Low Incidence of Fatal Coronary Heart Disease in Pima Indians Despite High Prevalence of Non-Insulin-Dependent Diabetes

Transcription

1 987 Low Incidence of Fatal Coronary Heart Disease in Pima Indians Despite High Prevalence of Non-Insulin-Dependent Diabetes Robert G. Nelson, MD, MPH, Maurice L. Sievers, MD, William C. Knowler, MD, DrPH, Boyd A. Swinburn, MB, ChB, FRACP, David J. Pettitt, MD, Mohammed F. Saad, MD, MRCP (UK), Irving M. Liebow, MD, Barbara V. Howard, PhD, and Peter H. Bennett, MB, FRCP, FFCM The incidence of fatal coronary heart disease (CHD) was determined in a population of Pima Indians from the Gila River Indian Community in Arizona. Between 1975 and 1984, 394 deaths occurred among 4,828 subjects aged 5 years or older, and 199 of these occurred in the 1,093 persons with non-insulin-dependent diabetes. Only 28 deaths were attributed to CHD; all occurred among the 689 diabetic persons 45 years of age or older. No CHD deaths occurred among the 419 nondiabetic subjects 45 years of age or older. The rate of fatal CHD among the diabetic subjects was higher in men than in women and increased with advancing age and duration of diabetes. A higher incidence of fatal CHD was associated with proteinuria, renal insufficiency, medial arterial calcification, diabetic retinopathy, insulin therapy, and an abnormal electrocardiogram. In Pima Indians aged years, the incidence of fatal CHD was less than half that found in the Framingham population after controlling for age, sex, and diabetes (incidence rate ratio, 0.4; 95% confidence interval, ). Factors protecting Pima Indians from fatal CHD may include racial heritage, low serum concentrations of total and low density lipoprotein cholesterol, and rarity of heavy smoking. Among the diabetic subjects, mortality from diabetic renal disease, which shows many of the same risk factors, may selectively compete and remove those at risk for fatal CHD. This would not, however, explain the lack of fatal CHD among the nondiabetic subjects. Fatal CHD shares many of the risk factors associated with the specific microvascular complications of diabetes, and diabetes and its associated attributes are the major predictors of fatal CHD in this population. (Circulation 1990;81: ) A therosclerotic coronary heart disease (CHD) occurs more frequently in persons with diabetes mellitus than in those without it1,2 and is the most common underlying cause of death in diabetic adults in the United States.2 Although the Pima Indians of the Gila River Indian Community in Arizona have the world's highest reported prevalence of non-insulin-dependent diabetes,3 they have a low frequency of clinically apparent myocardial infarction4 and a low prevalence of electrocardiographic (ECG) and necropsy-proven CHD.4-6 From the Department of Biostatistics and Epidemiology (R.G.N.), The Cleveland Clinic Foundation, Phoenix; and the Phoenix Epidemiology and Clinical Research Branch (M.L.S., W.C.K., B.A.S., D.J.P., M.F.S., I.M.L., B.V.H., P.H.B.), National Institute of Diabetes and Digestive and Kidney Diseases, Phoenix, Arizona. Address for correspondence: Dr. Robert G. Nelson, The Cleveland Clinic Foundation, 1616 E. Indian School Road, Suite 140, Phoenix, AZ Received July 20, 1989; revision accepted November 28, In the present study, the incidence of CHD as an underlying cause of death was determined in the Pima Indians with and without non-insulindependent diabetes. Risk factors for CHD death were identified, and the incidence of fatal CHD was compared with that of the primarily caucasian population participating in the Framingham study.7 Methods A longitudinal study of diabetes and its complications has been conducted in the Gila River Indian Community of Arizona since This community is inhabited primarily by Pima and Papago Indians. Approximately every 2 years, every resident who is 5 years of age or older, regardless of health, is asked to participate in a standardized medical examination. Informed consent is obtained before each examination. The physical examination included measurement of blood pressure in the right arm with the subject in the supine position. Diastolic blood pressure was

2 988 Circulation Vol 81, No 3, March 1990 measured at the fourth Korotkoff sound. After pupillary dilation, direct ophthalmoscopic examination was performed by a physician, and diabetic retinopathy was considered present if the subject had one or more exudates, microaneurysms, or hemorrhages or had proliferative retinopathy. Subjects were interviewed at each examination to determine their smoking habits. Persons who smoked any amount within 1 year of the examination were considered to be smokers. Each biennial examination included determination of the serum creatinine and plasma glucose concentrations in venous blood drawn 2 hours after the ingestion of a 75 -g carbohydrate load (Glucola, Ames Co., Elkhart, Indiana, or Dexcola, Custom Laboratories, Baltimore, Maryland). Diabetes was diagnosed according to World Health Organization (WHO) criteria for epidemiologic studies9 (i.e., if the 2-hour postload plasma glucose concentration was.11.1 mm [200 mg/dl]). The date of diagnosis of diabetes was determined from biennial examinations and from review of clinical records. Subjects were asked to void at the beginning of the glucose tolerance test, and a urine specimen was collected 2 hours later. A screening test for protein in the 2-hour urine was done by dipstick (Labstix, Ames Co.). Urine specimens containing at least a trace of protein on dipstick were tested quantitatively for protein by the Shevky- Stafford method.10 The concentration of creatinine was measured in the same urine specimen, and the protein-to-creatinine ratio was calculated. Proteinuria was defined as a protein-to-creatinine ratio of at least 1.0 (>1.0 g protein/g creatinine); this is equivalent to a total protein excretion rate of approximately 1 g/day under the assumption that the daily urinary excretion of creatinine averages about 1 g."'l2 Renal insufficiency was defined as a serum creatinine concentration of 177 gm (2.0 mg/dl) or more. Twelve-lead ECGs were obtained on all subjects at their first biennial examination and at each examination subsequent to their 15th birthday. All ECGs were interpreted and classified according to the Minnesota Code13'4 by the same cardiologist (I.L.), who had no knowledge of the clinical data. Major ECG abnormalities were identified according to the Tecumseh criteria.15 Radiographic examinations of the hands, feet, left calf, left thigh, and pelvis were performed every 2-4 years after the age of 15 years. The radiographs were examined for arterial calcification by a radiologist who had no knowledge of the clinical data.16 Linear calcification in the arteries was categorized as medial calcification and any patchy calcification was categorized as intimal calcification.17 The Pima Indian study population (4,828 subjects) consisted of all persons 5 years of age or older who lived in the Gila River Indian Community at any time between January 1, 1975, and December 31, 1984, whose heritage was at least 50% Pima, Papago, or a mixture of these two closely related tribes and who had undergone one or more biennial research examinations. The underlying cause of death for each of the 394 subjects who died during the study period was determined by review of all available medical information. This review included inpatient and outpatient clinical records from the Public Health Service Indian Hospital at Sacaton, Arizona, where primary care is provided for this population; from the Phoenix Indian Medical Center, the principal referral hospital; and from any other hospital that provided care during the period immediately preceding death. In addition, death certificates, autopsy findings, and medical examiners' reports were reviewed. Based on this information, the underlying cause of death was determined for each subject in accordance with published guidelines of the National Center for Health Statistics (NCHS) Terminology and codes of the ninth revision of the Intemational Classification of Disease (ICD-9)23 were used to record underlying cause of death. CHD death was defined by the Lipid Research Clinics criteria for definite atherosclerotic coronary heart disease death as death from 1) acute CHD (ICD-9 codes ) (i.e., preterminal hospitalization with myocardial infarction), 2) chronic CHD (ICD-9 codes ) (i.e., previous angina pectoris or myocardial infarction when no cause other than atherosclerotic CHD could be ascribed as cause of death), or 3) observed sudden death occurring within 1 hour and not attributable to another cause (ICD-9 code 798.1).24 Data on CHD deaths and death rates during the same time period in the Framingham population7 for subjects aged years were compared with those of the Pima Indians. Similar criteria for fatal CHD were used in the Framingham population.7 Statistical Analysis Incidence was expressed as the number of deaths from CHD per 1,000 person-years of observation for subjects within the community. The numerator was the number of subjects dying from CHD between January 1, 1975, and December 31, The denominator was the number of person-years at risk for death. The period of risk began on either January 1, 1975, or the date of the first biennial examination for subjects entering the study after that date and extended to the date of death, emigration from the community, or close of the study (December 31, 1984) -whichever was earlier. Deaths and personyears were stratified by age, sex, and variables measured at biennial examinations. For subjects who entered the study on January 1, 1975, the persontime up to the next biennial examination after that date was stratified according to data from their most recent previous biennial examination. Subjects who developed diabetes during the study period accumulated person-time in the nondiabetic stratum up to the date of diabetes diagnosis and in the diabetic stratum thereafter. Likewise, as subjects changed from one age or diabetes-duration stratum to another, their person-time within each category accumulated within the corresponding stratum.

3 Incidence-rate ratios were used to assess the effect of dichotomous variables on the incidence of CHD death and to compare death rates of the Pima and Framingham populations. These ratios and their confidence intervals were computed for person-time denominators by a modification of the Mantel and Haenszel procedure25 proposed by Rothman and Boice.26 The effect of continuous variables was evaluated by grouping the values of the variables and assessing average partial association in three-way contingency tables with a x2 test under the multiple hypergeometric model while controlling for the effects of a set of covariables.27 Variables representing potential risk factors for CHD death were analyzed and controlled for age, sex, and duration of diabetes. The small number of subjects with fatal CHD did not permit multivariate analysis, controlled for all variables simultaneously. Age- and sex-adjtusted CHD death rates for the Pima and Framingham populations were computed by direct standardization to the 1980 US population (age, years), and confidence intervals were computed as described previously.3 Results During the study, 394 subjects died, of whom 199 were among the 1,093 subjects with diabetes and 195 were among the 3,735 subjects without diabetes. Only 28 deaths were attributed to CHD, and all were among the 689 diabetic subjects 45 years of age or older. No CHD deaths occurred among the 419 nondiabetic subjects 45 years of age or older. Ten deaths (seven men and three women) were attributed to acute myocardial infarction, and 18 (12 men and six women) were attributed to chronic ischemic heart disease in subjects with a history of angina pectoris or myocardial infarction. Of the CHD deaths, 17 occurred in hospitals, three in nursing homes, and eight elsewhere on the reservation. Crude (unadjusted) incidence rates of fatal CHD are presented in Table 1. Risk Factors for Fatal CHD Age- and sex-specific CHD death rates for diabetic Pimas are presented in Table 2. The incidence of fatal CHD was significantly higher in men than in women, with a male-to-female ratio of 3.0 (95% confidence interval, ) when controlled for age and duration of diabetes; and the incidence increased with age (,2=21.8, df=3, p<0.001; controlled for sex and duration of diabetes). Duration of diabetes was strongly associated with fatal CHD (X2=25.9, df=4, p<0.001; controlled for age and sex), with 21 (75%) of the 28 deaths attributed to CHD occurring in subjects with diabetes of 15 years or longer. Figure 1 presents the incidence of fatal CHD, according to duration of diabetes, in men and women. The incidence not only was greater with increasing age but also was consistently and significantly higher in subjects with diabetes of longer duration, regardless of age. The effects of other variables that are potential risk factors for fatal CHD in diabetic Pima Indians Nelson et al Fatal Coronary Heart Disease 989 TABLE 1. Person-Years at Risk, Number of Deaths From Coronary Heart Disease, and Incidence of Fatal Coronary Heart Disease, by Sex, in Diabetic and Nondiabetic Pima Indians Aged 5 Years or Older CHD Subjects Person-years deaths Diabetes Sex (n) at risk (n) Incidence* No Men 1,857 12, Women 1,878 12, Yes Men 452 2, Women 641 4, Total 4,828 32, CHD, coronary heart disease. *Deaths from coronary heart disease per 1,000 person-years at risk. aged 45 years or older (no CHD deaths occurred in younger subjects) are given in Tables 3 and 4. Of these variables, proteinuria, renal insufficiency, medial arterial calcification, diabetic retinopathy, an abnormal ECG indicative of myocardial ischemia or previous infarction by the Tecumseh criteria,15 and treatment with insulin were each significant predictors of fatal CHD (controlled for age, sex, and diabetes duration). Major ECG abnormalities, either by the Whitehall28 or the US Pooling Project criteria,29 had similar predictive value for CHD death (data not shown). On the other hand, smoking, hypertension (systolic blood pressure.140 mm Hg or diastolic blood pressure >90 mm Hg or treatment with antihypertensive medicine), 2-hour postload plasma glucose and total serum cholesterol concentrations, body mass index (weight in kilograms divided by the square of height in meters), or treatment with oral hypoglycemic medicine were not significant predictors of fatal CHD. Table S presents the distribution at baseline examination of blood pressure and total serum cholesterol concentrations in Pima Indians aged 45 years or older. Comparison With the Framingham Study Population Age- and sex-specific incidence rates of fatal CHD in Pima Indians aged years were compared with rates from Framingham, Massachusetts (Table 6 and Figure 2) for the period from 1975 through In the Framingham study population, the CHD death rate in diabetic subjects was 3.0 (95% confidence interval, ) that in those without diabetes, whereas among the Pima Indians, fatal CHD occurred only in the diabetic subjects. After controlling for diabetes, age, and sex, the incidence rate of fatal CHD in Pima Indians was less than one half that found in the Framingham population (incidence rate ratio, 0.4; 95% confidence interval, ). Furthermore, the rate in the diabetic Pima Indians was only half that in diabetic subjects from Framingham (incidence rate ratio=0.5; 95% confidence interval, 0.3 to 1.0, controlled for age and sex).

4 990 Circulation Vol 81, No 3, March 1990 TABLE 2. Incidence of Fatal Coronary Heart Disease in Diabetic Pima Indians by Age and Sex Men Women Age Person-years CHD deaths Person-years CHD deaths (yr) at risk (n) Incidence* at risk (n) Incidence* , , , > Total 2, , CHD, coronary heart disease. *Deaths from coronary heart disease per 1,000 person-years at risk. Discussion Comparisons With Other Populations In the United States, CHD is the leading underlying cause of death, and accounts for more than one third of the deaths among diabetic adults more than 40 years of age.1 In nondiabetic Pima Indians, deaths attributable to CHD are exceedingly rare; none occurred in the present study among 3,735 nondiabetic subjects, 419 of whom were 45 years of age or older, observed for as long as 10 years. Moreover, even among diabetic subjects, the CHD death rate in Pima Indians was only one half that in the primarily Caucasian population of Framingham. These results concur with previous observations in Pima Indians of a low prevalence of CHD in life and at death, based on clinical, ECG, and postmortem findings.4-6 Lower rates of CHD mortality have also been reported in other southwestern US Indian populations.30'31 The rate of nonfatal myocardial infarction, however, appears to be rising in the Pima, Navajo, and other southwestern US Indian tribes.32'33 In the Pimas, the overall mortality is nearly twice that observed in the general US population.3 The leading underlying causes of death, grouped according to major categories, are accidents, cirrhosis, homicide, suicide, and diabetes; most deaths in the diabetes category are due to diabetic nephropathy j Mn Dkbbe Dumbn v~ --- ysams ( < k75 Age (b) FIGURE 1. Plots of incidence (deaths/1,000 person-yr) of fatal coronary heart disease in Pima Indians, by age and sex, as a function of diabetes duration. In the present study, 41 of 42 deaths included in the diabetes category were due to diabetic nephropathy, which accounted for 21% of deaths in the diabetic subjects. By contrast, fewer than 5% of deaths in Caucasians with non-insulin-dependent diabetes are attributable to diabetic renal disease.35'36 Although the incidence analysis used in this study would not be influenced by deaths from causes unrelated to CHD (e.g., accidental and violent deaths), it is possible that certain causes of death with similar determinants to those of CHD, that, therefore, may compete with CHD as a cause of death, could lower the rate of fatal CHD. Factors that were shown to predict fatal CHD among the diabetic subjects in the present study have also been shown to predict the development of diabetic nephropathy in the Pimas.37 Thus, among persons with diabetes, those with proteinuria, renal insufficiency, medial arterial calcifications, longer duration of diabetes, and retinopathy were at higher risk of dying from either end-stage renal disease or CHD. Hence, those at greatest risk for fatal CHD in the present study could have been selectively removed by dying from renal disease before developing CHD. This competing cause of death, however, would not explain the lack of fatal CHD among the nondiabetic Pimas. Racial differences may be partially responsible for the differences in CHD mortality observed in this study because wide variations in CHD incidence have been reported around the world.38 In contrast with the experience of southwestern Indians, the Sioux and other Northern Plains Indians have higher rates of morbidity and mortality from acute myocardial infarction than the general US population.39 On the other hand, the incidence of nonfatal myocardial infarction and fatal CHD in men of Japanese heritage and from Puerto Rico is less than one half that reported for US Caucasian men.40,41 Furthermore, in the United States, a lower incidence of fatal and nonfatal CHD has been noted for blacks than whites42; in Trinidad, a higher incidence has been reported for Asian Indians than for persons of European descent.43 Although these differences may, in part, be attributed to differences in environmental risk factors, especially diet,40,44'45 racial factors may

5 TABLE 3. Person-Years at Risk, Number of Deaths From Coronary Heart Disease, and Incidence-Rate Ratios for Dichotomous Variables Representing Potential Risk Factors for Fatal Coronary Heart Disease in Diabetic Pima Indians 45 Years of Age or Older CHD Adjusted Variable Person-years at risk* deaths (n) incidence-rate ratio Proteinuriat No 3, Yes ( ) Renal insufficiency* No 4, Yes ( ) Medical arterial calcification'6 No 2,836 5 Yes 1, ( ) Hypertension No 1,744 7 Yes 2, ( ) Retinopathy H No 2, Yes 1, ( ) Smoking No 3, Yes ( ) Abnormal ECG by Tecumseh criteria15 No 3, Yes ( ) Treatment protocol** No drug 1,996 8 Oral medicine 1, ( ) Insulin ( ) CHD, coronary heart disease; ECG, electrocardiogram. Incidence-rate ratios are adjusted for age, sex, and duration of diabetes. *Total number of person-years for each variable differ slightly because of missing values. turine protein-to-creatinine ratio of 1.0 mg protein/mg creatinine, equivalent to a total protein excretion rate of approximately 1 g/day.11,12 tserum creatinine concentration >177,uM (.2.0 mg/dl). Systolic blood pressure. 140 mm Hg, diastolic blood pressure.90 mm Hg, or treatment with antihypertensive medicine. litwo cases of fatal CHD not included due to missing retinopathy data. History of smoking any amount within 1 year of biennial examination. **Incidence-rate ratios computed for oral medicines/no drug; and insulin/no insulin. also be major influences. As descendants of the Paleo Indians who migrated from northeastern Asia,46,47 the Pimas have Asiatic heritage that may contribute to their low CHD death rate, compared with Caucasians. Different criteria for diabetes may account for some of the differences between the Pima Indian and Framingham populations in fatal CHD rates among persons classified as having diabetes. In the Framingham population, the diagnosis of diabetes was based on a history of treatment with oral hypoglycemic Nelson et al Fatal Coronary Heart Disease 991 TABLE 4. Person-Years at Risk, Number of Deaths From Coronary Heart Disease, and Tests of Association for Continuous Variables Representing Potential Risk Factors for Fatal Coronary Heart Disease in Diabetic Pima Indians 45 Years of Age or Older CHD Person-years deaths Significance Variable at risk* (n) test Two-hour postload plasma glucose (mm)t <16.0 (<288) 1, ( ) 1,436 7 X2 =2.8,p=0.2 >23.0 (>414) 1,536 9 Serum cholesterol (mm)t <4.2 (<163) 1, ( ) 1,513 6 x2=2.6,p=0.3 >4.9 (>190) 1, Body mass index (kg/m2) <27 1, ,475 7 x2=3.3, p=0.2 >32 1,383 6 CHD, coronary heart disease. Age, sex, and duration of diabetes are controlled in the test of statistical significance. *Total number of person-years for each variable differ slightly because of missing values. ttwo-hour postload plasma glucose and total serum cholesterol concentrations in mg/dl are shown in brackets. agents or insulin or on having elevated casual blood glucose concentration (.8.4 mm [150 mg/dl]) on two successive research examinations combined with other corroborative evidence.48 In the Pimas, diabetes was diagnosed according to WHO criteria,9 based on the plasma glucose concentration 2 hours after the 75 -g carbohydrate load administered at each biennial examination. Indeed, some subjects diagnosed by either Framingham or WHO criteria might not be considered diabetic by the other criteria. Moreover, possible differences in diabetes duration between Pima Indians and subjects from Framingham could not be evaluated because information on duration was unavailable for the Framingham population. During the study, eight Pimas died from undetermined causes (ICD-9 codes and 799.9); six (five men and one woman) had diabetes, and two (both men) did not. It is possible that CHD was the underlying cause of death in some of these subjects. If so, their omission would cause an underestimation of the rate of fatal CHD among the Pimas, but such underestimation could also occur in the Framingham population. Nevertheless, even if the seven illdefined deaths occurring in Pima subjects aged years were considered to be deaths from CHD, the rate of death attributed to CHD in Pima Indians would still be significantly lower than that observed in the Framingham population. Risk Factors Diabetes is the main predictor of CHD death in Pima Indians as all cases occurred in diabetic subjects. Previous studies have shown a relation between

6 992 Circulation Vol 81, No 3, March 1990 TABLE 5. Mean Values at Baseline Examination of Systolic and Diastolic Blood Pressures and Total Serum Cholesterol Concentrations in Pima Indians Aged 45 Years or Older by Age, Sex, and Presence or Absence of Diabetes Diabetic Nondiabetic Systolic Diastolic Total serum Systolic Diastolic Total serum Age pressure pressure cholesterol* pressure pressure cholesterol* (yr) Sex (mm Hg) (mm Hg) (mm) (mm Hg) (mm Hg) (mm) Men 141 ±21 87± ±+ 1.1 (195+44) (192+32) Women (189+40) ( ) Men (181±35) 137±18 83± (176+37) Women ±0.9 (190±35) 140±23 81±13 4.7±0.9 (182+34) Men 147± ±0.8 (167±30) 146± (188±37) Women 151± ±1.0 (183±40) 140±23 74± (182±32).75 Men 152±24 81±9 4.6±0.9 (177±35) ±14 4.4±0.9 (171±35) Women 161±26 80±10 5.0±1.0 (193+37) 156±+15 78±11 4.7±0.9 (179±36) Values are given as mean±sd. *Total serum cholesterol concentration in mg/dl shown in parentheses. diabetes and CHD mortality.1,2 This finding may result from a greater frequency of CHD among diabetic subjects, a higher case-fatality ratio from CHD in diabetic persons, or both. Risk factors for fatal CHD in diabetic Pima Indians include advancing age, male sex, duration of diabetes, treatment with insulin, and presence of proteinuria, renal insufficiency, medial arterial calcification, diabetic retinopathy, or abnormal ECGs. Some factors commonly believed to be associated with CHD mortality in persons with and without diabetes, such as total serum cholesterol concentration, hypertension, and smoking,49-51 were not, however, found to be predictive in Pima Indians. Because less than 1% of adult Pimas smoke one pack or more of cigarettes a day, an association, if one were to exist, might not be recognized. Furthermore, the absence of an association between fatal CHD and serum cholesterol concentration or hypertension may be due to the small size of the study. Hypertension, clearly a risk factor for CHD in nondiabetic persons has been found in other studies to have a similar or higher relative impact in persons with diabetes.49,52,53 Pima Indian men and women have lower serum concentrations of total and low density lipoprotein cholesterol than US Caucasians, and the levels do not increase with age in Pima men more than 25 years old, as occurs in Caucasians.54 On the other hand, high density lipoprotein cholesterol concentrations are also low in the Pimas and are similar in both sexes despite the higher incidence of death from CHD in men. Plasma triglyceride concentrations are higher in Pimas than in Caucasians, and diabetes is associated with elevated total and very low density lipoprotein triglyceride concentrations.55 In the present study, the only lipid measurement was of total serum cholesterol concentration (Table 5). Differences in plasma lipoprotein concentrations between Pimas TABLE 6. Comparison of Person-Years at Risk, Number of Deaths From Coronary Heart Disease, and Incidence Rates of Fatal Coronary Heart Disease in Pima Indian and Framingham Populations, by Age, Sex, and Presence or Absence of Diabetes Pima Indians Framingham population Diabetic Diabetic Nondiabetic CHD CHD CHD Age Person-years deaths Incidence* Person-years deaths Incidence* Person-years deaths Incidence* (yr) Sex at risk (n) (CI) at risk (n) (CI) at risk (n) (CI) Men , Women 1, , Men , Women , Men , Women , All ages Men 1, , , Women 2, , , Both 3, , , Age- and sex-adjusted t 3, ( ) 2, ( ) 27, ( ) CHD, coronary heart disease; CI, confidence interval. *Deaths from CHD per 1,000 person-years at risk. tage- and sex-adjusted to the 1980 US population (95% confidence interval).

7 '~0 15 D-abetes No Nws Pmo From Pio From FIGURE 2. of fatal coronary heart disease in Pima Indians and in the Framingham population (Fram) aged years, ageadjusted to the 1980 US population, according to sex, and according to the presence or absence of diabetes. Bar graphs of incidence (deaths/1,000person-yr) and Caucasians may be partly responsible for the differences in CHD death rates observed in this study, although no significant relation to total serum cholesterol was found in Pimas. Urinary albumin excretion and CHD are associated with both insulin-49,56 and non-insulin-dependent diabetes,5758 and proteinuria was associated with increased total and cardiovascular mortality in Pima Indians.34 Further, diabetic retinopathy and cardiovascular disease were associated in persons with noninsulin-dependent diabetes in the Framingham population59 as well as in the present study. The mechanism of the association between cardiovascular or, more specifically, CHD mortality and the small vessel diseases of diabetes is unknown. It has been proposed, however, that the microangiopathy of diabetes, which is commonly believed to affect the eyes and the kidneys simultaneously, could also affect the heart, could contribute to the pathogenesis of CHD, or could increase the risk of death from CHD in persons with diabetes.1"59,60 Microvascular changes that have been observed in the hearts of diabetic persons include intimal proliferation, thickened capillary walls, and perivascular fibrosis, as well as the typical diabetic changes (i.e., thickening of the capillary basement membrane and microaneurysms) Medial arterial calcification frequently is observed in diabetic Pima Indians.16 Its presence is associated with a greater incidence of many end-stage diabetic complications such as nephropathy, retinopathy, lower extremity amputations, and death16'67 as well as with fatal CHD, as shown in the present study. Treatment with insulin was associated with a substantially increased risk of death from CHD in Pima Indians. Similar observations have been reported by Kleinman et al,2 who found that diabetic persons treated with insulin, as compared with those not receiving medicines for the treatment of diabetes, had twice the risk of ischemic heart disease mortality. Nelson et al Fatal Coronary Heart Disease 993 The University Group Diabetes Program,68 in which treatment with insulin or other regimens for noninsulin-dependent diabetes was randomly assigned, reported that the occurrence of fatal and nonfatal myocardial infarction did not differ substantially between subjects treated with insulin in a fixed or variable dose and those on no drug therapy. The lack of difference between these treatment groups suggests that insulin treatment per se is not a risk factor for CHD but that the association of CHD with insulin treatment is the result of its more frequent use in those with more severe diabetes. On the other hand, Stout and Vallance-Owen69 suggested that insulin is atherogenic, and a number of studies have lent support to this hypothesis (reviewed in Stout).70 Resistance to the action of insulin has been suggested as a factor contributing to increased CHD in subjects with impaired glucose tolerance in the Whitehall study.71 However, in the present study, fatal CHD did not occur in any subjects with impaired glucose tolerance, although Pima Indians with impaired glucose tolerance are characterized by insulin resistance and hyperinsulinemia.72,73 Longer duration of diabetes was associated with an increased incidence of fatal CHD in Pima Indians. Seventy-five percent of the diabetic subjects whose deaths were attributed to CHD had had diabetes for at least 15 years. An association between cardiovascular mortality and diabetes duration was also observed by Kleinman et a12 for the US Caucasian population aged years. Nevertheless, Jarrett and Shipley74 have argued that increased cardiovascular disease mortality among persons with noninsulin-dependent diabetes could be the result of greater exposure to other cardiovascular risk factors that might precede the onset of diabetes and that a relation with diabetes duration reflects increasing age and exposure to other factors rather than to diabetes per se. The duration of diabetes, as determined in many studies, however, is questionable. On the other hand, the duration of diabetes in Pima Indians was more precisely estimated because of the periodic glucose tolerance testing of the population, and a relation with diabetes duration was found that was independent of age and sex. Conclusions Pima Indians, despite their high prevalence of diabetes, had a low incidence of fatal CHD when compared with the primarily Caucasian population of Framingham. Diabetes was a major risk factor for fatal CHD in Pimas, as no CHD deaths occurred in the nondiabetic population during 10 years of observation. Risk factors for fatal CHD in diabetic Pima Indians included older age, male sex, diabetes duration, insulin treatment, proteinuria, renal insufficiency, medial arterial calcification, diabetic retinopathy, and ECG abnormalities. On the other hand, the traditional CHD risk factors, such as total serum cholesterol concentration, hypertension, and smoking, did not predict CHD in the diabetic Pimas. The smaller contribution of these risk factors in this population apparently magnifies the importance of

8 994 Circulation Vol 81, No 3, March 1990 diabetes as a risk factor for CHD. As CHD death occurred only among diabetic subjects and was strongly associated with the duration and type of treatment of diabetes and with other specific diabetes complications, fatal CHD in the Pimas showed a pattern of risk factors closely resembling those of the specific microvascular complications of diabetes. Comparison of the Pima and Framingham populations suggests that the impact of non-insulindependent diabetes on CHD mortality in different populations is modulated by other factors whose identities are unknown. Acknowledgments The authors are indebted to the members of the Gila River Indian Community for participating in this investigation; to the staff of the Diabetes and Arthritis Epidemiology Section, NIDDK, for conducting the examinations and processing data; to the staff of the Phoenix Indian Medical Center and Public Health Service Indian Hospital, Sacaton, for use of medical records; to Dr. Fredrick A. Rose for interpreting the soft tissue radiographs; and to Robert Garrison and the other Framingham Study investigators for making summary estimates of the Framingham cohort CHD experience available. References 1. Barrett-Connor E, Orchard T: Diabetes and heart disease, in Diabetes in America: Diabetes Data Compiled National Diabetes Data Group, US Dept of Health and Human Services, NIH publication No , 1985, pp Kleinman JC, Donahue RP, Harris MI, Finucane FF, Madans JH, Brock DB: Mortality among diabetics in a national sample. Am J Epidemiol 1988;128: 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: Sievers ML: Myocardial infarction among southwestern American Indians. Ann Intem Med 1967;67: Ingelfinger JA, Bennett PH, Liebow IM, Miller M: Coronary heart disease in the Pima Indians: Electrocardiographic findings and postmortem evidence of myocardial infarction in a population with a high prevalence of diabetes mellitus. Diabetes 1976;25: Howard BV, Lisse JR, Knowler WC, Davis MP, Pettitt DJ, Bennett PH: Diabetes and atherosclerosis in the Pima Indians. Mt Sinai J Med (NY) 1982;49: The Framingham Study: Some Risk Factors Related to the Annual Incidence of Cardiovascular Disease and Death Using Pooled Repeated Biennial Measurements: Framingham Heart Study, 30-Year Followup. DHHS, NIH publication No , February Bennett PH, Burch TA, Miller M: Diabetes mellitus in American (Pima) Indians. Lancet 1971;2: World Health Organization: Diabetes Mellitus. Technical report series 727. Geneva, WHO, 1985, pp Shevky MC, Stafford DD: A clinical method for the estimation of protein in urine and other body fluids. Arch Intern Med 1923;32: Shaw AB, Risdon P, Lewis-Jackson JD: Protein creatinine index and Albustix in assessment of proteinuria. Br Med J 1983;287: Ginsberg JM, Chang BS, Matarese RA, Garella S: Use of single voided urine samples to estimate quantitative proteinuria. N Engl J Med 1983;309: Blackburn H, Keys A, Simonson E, Rautaharju P, Punsar S: The electrocardiogram in population studies: A classification system. Circulation 1960;21: Rose GA, Blackburn H: Cardiovascular Survey Methods. World Health Organization, Geneva, 1968, monograph series No. 56, pp Epstein FH, Ostrander LD Jr, Johnston BC, Payne MW, Hayner NS, Keller JB, Francis T Jr: Epidemiological studies of cardiovascular disease in a total community-tecumseh, Michigan, Ann Intem Med 1965;62: Everhart JE, Pettitt DJ, Knowler WC, Rose FA, Bennett PH: Medial arterial calcification and its association with mortality and complications of diabetes. Diabetologia 1988;31: Lachman AS, Spray TL, Kerwin DM, Shugell GI, Roberts WC: Medical calcinosis of Monckeberg: A review of the problem and a description of a patient with involvement of peripheral, visceral and coronary arteries. Am J Med 1977; 63: Data Preparation Branch: Vital Statistics Instruction Manual: Part 2a. Instructions for Classifying the Underlying Cause of Death, Hyattsville, Md, DHHS, PHS, National Center for Health Statistics, Data Preparation Branch: Vital Statistics Instruction Manual: Part 2b. Instructions for Classifying Multiple Causes of Death, Hyattsville, Md, DHHS, PHS, National Center for Health Statistics, Data Preparation Branch: Vital Statistics Instruction Manual: Part 2c. ICD-9 ACNE Decision Tables for Classifying Underlving Causes of Death, Hyattsville, Md, DHHS, PHS. National Center for Health Statistics, Data Preparation Branch: Vital Statistics Instruction Manual: Part 4. Demographic Classification and Coding Instructions for Death Records, Hyattsville, Md, DHHS, PHS, National Center for Health Statistics, Physicians' Handbook on Medical Certification: Death, Birth, Fetal Death. Hyattsville, Md, DHHS, PHS, National Center for Health Statistics, 1978 (reprinted 1980), pp 1-30 (DHHS publication No. PHS ) 23. The Intemational Classification of Diseases, Clinical Modification, ed 9. Ann Arbor, Mich, Edwards, The Lipid Research Clinics Program: The coronary primary prevention trial: Design and implementation. J Chronic Dis 1979;32: Mantel N, Haenszel W: Statistical aspects of the analysis of data from retrospective studies of disease. J Nat Cancer Inst 1959;22: Rothman KJ, Boice JD: Follow-up (cohort) studies, in: Epidemiologic Analysis With a Programmable Calculator. Washington, DC, National Institutes of Health, 1979, NIH publication No Landis JR, Heyman ER, Koch GG: Average partial association in three-way contingency tables: A review and discussion of alternative tests. Int Stat Rev 1978;46: Reid DD, Brett GZ, Hamilton PJS, Jarrett RJ, Keen H, Rose G: Cardiorespiratory disease and diabetes among middleaged civil sei :ants: A study of screening and intervention. Lancet 1974;1: Pooling Project Research Group: Relationship of blood pressure, serum cholesterol, smoking habit, relative weight and ECG abnormalities to incidence of major coronary events: Final report of the Pooling Project. J Chronic Dis 1978;31: Becker TM, Wiggins C, Key CR, Samet JM: Ischemic heart disease mortality in Hispanics, American Indians, and non- Hispanic Whites in New Mexico. Circulation 1988; 78: Coulehan J, Lerner G, Helzlsouer K, Welty TK, McLaughlin J: Acute myocardial infarction among Navajo Indians, Am J Public Health 1986;76: Sievers ML, Fisher JR: Increasing rate of acute myocardial infarction in southwestern American Indians. Ariz Med 1979; 36:

9 33. Klain M, Coulehan JL, Arena VC, Janett R: More frequent diagnosis of acute myocardial infarction among Navajo Indians. Am J Public Health 1988;78: Nelson RG, Pettitt DJ, Carraher MJ, Baird HR, Knowler WC: Effect of proteinuria on mortality in NIDDM. Diabetes 1988;37: Balodimus MC: Diabetic nephropathy, in Marble A, White P, Bradley RF, Krall LP (ed): Joslin's Diabetes Mellitus, ed 11. Philadelphia, Lea & Febiger, 1971, p Fabre J, Balant LP, Dayer PG, Fox HM, Vernet AT: The kidney in maturity onset diabetes mellitus: A clinical study of 510 patients. Kidney Int 1982;21: Kunzelman CL, Knowler WC, Pettitt DJ, Bennett PH: Incidence of proteinuria in type 2 diabetes mellitus in the Pima Indians. Kidney Int 1989;35: Keys A: Coronary heart disease in seven countries. Circulation 1970;41(suppl I):I Hrabovsky SL, Welty TK, Coulehan JL: Acute myocardial infarction and sudden death in Sioux Indians. West J Med 1989;150: Knutsen R, Knutsen SF, Curb JD, Reed DM, Kautz JA, Yano K: The predictive value of resting electrocardiograms for 12-year incidence of coronary heart disease in the Honolulu Heart Program. J Clin Epidemiol 1988;41: Gordon T, Garcia-Palmieri MR, Kagan A, Kannel WB, Schiffman J: Differences in coronary heart disease in Framingham, Honolulu and Puerto Rico. J Chronic Dis 1974;27: Watkins LO, Neaton JD, Kuller LH: Racial differences in high-density lipoprotein cholesterol and coronary heart disease incidence in the usual-care group of the Multiple Risk Factor Intervention Trial. Am J Cardiol 1986;57: Beckles GLA, Miller GJ, Kirkwood BR, Alexis SD, Carson DC, Byam NTA: High total and cardiovascular disease mortality in adults of Indian descent in Trinidad, unexplained by major coronary risk factors. Lancet 1986;1: Yano K, Reed DM, McGee DL: Ten-year incidence of coronary heart disease in the Honolulu Heart Program: Relationship to biologic and lifestyle characteristics. Am J Epidemiol 1984;119: Gordon T, Kagan A, Garcia-Palmieri M, Kannel WB, Zukel WJ, Tillotson J, Sorlie P, Hjortland M: Diet and its relationship to coronary heart disease and death in three populations. Circulation 1981;63: Turner CG: Dental evidence for the peopling of the Americas, in Shutler R (ed): Early Man in the New World. Beverly Hills, Calif, Sage Publications, 1983, pp Williams RC, Steinberg AG, Gershowitz H, Bennett PH, Knowler WC, Pettitt DJ, Butler W, Baird HR, Dowda-Rea L, Burch TA, Morse HG, Smith C: Gm allotypes in Native Americans: Evidence for three distinct migrations across the Bering land bridge. Am J Phys Anthropol 1985;66: Shurtleff D: Some Characteristics Related to the Incidence of Cardiovascular Disease and Death: Framingham Study, 18 Year Follow-up. DHEW publication No. (NIH) , Jensen T, Borch-Johnsen K, Kofoed-Enevoldsen A, Deckert T: Coronary heart disease in young type 1 (insulindependent) diabetic patients with and without diabetic nephropathy: Incidence and risk factors. Diabetologia 1987; 30: Kannel WB, McGee DL: Diabetes and cardiovascular risk factors: The Framingham study. Circulation 1979;59: Uusitupa M, Siitonen 0, Pydrali K, Aro A, Hersio K, Penttila I, Voutilainen E: The relationship of cardiovascular risk factors to the prevalence of coronary heart disease in newly diagnosed type 2 (non-insulin-dependent) diabetes. Diabetologia 1985;28: 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. Br Med J 1983;2: Nelson et al Fatal Coronary Heart Disease Aromaa A, Reunanen A, Pydralai K: Hypertension and mortality in diabetic and non-diabetic Finnish men. J Hypertens 1984;2:S205-S 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: Howard BV, Knowler WC, Vasquez B, Kennedy AL, Pettitt DJ, Bennett PH: Plasma and lipoprotein cholesterol and triglyceride in the Pima Indian population: Comparison of diabetics and nondiabetics. Arteriosclerosis 1984;4: Borch-Johnsen K, Andersen PK, Deckert T: The effect of proteinuria on relative mortality in type 1 (insulindependent) diabetes mellitus. Diabetologia 1985;28: Mattock MB, Keen H, Viberti GC, El-Gohari MR, Murrells TJ, Scott GS, Wing JR, Jackson PG: Coronary heart disease and urinary albumin excretion rate in type 2 (noninsulin-dependent) diabetic patients. Diabetologia 1988; 31: Jarrett RJ, Viberti GC, Argyropoulos A, Hill RD, Mahmud U, Murrells TJ: Microalbuminuria predicts mortality in noninsulin-dependent diabetes. Diabetic Med 1984;1: Hiller R, Sperduto RD, Podgor MJ, Ferris FL, Wilson PWF: Diabetic retinopathy and cardiovascular disease in type II diabetics. Am J Epidemiol 1988;128: Deckert T, Feldt-Rasmussen B, Borch-Johnsen K, Jensen T, Kofoed-Enevoldsen A: Albuminuria reflects widespread vascular damage: The Steno hypothesis. Diabetologia 1989; 32: Hamby RI, Zoneraich S, Sherman L: Diabetic cardiomyopathy. JAMA 1974;229: Pearce MB, Bulloch RT, Kizziar JC: Myocardial small vessel disease in patients with diabetes mellitus. Circulation 1973; 48(suppl IV):IV Zoneraich S, Silverman G, Zoneraich 0: Primary myocardial disease, diabetes mellitus, and small vessel disease. Am Heart J 1980;100: Ledet T: Diabetic cardiopathy. Quantitative histological studies of the heart amongst young juvenile diabetics. Acta Pathol Microbiol Scand 1976;84: Fischer VW, Barner HB, Leskiw ML: Capillary basal laminar thickness in diabetic human myocardium. Diabetes 1979; 28: Factor SM, Okun EM, Minase T: Capillary microaneurysms in the human diabetic heart. N Engl J Med 1980;302: Nelson RG, Gohdes DM, Everhart JE, Hartner JA, Zwemer FL, Pettitt DJ, Knowler WC: Lower-extremity amputations in NIDDM: 12-yr follow-up study in Pima Indians. Diabetes Care 1988;11: The University Group Diabetes Program: Effects of hypoglycemic agents on vascular complications in patients with adult-onset diabetes: VIII. Evaluation of insulin therapy: Final report. Diabetes 1982;31: Stout RW, Vallance-Owen J: Insulin and atheroma. Lancet 1979;1: Stout RW: Insulin and atheroma-an update. Lancet 1987; 1: Fuller JH, Shipley MJ, Rose G, Jarrett RJ, Keen H: Coronary-heart-disease risk and impaired glucose tolerance: The Whitehall Study. Lancet 1980;1: Lillioja S, Mott DM, Howard BV, Bennett PH, Yki-Jarvinen H, Freymond D, Nyomba BL, Zurlo F, Swinburn B, Bogardus C: Impaired glucose tolerance as a disorder of insulin action: Longitudinal and cross-sectional studies in Pima Indians.N EnglJMed 1988;318: Saad MF, Knowler WC, Pettitt DJ, Nelson RG, Mott DM, Bennett PH: Sequential changes in serum insulin concentration during development of non-insulin-dependent diabetes. Lancet 1989;1: Jarrett RJ, Shipley MJ: Type 2 (non-insulin-dependent) diabetes mellitus and cardiovascular disease-putative association via common antecedents: Further evidence from the Whitehall Study. Diabetologia 1988;31: KEY WORDS * risk factors * coronary heart disease * mortality * diabetes mellitus * Indians, American

10 Low incidence of fatal coronary heart disease in Pima Indians despite high prevalence of non-insulin-dependent diabetes. R G Nelson, M L Sievers, W C Knowler, B A Swinburn, D J Pettitt, M F Saad, I M Liebow, B V Howard and P H Bennett Circulation. 1990;81: doi: /01.CIR Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 1990 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: published Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Circulation is online at: