Hypertension is an important factor in premature death,

Transcription

1 Effect of Hypertension on Mortality in Pima Indians Maurice L. Sievers, MD; Peter H. Bennett, MB, MRCP, FFCM; Janine Roumain, MD, MPH; Robert G. Nelson, MD, PhD Background The effect of hypertension on mortality was examined in 5284 Pima Indians, 1698 of whom had type 2 diabetes at baseline or developed it during follow-up. Methods and Results During a median follow-up of 12.2 years (range, 0.01 to 24.8 years), 470 nondiabetic subjects and 488 diabetic subjects died. In the nondiabetic subjects, 45 of the deaths were due to cardiovascular disease, 208 to other natural causes, and 217 to external causes; in the diabetic subjects, 106 of the deaths were due to cardiovascular disease, 85 to diabetic nephropathy, 226 to other natural causes, and 71 to external causes. In the nondiabetic subjects, after adjusting for age, sex, body mass index, and serum cholesterol concentration in a proportional hazards model, hypertension predicted death from cardiovascular disease (death rate ratio [DRR] 2.8; 95% CI, 1.4 to 5.6; P 0.003). In the diabetic subjects, after additional adjustment for duration of diabetes, plasma glucose concentration, and proteinuria, hypertension strongly predicted deaths from diabetic nephropathy (DRR 3.5; 95% CI, 1.7 to 7.2; P 0.001), but it had little effect on deaths from cardiovascular disease (DRR 1.4; 95% CI, 0.88 to 2.3; P 0.15). Conclusions We propose that the weak relationship between hypertension and cardiovascular disease in diabetic Pima Indians is not because of a diminished effect of hypertension on cardiovascular disease in diabetes, but because of a relatively greater effect of hypertension on the progression of diabetic nephropathy. Factors that may account for this finding in Pima Indians include a younger age at onset of type 2 diabetes, a low frequency of heavy smoking, favorable lipoprotein profiles and, possibly, enhanced susceptibility to renal disease. (Circulation. 1999;100:33-40.) Key Words: cardiovascular diseases diabetes mellitus epidemiology hypertension Indians, North American mortality proportional hazards models Hypertension is an important factor in premature death, and it contributes to several of the leading causes of death in the United States, including heart disease, stroke, and renal disease. 1 4 Differences in the age at onset, frequency, and severity of hypertension are responsible for substantial variations in death rates and causes of death in different racial and ethnic groups. 5 Furthermore, coexistent diseases may modify or obscure the effect of hypertension on mortality. In particular, diabetes, which is associated with a high frequency of hypertension, 6,7 is independently associated with increased death rates from heart disease, stroke, and renal disease Pima Indians from the Gila River Indian Community in Arizona have a lower prevalence of hypertension than the general US population 7 and a substantially lower prevalence and incidence rate of heart disease. 13,14 However, they have the world s highest incidence rate of type 2 diabetes, and more than 20 times the rate of end-stage renal disease. 15 Thus, the relative influence of hypertension on noncardiovascular mortality may be greater in this population than in others. In the present study, we examined the effect of hypertension on overall and cause-specific mortality in the Pima Indians. Subjects and Methods Pima and the closely-related Tohono O odham Indians from the Gila River Indian Community are participating in a longitudinal diabetes study. 16 This study was approved by the review boards of the National Institute of Diabetes and Digestive and Kidney Diseases and by the Tribal Council. Since 1965, each member of the Community 5 years of age is asked to have a research examination every 2 years. Informed consent is given at each examination. These examinations include a glucose tolerance test, with determination of the glucose concentration in venous plasma drawn 2 hours after a 75-g oral carbohydrate load. Diabetes was defined according to World Health Organization criteria. 17 The date of diagnosis was determined from the research examinations or from review of clinical records if diabetes was diagnosed in the course of routine medical care. Blood pressure was measured to the nearest 2 mm Hg with a mercury sphygmomanometer while the subject rested in the supine position. Hypertension was defined (using the criteria of the sixth report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure 18 ) as a systolic blood pressure 140 mm Hg, a diastolic blood pressure 90 mm Hg, or treatment with antihypertensive medicine. Proteinuria was defined as a protein-to-creatinine ratio of at least 0.5 g/g, equivalent to a total protein excretion rate of 500 mg or more per day. Received December 14, 1998; revision received March 24, 1999; accepted April 15, From the Phoenix Epidemiology and Clinical Research Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Phoenix, Arizona. Correspondence to Dr Robert G. Nelson, National Institutes of Health, 1550 East Indian School Road, Phoenix, AZ rnelson@phx.niddk.nih.gov 1999 American Heart Association, Inc. Circulation is available at 33

2 34 Circulation July 6, 1999 TABLE 1. Baseline Clinical and Demographic Features of the Study Population Nondiabetic Subjects* (2011 men, 2285 women) Diabetic Subjects (441 men, 547 women) Mean (SD) Range Mean (SD) Range Age, y 26 (14) (15) Duration of diabetes, y 4.6 (5.8) Body mass index, kg/m 2 Men 28.2 (6.6) (6.7) Women 29.4 (6.8) (6.5) hr plasma glucose, mmol/l 6.0 (1.6) (7.1) Mean arterial pressure, mm Hg 88 (13) (15) Serum cholesterol, mmol/l 4.19 (0.91) (1.16) *Of the nondiabetic Pima Indians, 710 subsequently developed diabetes during the study period. Between January 1, 1965 and December 31, 1989, 5506 subjects who were at least half Pima or Tohono O odham Indian and who had 1 research examination(s) after 15 years of age resided in the Community. A total of 222 of these subjects were excluded from the study because of missing blood pressure data. The vital status of the remaining 5284 subjects as of December 31, 1989 was determined. For all deaths, the accuracy and completeness of the underlying and contributory causes reported on death certificates were assessed by review of clinical records and reports of autopsy and medical examiner findings. 19,20 The adjudicated causes were coded in accordance with the guidelines of the National Center for Health Statistics. The terminology and codes of the ninth revision of the International Classification of Disease 21 were used for recording causes of death and other diagnoses. Statistical Analysis Death rates were calculated as the number of subjects who died divided by the person-years of follow-up and expressed per 1000/ year. If subjects developed diabetes during follow-up, their personyears were apportioned to the appropriate stratum. The period of risk extended from the date of the first examination to death or December 31, Death rates were standardized to the 1980 Pima Indian population, and death rate ratios (DRRs) comparing overall and cause-specific death rates were computed from the age- and age/sexadjusted rates. Confidence intervals were computed from the logarithms of the rate ratios. 22 The effect of hypertension on mortality was examined using a Cox proportional hazards model to control for the effects of potentially confounding variables. Regression analyses were done separately in the nondiabetic and diabetic subjects. Subjects who had eligible follow-up both when nondiabetic and later when diabetic were included in both analyses. Of the 4296 subjects (17%) who were nondiabetic at baseline, 710 developed diabetes during the study period; 674 of them (95%) had at least 1 research examination after the onset of diabetes. For diabetic subjects, the baseline examination was the first examination at or after the diagnosis of diabetes at which blood pressure was measured. For nondiabetic subjects, the baseline examination was the first examination after 15 years of age at which blood pressure was measured. Continuous variables included in the regression models were centered at their mean baseline values. The effects of time-dependent variables were assessed at baseline and as they changed at subsequent examinations. If the values changed, the new values were included for the appropriate time periods. Squared (quadratic) terms for body mass index, plasma glucose concentration, and serum cholesterol concentration were included in some models because the effect of these variables for specific causes of death was shown previously to be U-shaped Product terms of predictor variables did not improve the regression models and were not included. The adequacy of the fit of each model to individual observations was assessed by inspection of deviance residuals. Results Of the 5284 subjects (2452 men, 2832 women) 15 years old, 958 (578 men, 380 women) died during a median follow-up of 12.2 years (range, 0.01 to 24.8 years); 470 of the deaths (329 men, 141 women) were in the 4296 subjects without diabetes, and 488 (249 men, 239 women) were in the 1698 subjects who had diabetes at the baseline examination or developed it during follow-up. Clinical and demographic features of the study population at baseline are shown in Table 1. The diabetic subjects were older, on average, and more obese than those without diabetes, and their mean arterial pressure and serum cholesterol concentrations were higher. Hypertension was present in 17% of the nondiabetic subjects and in 53% of those with diabetes at their initial examination. A total of 217 of the 470 deaths (46%) in the nondiabetic subjects and 71 of the 488 deaths (15%) in the diabetic subjects were due to external causes (Table 2), reflecting the higher proportion of traumatic deaths in the younger, nondiabetic subjects. Age-specific rates for natural causes of death were higher in diabetic subjects than nondiabetic ones and higher in men than women. The death rate rose with increasing age in both sexes, and it was generally higher in the hypertensive than the normotensive subjects (Figure 1). In those subjects without diabetes, the effect of hypertension on natural causes of death was greater in men than in women (men: DRR 1.5; 95% CI, 1.1 to 2.0; women: DRR 1.0; 95% CI, 0.64 to 1.7). However, in the diabetic subjects, the effect was greater in women (men: DRR 1.1; 95% CI, 0.83 to 1.6; women: DRR 1.8; 95% CI, 1.3 to 2.6). Cardiovascular disease, which accounted for 18% (n 45) of the natural causes of death in nondiabetic Pima Indians and 25% (n 106) of the natural deaths in those with diabetes, was more strongly related to hypertension in the nondiabetic subjects (Table 2; nondiabetic subjects: DRR 2.6; 95% CI, 1.2 to 5.6; diabetic subjects: DRR 1.2; 95% CI, 0.72 to 1.9). Furthermore, the effect of hypertension on cardiovascular disease in the nondiabetic subjects was greater in men than women (men: DRR 4.9; 95% CI, 1.9 to 13.1; women:

3 Sievers et al Effect of Hypertension on Mortality 35 TABLE 2. Number of Deaths and Death Rates* for Underlying Causes of Death in 5284 Pima Indians, According to Diabetes and Blood Pressure Status Nondiabetic Subjects Diabetic Subjects Normotensive Hypertensive Normotensive Hypertensive Underlying Cause of Death n Rate n Rate DRR (95% CI) n Rate n Rate DRR (95% CI) Cardiovascular disease ( ) ( ) ( ) Ischemic heart disease ( ) ( ) ( ) Stroke ( , 436) (1.4 15) ( ) Other cardiovascular diseases ( ) ( ) Diabetes-related diseases ( ) ( ) Diabetic nephropathy (250.4) ( ) Other diabetes-related diseases ( ) Other natural causes ( ) ( ) Infections ( ) ( ) Malignant neoplasms ( ) ( ) ( ) Alcoholic liver disease ( ) ( ) ( ) Other causes ( ) ( ) All natural causes ( ) ( ) All external causes ( ) ( ) All causes ( ) ( ) *Age/sex-adjusted rate per 1000/y. Numbers in parentheses are International Classification of Disease, revision 9 (ICD-9) codes. ICD-9 codes for infections: ; ; ; ; ; 572.0; ; 599.0; ; DRR 1.4; 95% CI, 0.45 to 4.3); however, in the diabetic subjects, hypertension had a slightly greater effect in women (men: DRR 1.1; 95% CI, 0.62 to 2.0; women: DRR 1.4; 95% CI; 0.67 to 2.9) (Figure 2). The effect of hypertension on deaths from cardiovascular disease in nondiabetic subjects was due principally to its effect on deaths from stroke (Table 2). Hypertensive subjects, compared with those with normal blood pressure, Figure 1. Age-specific death rates for natural causes by sex and diabetes (top, nondiabetic subjects; bottom, diabetic subjects) in hypertensive (HTN) and normotensive (no HTN) Pima Indians.

4 36 Circulation July 6, 1999 Figure 2. Age-adjusted cause-specific death rates by sex and diabetes (top, nondiabetic subjects; bottom, diabetic subjects) in hypertensive (HTN) and normotensive (no HTN) Pima Indians. Deaths from diabetic nephropathy are not included in the category other natural causes of death in diabetic subjects. were 4.6 times (95% CI, 1.4 to 15) more likely to die from a stroke. However, in diabetic subjects, those with hypertension were only 1.3 times (95% CI, 0.5 to 3.3) as likely to die from a stroke. The effect of hypertension on deaths from stroke was greater in the nondiabetic men than women (men: DRR 5.6; 95% CI, 1.2 to 25; women: DRR 3.7; 95% CI, 0.54 to 25), but it was similar in both sexes in the diabetic subjects (men: DRR 1.2; 95% CI, 0.31 to 4.5; women: DRR 1.3; 95% CI, 0.34 to 4.9) (Figure 3). Diabetic nephropathy, which accounted for 20% (n 85) of the natural causes of death in diabetic Pima Indians, was strongly related to hypertension (DRR 4.3; 95% CI, 2.2 to 8.4). The effect of hypertension was greater in women than men (men: DRR 3.0; 95% CI, 1.1 to 8.1; women: DRR 5.4; 95% CI, 1.9 to 15.5) (Figure 4). When examined in a Cox proportional hazards model adjusting for age, sex, serum cholesterol concentration, and body mass index, hypertension in nondiabetic subjects was strongly associated with cardiovascular deaths (P 0.003) but less associated with other natural causes of death (P 0.09; Table 3). In the diabetic subjects, hypertension was strongly associated with deaths from diabetic nephropathy (P 0.001) but not with deaths from cardiovascular disease (P 0.15) or other natural causes (P 0.13), after adjusting for age, sex, duration of diabetes, 2-hour post-load plasma glucose concentration, serum cholesterol concentration, body mass index, and proteinuria (Table 4). The number of stroke deaths was too small (n 20) to make a valid estimate of the independent effect of hypertension in a multiple regression model. Hypertension did not affect deaths from external causes in either the nondiabetic or diabetic subjects (data not shown). Figure 3. Age-adjusted death rates for stroke by sex and diabetes in hypertensive (HTN) and normotensive (no HTN) Pima Indians.

5 Sievers et al Effect of Hypertension on Mortality 37 Figure 4. Age-adjusted death rates for diabetic nephropathy by sex in hypertensive (HTN) and normotensive (no HTN) Pima Indians. Discussion Hypertension increased the death rate from cardiovascular disease nearly 3-fold in nondiabetic Pima Indians; however, in contrast with findings from other populations, 8 11 it had little effect on cardiovascular deaths in those with diabetes. The apparent lack of effect in the diabetic subjects may be due to competing causes of death. Cardiovascular disease and diabetic renal disease share many of the same risk factors, including hypertension; and hypertension had a major effect on deaths from diabetic nephropathy, increasing the death rate 4-fold. Hypertension in the presence of inherited susceptibility to renal disease may accelerate the development of diabetic nephropathy relative to cardiovascular disease and may explain its reduced effect on cardiovascular deaths. In addition, the younger age at onset of type 2 diabetes in the Pima Indians, 22,26 the lower prevalence of heavy smoking (see below), 13,27 and the lower serum concentrations of total and low-density lipoprotein cholesterol when compared with US Caucasians 28 may limit the number of Pima Indians with preexisting macrovascular disease at the onset of diabetes and provide a greater opportunity for renal disease to develop before life-threatening cardiovascular disease has time to evolve. Hypertension had almost no effect on deaths from ischemic heart disease in either the nondiabetic or diabetic subjects, presumably because these TABLE 3. Estimated Adjusted Effects of Hypertension and Other Factors on Deaths From Cardiovascular Disease and Other Natural Causes of Death in Nondiabetic Pima Indians: Results of Proportional Hazards Regression Analysis Model Covariates Rate Ratio* (95% CI) P Value Cardiovascular disease (44 deaths) Age (per 10 years) 2.3 ( ) Sex (male/female) 1.4 ( ) 0.33 Hypertension (yes/no) 2.8 ( ) Serum cholesterol (per 1.29 mmol/l [50 mg/dl]) 1.4 ( ) 0.14 Body mass index (per 5 kg/m 2 ) } 0.01 Body mass index body mass index Other natural causes of death (204 deaths) Age (per 10 years) 2.1 ( ) Sex (male/female) 1.5 ( ) Hypertension (yes/no) 1.3 ( ) 0.09 Serum cholesterol (per 1.29 mmol/l [50 mg/dl]) } Serum cholesterol serum cholesterol Body mass index (per 5 kg/m 2 ) Body mass index body mass index } 0.04 Missing values for body mass index and/or serum cholesterol reduced the number of cardiovascular deaths by 1 and other natural causes of deaths by 4 in comparison with Table 2. *Rate ratio is for the number of units shown in the first column. All rate ratios are adjusted for the other variables in the model. Squared terms for these variables precluded estimates of their effects.

6 38 Circulation July 6, 1999 TABLE 4. Estimated Adjusted Effects of Hypertension and Other Factors on Deaths From Cardiovascular Disease, Diabetic Nephropathy, and Other Natural Causes of Death in Diabetic Pima Indians: Results of Proportional Hazards Regression Analysis Model Covariates Rate Ratio* (95% CI) P Value Cardiovascular disease (100 deaths) Age (per 10 years) 2.1 ( ) Sex (male/female) 2.6 ( ) Duration of diabetes (per 5 years) 1.2 ( ) Hypertension (yes/no) 1.4 ( ) 0.15 Plasma glucose (per 5.6 mmol/l [100 mg/dl]) 1.1 ( ) 0.15 Serum cholesterol (per 1.29 mmol/l [50 mg/dl]) 1.4 ( ) Body mass index (per 5 kg/m 2 ) } 0.66 Body mass index body mass index Proteinuria (yes/no) 1.9 ( ) Diabetic nephropathy (83 deaths) Age (per 10 years) 1.1 ( ) 0.19 Sex (male/female) 0.85 ( ) 0.51 Duration of diabetes (per 5 years) 1.4 ( ) Hypertension (yes/no) 3.5 ( ) Plasma glucose (per 5.6 mmol/l [100 mg/dl]) } 0.11 Plasma glucose plasma glucose Serum cholesterol (per 1.29 mmol/l [50 mg/dl]) 1.3 ( ) Body mass index (per 5 kg/m 2 ) 0.97 ( ) 0.83 Proteinuria (yes/no) 14 (7.7 27) Other natural causes of death (218 deaths) Age (per 10 years) 1.6 ( ) Sex (male/female) 1.5 ( ) Duration of diabetes (per 5 years) 1.0 ( ) 0.62 Hypertension (yes/no) 1.3 ( ) 0.13 Plasma glucose (per 5.6 mmol/l [100 mg/dl]) 0.98 ( ) 0.70 Serum cholesterol (per 1.29 mmol/l [50 mg/dl]) } Serum cholesterol serum cholesterol Body mass index (per 5 kg/m 2 ) } Body mass index body mass index Proteinuria (yes/no) 1.7 ( ) Missing values for body mass index, plasma glucose, serum cholesterol, and/or proteinuria reduced the number of cardiovascular deaths by 6, diabetic nephropathy deaths by 2, and other natural causes of deaths by 4 in comparison with Table 2. *Rate ratio is for the number of units shown in the first column. All rate ratios are adjusted for the other variables in the model. Squared terms for these variables precluded estimates of their effects. other risk factors were missing. Nevertheless, death rates from ischemic heart disease were much higher in the diabetic subjects, presumably because risk factors other than hypertension were overwhelmingly more important. In Pima Indians, the lower prevalence of clinically diagnosed and electrocardiographically and autopsy-proven atherosclerotic coronary heart disease 29,30 and the lower incidence of fatal coronary heart disease compared with US Caucasians 14 seems paradoxical in view of the high rates of type 2 diabetes. Competing causes of death in those with diabetes, as proposed in the present study, may explain this paradox. Recent studies indicate that deaths from ischemic heart disease in American Indians are increasing, 31,32 a finding that is attributable, at least in part, to greater use of renal replacement therapy and to a substantial improvement in survival during such therapy. 31,33,34 The resulting enhanced survival of diabetic patients, who otherwise would have died from renal disease, permits many of them to die from cardiovascular disease. The evidence that coronary heart disease rates in American Indians actually exceed rates in

7 Sievers et al Effect of Hypertension on Mortality 39 other US populations 35 is based on comparisons that do not account for differences in the prevalence and duration of diabetes. In nondiabetic subjects, hypertension had a greater adverse effect on mortality in men than in women, but in those with diabetes, the effect was greater in the women. Among the diabetic subjects, this was particularly striking for deaths from diabetic nephropathy; hypertension increased the age-adjusted death rate from diabetic nephropathy 5-fold in the women and 3-fold in the men. Studies in this population and others indicate that diabetes has a relatively greater impact on mortality in women than men, 8,9,19,36,37 and the present study suggests that this effect is mediated, at least in part, by the differential effect of hypertension in the sexes. In both diabetic and nondiabetic subjects, hypertension was generally associated with higher death rates from other natural causes of death unrelated to renal or cardiovascular disease, but the effect was small and not consistent for specific causes of death (Table 2). In addition, hypertension had no effect on external causes of death. Less than 1% of adult Pima Indians smoke 1 pack of cigarettes per day, 13,27 and previous studies in this population found that smoking did not predict either diabetic microvascular disease 27 or heart disease. 13 Because of the low frequency of heavy smoking, it was not included as a covariate in the present study. In addition, serum insulin concentration was not included as a covariate because neither fasting nor 2-hour post-load insulin concentration predicted the development of electrocardiographic abnormalities in Pima Indians. 38 In conclusion, hypertension strongly predicted deaths from stroke in nondiabetic Pima Indians and from nephropathy in those with type 2 diabetes. Hypertension had no identifiable effect on deaths from ischemic heart disease in either the nondiabetic or diabetic subjects or on deaths from stroke in the diabetic subjects. These observations are probably explained by the limited exposure to other risk factors for cardiovascular disease in this population and by competing causes of death, whereby the diabetic subjects die from renal disease before they develop life-threatening cardiovascular disease. Without the risk of diabetic renal disease, the nondiabetic subjects lived long enough for hypertension to promote fatal strokes. Acknowledgments The authors are indebted to the members of the Gila River Indian Community for participating in this investigation. References 1. Stamler J. Blood pressure and high blood pressure: aspects of risk. Hypertension. 1991;18:I-95 I Flack JM, Neaton J, Grimm R Jr, Shih J, Cutler J, Ensrud K, MacMahon S. Blood pressure and mortality among men with prior myocardial infarction: Multiple Risk Factor Intervention Trial Research Group. Circulation. 1995;92: National Heart, Lung, and Blood Institute. Fact Book Fiscal Year Bethesda, Md: US Department of Health and Human Services; US Renal Data System. USRDS 1997 Annual Data Report. Bethesda, Md: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; Klag MJ, Whelton PK, Randall BL, Neaton JD, Brancati FL, Stamler J. End-stage renal disease in African-American and white men: 16-year MRFIT findings. JAMA. 1997;277: Reaven GM, Hoffman BB. A role for insulin in the aetiology and course of hypertension. Lancet. 1987;2: Saad MF, Knowler WC, Pettitt DJ, Nelson RG, Mott DM, Bennett PH. Insulin and hypertension: relationship to obesity and glucose intolerance in Pima Indians. Diabetes. 1990;39: Kannel WB, McGee DL. Diabetes and cardiovascular disease: the Framingham study. JAMA. 1979;241: Fuller JH, Shipley MJ, Rose G, Jarrett RJ, Keen H. Mortality from coronary heart disease and stroke in relation to degree of glycaemia: the Whitehall study. BMJ. 1983;287: Stamler J, Vaccaro O, Neaton JD, Wentworth D. Diabetes, other risk factors, and 12-yr cardiovascular mortality for men screened in the Multiple Risk Factor Intervention Trial. Diabetes Care. 1993;16: Kuller LH. Stroke and diabetes. In: Diabetes in America. 2nd ed. Bethesda, Md: National Diabetes Data Group, US Department of Health and Human Services; 1995: NIH Publication No Nelson RG, Pettitt DJ, Carraher MJ, Baird HR, Knowler WC. Effect of proteinuria on mortality in NIDDM. Diabetes. 1988;37: Howard BV, Lee ET, Cowan LD, Fabsitz RR, Howard WJ, Oopik AJ, Robbins DC, Savage PJ, Yeh JL, Welty TK. Coronary heart disease prevalence and its relation to risk factors in American Indians: the Strong Heart Study. Am J Epidemiol. 1995;142: Nelson RG, Sievers ML, Knowler WC, Swinburn BA, Pettitt DJ, Saad MF, Liebow IM, Howard BV, Bennett PH. Low incidence of fatal coronary heart disease in Pima Indians despite high prevalence of noninsulin-dependent diabetes. Circulation. 1990;81: Nelson RG, Newman JM, Knowler WC, Sievers ML, Kunzelman CL, Pettitt DJ, Moffett CD, Teutsch SM, Bennett PH. Incidence of end-stage renal disease in Type 2 (non-insulin-dependent) diabetes mellitus in Pima Indians. Diabetologia. 1988;31: Bennett PH, Burch TA, Miller M. Diabetes mellitus in American (Pima) Indians. Lancet. 1971;2: World Health Organization. Diabetes mellitus. Geneva: World Health Organization; Technical Report Series No Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. The sixth report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med. 1997;157: Sievers ML, Nelson RG, Knowler WC, Bennett PH. Impact of NIDDM on mortality and causes of death in Pima Indians. Diabetes Care. 1992; 15: Sievers ML, Nelson RG, Bennett PH. Adverse mortality experience of a southwestern American Indian Community: overall death rates and underlying causes of death in Pima Indians. J Clin Epidemiol. 1990;43: US National Center for Health Statistics. The International Classification of Diseases: Clinical Modification. 9th revision, Vol 1. Ann Arbor, Mich: Edwards Brothers; Knowler WC, Bennett PH, Hamman RF, Miller M. Diabetes incidence and prevalence in Pima Indians: a 19-fold greater incidence than in Rochester, Minnesota. Am J Epidemiol. 1978;108: Hanson RL, McCance DR, Jacobsson LTH, Narayan KMV, Nelson RG, Pettitt DJ, Bennett PH, Knowler WC. The u-shaped association between body mass index and mortality: relationship with weight gain in a Native American population. J Clin Epidemiol. 1995;48: Fagot-Campagna A, Hanson RL, Narayan KMV, Sievers ML, Pettitt DJ, Nelson RG, Knowler WC. Serum cholesterol and mortality rates in a Native American population with low cholesterol concentrations: a u-shaped association. Circulation. 1997;96: Sievers ML, Bennett PH, Nelson RG. Effect of glycemia on mortality in Pima Indians with type 2 diabetes. Diabetes. 1999;48: Knowler WC, Pettitt DJ, Saad MF, Bennett PH. Diabetes mellitus in the Pima Indians: incidence, risk factors and pathogenesis. Diabetes Metab Rev. 1990;6: Nelson RG, Wolfe JA, Horton MB, Pettitt DJ, Bennett PH, Knowler WC. Proliferative retinopathy in NIDDM: incidence and risk factors in Pima Indians. Diabetes. 1989;38:

8 40 Circulation July 6, Howard BV, Davis MP, Pettitt DJ, Knowler WC, Bennett PH. Plasma and lipoprotein cholesterol and triglyceride concentrations in the Pima Indians: distributions differing from those of Caucasians. Circulation. 1983;68: Inglefinger JA, Bennett PH, Liebow IM, Miller M. Coronary heart disease in the Pima Indians: electrocardiographic findings and postmortem evidence of myocardial infarction. Diabetes. 1976;25: Howard BV, Lee ET, Cowan LD, Fabsitz RR, Howard WJ, Oopik AJ, Robbins DC, Savage PJ, Yeh JL, Welty TK. Coronary heart disease prevalence and its relation to risk factors in American Indians: The Strong Heart Study. Am J Epidemiol. 1995;142: Sievers ML, Nelson RG, Bennett PH. Sequential trends in overall and cause-specific mortality in diabetic and nondiabetic Pima Indians. Diabetes Care. 1996;19: Lee ET, Cowan LD, Welty TK, Sievers M, Howard WJ, Oopik A, Wang W, Yeh J, Devereux RB, Rhoades ER, Fabsitz RR, Go O, Howard BV. All-cause mortality and cardiovascular disease mortality in 3 American Indian populations, aged years, : The Strong Heart Study. Am J Epidemiol. 1998;147: Nelson RG, Hanson RL, Pettitt DJ, Knowler WC, Bennett PH. Survival during renal replacement therapy for diabetic end-stage renal disease in Pima Indians. Diabetes Care. 1996;19: Sievers ML, Nelson RG, Bennett PH. Decrease in mortality from diabetic nephropathy in Pima Indians. Diabetes Care. 1996;19:777. Letter. 35. Howard BV, Lee ET, Cowan LD, Devereux RB, Galloway JM, Go OT, Howard WJ, Rhoades ER, Robbins DC, Sievers ML, Welty TK. The rising tide of cardiovascular disease in American Indians: The Strong Heart Study. Circulation. 1999;99: O Sullivan JB, Mahan CM. Mortality related to diabetes and blood glucose levels in a community study. Am J Epidemiol. 1982;116: Panzram G. Mortality and survival in type 2 (non-insulin-dependent) diabetes mellitus. Diabetologia. 1987;30: Liu QZ, Knowler WC, Nelson RG, Saad MF, Charles MA, Liebow IM, Bennett PH, Pettitt DJ. Insulin treatment but not endogenous insulin concentration is associated with ECG abnormalities in diabetic Pima Indians. Diabetes. 1992;41: