briefly, a three-stage cluster design was used to select a representative sample of adolescent and adult residents from households in each of the eight Indian Health Service units serving the Navajo Nation. A total of 985 subjects were examined from October 1991 to December 1992, and 788 nonpregnant adults aged 20 through 91 were considered for these analyses.
Blood pressure measurements.
After the participant had been seated for at least 10 min, blood pressure measurements were made with a random-zero sphygmomanometer with appropriate-size cuff (Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure 1988). Three measurements were taken, and the mean of the second and third blood pressure measurements was used in the analyses for most participants. These blood pressure measurements were moderately repeatable, with CV of 5% (systolic) and 8% (diastolic); intraclass correlation coefficients were 0.80 (systolic) and 0.73 (diastolic). For the 23 persons who had only two systolic or diastolic measurements, the mean of these measurements was used; 8 persons with fewer than two measurements were excluded, resulting in a total of 780 persons (301 men and 479 women).
Following the criteria endorsed by the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure (1993), hypertension was defined as a mean of two systolic blood pressures of 140 mm Hg or more, a mean of two diastolic blood pressures of 90 mm Hg or more, or currently having a prescription for antihypertensive medication. Participants were asked to bring containers of all medications to the examination, and these samples were examined by the interviewer to determine if antihypertensive medication had been prescribed.
Other characteristics.
Anthropometric measurements taken during the survey included height, weight, triceps and subscapular skinfolds; all were measured by using the standard protocols of the Second National Health and Nutrition Examination Survey (McDowell et al. 1985) and have been described (White et al. 1997a). The mean of two replicate measurements was used in the analysis. Height was measured without shoes, body weight was measured in light clothing, and skinfolds were measured to the nearest 0.5 mm with Lange skinfold calipers (Cambridge Scientific Instruments, Cambridge, MD). Body mass index (BMI), an index of relative weight, was computed as weight(kg)/[height(m)]2, and the ratio of subscapular to triceps skinfolds was used as a measure of central (truncal) obesity.
Tobacco use (smoking cigarettes and chewing tobacco) was self-reported as current, ever or never; no biochemical markers were used to validate reported tobacco use. Persons were considered to be sedentary if they reported not having participated in any activity for exercise, including running, basketball, softball, aerobics, walking, hunting, getting wood and swimming, during the previous month.
Individuals were classified as having diabetes if they had a confirmed medical diagnosis or if they met the WHO definition for noninsulin-dependent diabetes mellitus on the basis of a fasting glucose level 140 mg/dL4 or, after an oral glucose tolerance test (Will et al. 1997), a 2-h level >200 mg/dL; persons with a normal fasting level and a 2-h level between 140 and 199 mg/dL were considered to have impaired glucose tolerance.
Statistical methods.
All proportions, means and correlations were calculated by using sample weights (White et al. 1997a); because of the sampling design, SUDAAN (Shah 1991) was used to estimate standard errors. Locally weighted scatterplot smoother curves, which are resistant to outliers and rely only on the data to specify the form of the model (Cleveland 1979), were used to examine the relation of blood pressure to age. Associations between hypertension status or blood pressures and age, diabetes, BMI, and smoking were examined using 
2 analysis or regression techniques. Because of the strong, and often nonlinear associations observed with age, most presented values were either age-standardized (by using the 1990 U.S. population) or age-adjusted by using regression techniques including both linear and quadratic terms for age. Because of the small sample sizes in some cross-tabulations, exact mid-P values based on Fisher's exact test have been calculated (Mehta and Patel 1995).
RESULTS
Blood pressure measurements were obtained for 780 nonpregnant adults (Table 1), and the unadjusted prevalence of hypertension was 17% (23% among men; 14% among women). Of the 136 persons with hypertension, 24 (18%) had an elevated blood pressure at the examination despite having a prescription for antihypertensive medications, 73 (54%) had an elevated blood pressure (only), and 39 (29%) were using antihypertensive medications (only). The prevalence of hypertension increased markedly with age (e.g., among women, from 5 to 28% across the three age groups), but age-standardized estimates (19% overall, 24% among women, 16% among men) were similar to the unadjusted values. Age-related increases were also seen in levels of systolic blood pressure, which were greater among women, and in the use of antihypertensive medications. Although mean diastolic blood pressures reached a maximum at about 40-50 y of age (Fig. 1), additional regression analyses (incorporating natural splines for both age and BMI) indicated that these curvilinear associations were, in part, due to the relatively low relative weight of persons 60 years of age, data not shown.
|
Table 1.
Levels of various characteristics by age group, Navajo Health and Nutrition Survey, 1991-92
[View Table]
|
Fig. 1.
Scatterplots and locally weighted scatter plot smoother (LOWESS) curves showing the relation of age to levels of systolic blood pressure (top) and diastolic blood pressure (bottom); separate plots are shown for men (left) and women (right) in the Navaho Health and Nutrition Survey, 1991-92. The horizontal lines at 90 or 140 mm Hg represent the cutpoints for hypertension. Each point along the fitted LOWESS curve represents the predicted level on the basis of a weighted least-squares regression using 67% of all data (see Materials and Methods for additional information). Because of the large number of subjects, one half were randomly selected to be shown (as open circles) in the plots. As assessed by using natural splines, the relation of age to diastolic blood pressure was nonlinear among both men and women.
[View Larger Version of this Image (31K GIF file)]
Individuals with hypertension were, on average, about 13 y older than individuals with normal blood pressure (Table 2). Compared with other, persons with hypertension also had a substantially higher age-adjusted BMI (+2.2 kg/m2 among men, +1.7 kg/m2 among women) and a higher subscapular/triceps skinfold ratio; the latter difference, which was statistically significant among women, persisted even after adjustment for BMI. The prevalence of cigarette smoking among women with hypertension was about threefold greater than among other women, but little association was seen among men. (The use of smokeless tobacco tended to be lower among persons with hypertension.) Among men, the prevalence of sedentary behavior tended to be slightly, but not significantly higher among those with hypertension.
|
Table 2.
Levels of selected characteristics by sex and hypertensive status, Navajo Health and Nutrition Survey, 1991-921
[View Table]
|
Hypertension was strongly associated with diabetes mellitus among men (Table 2), with an almost threefold difference (33 vs. 12%) between the two groups. Among women, a larger proportion of those with hypertension reported that they had been previously told that they had diabetes (33 vs. 15%), but none of the 34 women with hypertension was classified as having newly diagnosed diabetes mellitus on the basis of the oral glucose tolerance test. These contrasting findings resulted in the overall prevalence of diabetes mellitus among women showing little difference according to hypertension status (23 vs. 25%). The prevalence of impaired glucose tolerance among women with hypertension, however, was two- to threefold higher (33 vs. 14%) than among normotensive women.
The survey identified a relatively high number of persons with newly diagnosed hypertension. Of the 97 participants who had an elevated blood pressure level at the examination, only 50 reported that they had been previously told by a physician or other health professional that they had high blood pressure (data not shown). These 50 people tended to be slightly older (0.8 y) than the 47 persons with newly diagnosed hypertension. About three fourths of the 50 people with previously diagnosed hypertension reported that they had received advice about behavioral management of blood pressure, including weight loss, exercise, dietary sodium restriction or alcohol restriction.
The effectiveness of hypertension control was then examined among the 135 people who reported having been previously told by a doctor or health professional that they had high blood pressure (Table 3). Among these participants, 70% reported that they had been given behavioral advice concerning blood pressure management. However, the proportion of participants who reported receiving this advice did not differ significantly between those whose blood pressure at the examination was normal (n = 85) and those in whom it was elevated (n = 50). Furthermore, the proportion of the two groups who reported following this behavioral advice (62 and 56%) or using antihypertensive medications (35 and 42%) did not significantly differ. Older persons, however, were more likely to have an elevated blood pressure at the examination than were other participants.
|
Table 3.
Levels of various characteristics among the 135 participants who reported that they had previously been told that they had hypertension, Navajo Health and Nutrition Survey, 1991-921
[View Table]
|
DISCUSSION
The prevalence of hypertension among the Navajo people, like that of diabetes and heart disease, has increased markedly over the past half-century. A review of >4000 hospital admissions in the 1930s at an Arizona mission hospital serving the Navajo suggested that the prevalence of hypertension was <0.1% (Salsbury 1937). During the 1950s, community-based surveys, performed by Fulmer and Roberts (1963) and Darby et al. (1956), found that the prevalence of hypertension among the Navajo ranged from 4 to 7%. Two decades later, on the basis of a sample of clinic users, employees and community volunteers, DeStefano et al. (1979) estimated that the prevalence of hypertension among Navajo adults was 17%, similar to the age-standardized rate of 19% found in this study.
Comparing the current results with those of previous studies is difficult because of differences in the definition of hypertension, the survey and sampling methods used, and the proportion of persons using antihypertensive medications. For example, the use of clinic-based samples as in DeStefano's (1979) study could lead to an overestimation of the prevalence because persons with hypertension are more likely to be studied. Furthermore, much of the previously reported data was based on either single measurements or convenience samples (Alfred 1970, DeStefano et al. 1979, Fulmer and Roberts 1963, Kunitz and Levy 1986, Salsbury 1937). Several aspects of the design of this study, such as random selection of households, use of a random-zero sphygmomanometer and three separate blood pressure measurements, should lead to more accurate estimates of the prevalence of hypertension in the community. It is important to realize, however, that ~40% of the target households did not participate in the current study and that younger men are underrepresented (White et al. 1997b). Although it is difficult to quantify the effects of this nonparticipation, if working persons (who were possibly in better health) were less likely to participate, the prevalence of hypertension could be overestimated.
Our results are consistent with previous reports indicating that the prevalence of hypertension among the Navajo increases with age and is higher among men than women (Coulehan et al. 1986, Sugarman 1990). However, the conclusion of DeStefano et al. (1979) concerning the lack of association between systolic blood pressure and age among Navajo men is not supported by our data. Although the magnitude of the increase was greater among women, we found that systolic blood pressure levels increased with age among both men and women. The relatively small changes in diastolic blood pressure levels that we observed with age (~5 mm Hg increase between ages 20 and 60) were, in part, due to the inverse association between BMI and age. Somewhat similar findings have been reported in other populations (Clifford et al 1963, Joint National Committee on the Detection, Evaluation, and Treatment of High Blood Pressure 1993, Young 1991).
Although based in part on a relatively small number of hypertensives who underwent an oral glucose tolerance test, we also found that hypertension was associated with diabetes mellitus and impaired glucose tolerance. Other analyses of these data (Will et al. 1997) indicate that hypertension is ~2.5 times as prevalent among adults with a medical history of diabetes compared with other persons, and almost one half of Navajo adults with diabetes mellitus have been found to have hypertension in other studies (Sugarman 1990). The high rate of previously undiagnosed diabetes (22%) among men with hypertension in the current study, as well as the strong association between diabetes and hypertension in other studies (Broussard et al. 1993), suggests that routine screening for diabetes among Navajo adults with hypertension could have a high yield. Because of the importance of both diabetes and hypertension in heart disease and renal failure (Hoy et al. 1995), such a program could be a major part of prevention efforts.
The association between high blood pressure and obesity, well documented in other studies (Joint National Committee on the Detection, Evaluation, and Treatment of High Blood Pressure 1993), was also seen when BMI was used as a proxy for obesity in these analyses. Although persons with hypertension also tended to have a central distribution of body fat, as assessed by the ratio of the subscapular to triceps skinfolds, this difference was statistically significant only among women. Although there are some reports suggesting that the relation of body fat distribution to blood pressure may differ by sex or ethnic group (Blair et al. 1984), weaker association among Navajo men may be due to either the smaller number of participants or the skewed distribution of the subscapular/triceps skinfold ratio. Mendlein et al. (1997), however, found the association between central obesity and adverse lipid levels to be stronger among men than among women, and it is possible that these differences by sex are due to chance.
Only weak associations were seen between sedentary behavior and hypertension in the current analyses. An examination of survey responses, however, indicated that a substantial number of physically active persons considered their activity to be unrelated to exercise, and gave a negative response to the question. These complications, along with difficulties in coding several of the responses, precluded a more in-depth analysis of other measures of physical activity. However, it is likely that Navajo adults are more sedentary than those of only a generation ago (Broudy and May 1983), and further efforts to understand the relation of physical activity to hypertension (and obesity) among the Navajo would be helpful.
Compared with other populations, there appears to be fairly good control of hypertension among Navajo adults. About 60% of those who reported being previously told that they had high blood pressure had an average reading <140/90 in this study. However, restricting the definition of hypertension to include only persons with antihypertensive medications or an elevated systolic or diastolic measurement decreased the rate of blood pressure control to 29%. This estimate compares favorably with rates of 11 and 24% in national surveys (Burt et al. 1995), but is well below the Healthy People 2000 goal of 50% (U.S. Public Health Service 1991).
A potential weakness of the current study was the inability to verify prior diagnosis of hypertension against medical records. Some of the individuals who reported a previous diagnosis may have had only transient elevations of blood pressure, perhaps related to the stress of a physical examination or illness, whereas others likely had confirmed hypertension. Because we were not able to verify these self-reports, we chose to use possession of antihypertensive medications as a surrogate for a previous diagnosis of hypertension. However, if a substantial number of persons who had been previously diagnosed with hypertension were able to control their blood pressure through nonpharmacologic methods, the prevalence of hypertension would have been underestimated. In contrast, accepting all (unverified) self- reports of a previous diagnosis of hypertension as valid (rather than relying on the possession of antihypertensive medications) would have increased the estimate of the prevalence of hypertension to 25%. Because nonpharmacologic means are becoming more important in the management of high blood pressure, consideration should be given to including persons with diagnosed hypertension, but who were under control without medications in future surveys of hypertension and blood pressure control.
Results from this community-based survey of the Navajo Nation indicate that overall prevalence of hypertension among the Navajo is now similar to the overall U.S. prevalence (Burt 1995). This appears to represent a continuation of an increase in rates over several generations and parallels the increases that have been observed for obesity, diabetes and heart disease. Because the combination of hypertension, diabetes and obesity is a common syndrome among Navajo adults, paralleling the clustering of glucose intolerance and other characteristics in syndrome X (Reaven 1994), controlling these risk factors may mitigate future increases in coronary heart disease and chronic renal failure. Information gathered during the Navajo Health and Nutrition Survey shows that there are important opportunities for primary and secondary prevention of hypertension and its complications among the Navajo people.
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
,
,
ABSTRACT
