Hyperinsulinemia and Cardiovascular Disease in Elderly Men

The Honolulu Heart Program

From the Honolulu Epidemiology Research Unit, Epidemiology and Biometry Program, Division of Epidemiology and Clinical Applications, National Heart, Lung, and Blood Institute, Honolulu, HI (C.M.B., D.S.S.); the Honolulu Heart Program, Kuakini Medical Center (J.D.C., B.L.R., K.Y.); and Department of Medicine, the John A. Burns School of Medicine, University of Hawaii at Manoa, Honolulu, HI (J.D.C., B.L.R., R.A.); and the Division of Biostatistics, University of Virginia School of Medicine, Charlottesville (R.D.A.).

	Abstract

Top Abstract Introduction Methods Results Discussion References

 
Abstract—Hyperinsulinemia has been associated with cardiovascular disease (CVD), but whether this relation is independent of other CVD risk factors is uncertain. Most studies have focused on coronary heart disease (CHD), but few have included peripheral vascular disease (PVD) and stroke. Moreover, evidence in elderly and minority populations is limited. Between 1991 and 1993, 3562 elderly (71 to 93 years) Japanese-American men from the Honolulu Heart Program were examined and had fasting insulin levels measured. Hyperinsulinemia, defined as a fasting insulin 95th percentile among nonobese men with normal glucose tolerance and no diabetic history or medication use, was observed in 22% of the population. Subjects with hyperinsulinemia had a more adverse CVD risk factor profile and had higher age-adjusted prevalences of CHD, angina, PVD, thromboembolic stroke, and hemorrhagic stroke compared with those without hyperinsulinemia. Age-adjusted fasting insulin levels but not 2-hour levels were also significantly elevated (P<.01) in those with prevalent CVD compared with those without. In logistic regression analyses, adjustment for multiple CVD risk factors attenuated the relations of hyperinsulinemia with CHD, angina, and PVD to nonsignificant levels, whereas those involving thromboembolic and hemorrhagic stroke were strengthened and remained significant (odds ratios=2.27 and 7.53, 95% confidence intervals=1.25 to 4.13 and 1.65 to 34.25, respectively). When multivariate analyses were restricted to nondiabetic subjects, associations were slightly weaker and in general nonsignificant. Nondiabetic men with thromboembolic stroke were twice as likely to have hyperinsulinemia as those who were stroke-free, although this association was of borderline significance (odds ratio=1.99, 95% confidence interval=0.95 to 4.17, P=.069). In subjects with elevated total cholesterol levels, somewhat stronger associations were observed for PVD and stroke but not for CHD. Although further prospective studies are indicated, particularly for PVD and stroke, these cross-sectional results are consistent with an indirect role for insulin in CVD, wherein hyperinsulinemia or an underlying insulin-resistant state may adversely affect other CVD risk factors or serve as a marker for an atherogenic or thrombogenic state.

Key Words: Asian Americans • cardiovascular diseases • insulin

	Introduction

Top Abstract Introduction Methods Results Discussion References

 
The role of elevated insulin levels in CVD is uncertain. More than 15 years ago, several prospective epidemiological studies demonstrated an independent association between insulin concentrations and CHD.^{1 2 3} More recent investigations,^{4 5 6} but not all,⁷ suggest that the association may not be independent of other known CVD risk factors. Some of the discordant findings may be due to the fact that these early studies did not take into account HDL cholesterol levels.^{7 8 9}

Whether hyperinsulinemia is associated with CVD has not been widely studied in the elderly nor in subsets of the population who might be at increased risk. For example, relatively low levels of total and LDL cholesterol could account for the paradoxically low rates of CHD despite a high prevalence of glucose intolerance in some populations.^{10 11 12} Moreover, although associations of insulin with CHD have been examined in a number of studies, those involving stroke¹³ and PVD^{14 15 16} have been assessed infrequently.

The purpose of this investigation was to assess the association of hyperinsulinemia with several clinical and subclinical manifestations of CVD in elderly subjects. A recent examination of the entire Honolulu Heart Program cohort between 1991 and 1993 provided the opportunity to examine these relations in >3500 elderly Japanese-American men by using a cross-sectional design. Adjustment for other CVD risk factors was performed to determine whether these associations were independent or whether they might be mediated through an effect of insulin on other risk factors. Associations were also examined in nondiabetic subjects and in subsets of the population with elevated total cholesterol who might be at increased CVD risk.

	Methods

Top Abstract Introduction Methods Results Discussion References

 
Study Population
The Honolulu Heart Program is a prospective epidemiological study initiated in 1965 to determine risk factors for heart disease and stroke in a well-defined population of middle-aged Japanese-American men. Potential subjects identified from selective service records were born between 1900 and 1919 and were living on the island of Oahu in 1965.^{17 18} A total of 8006 men between 45 and 68 years of age completed the baseline examination (1965 to 1968). Subsequent examinations were completed an average of 2 years (1968 to 1970), 6 years (1971 to 1974), and 25 years (1991 to 1993) after the baseline examination.

Among the 8006 initially examined men, 3845 men were reexamined or completed an extended telephone interview between 1991 and 1993. Of 3741 men examined (80% of survivors at that time), 86% of the examinations were performed in a clinic setting, 13% at home, and 1% in nursing homes. A total of 3573 subjects provided fasting blood specimens and 3562 had fasting insulin concentrations measured. These 3562 men, aged 71 to 93 years between 1991 and 1993, constituted the study population.

Data Collection
The examination included demographic, lifestyle (smoking, alcohol, physical activity), medical history, medication use, and psychosocial information as well as physiological, anthropometric, and other laboratory measurements. The methods used in data collection were consistent with assessments made at previous examinations, were in accordance with institutional guidelines, and were approved by the Institutional Review Committee of Kuakini Medical Center.¹⁹ BMI was calculated as weight in kilograms divided by height in meters squared. An index of physical activity was based on the number of hours spent in five activity levels weighted by the estimated oxygen required.²⁰ Hypertension was considered present when blood pressure was 160/95 mm Hg on the basis of an average of two readings or when antihypertensive medication use was reported. Use of diabetic medication (insulin or oral hypoglycemic agents) was ascertained from subjects who reported that they had diabetes.

Subjects were instructed to fast overnight for at least 12 hours. Blood specimens were obtained and shipped to the University of Vermont. After separation of plasma, standard enzymatic measurements of total cholesterol, HDL cholesterol, and triglycerides were performed using the same laboratory and methods that were used in the Cardiovascular Health Study.^{21 22} Cholesterol and triglyceride analyses were performed using an Olympus Demand System (Olympus Corp) and were standardized according to the Centers for Disease Control and Prevention.²² LDL cholesterol was calculated for subjects with triglyceride levels 400 mg/dL on the basis of the method of Friedewald et al.²³ Plasma fibrinogen concentrations were measured with a clot-based end point method and a BBL fibrometer (Becton-Dickinson).²²

After the fasting specimen was obtained, a standard 75-g oral glucose load was administered and a second blood specimen collected 2 hours later. The oral glucose tolerance test was offered to participants who were examined in the clinic, were not taking insulin, and did not report any of the following conditions: stomach resection, active ulcer, stomach cancer, and severely elevated blood pressure (systolic >200 mm Hg or diastolic >115 mm Hg). A total of 2133 subjects fasted at least 12 hours and had both fasting and 2-hour glucose measurements. Glucose was measured by a glucose oxidase method (University of Vermont). Insulin was measured by a double-antibody radioimmunoassay method²⁴ at the University of Washington (Diabetes Endocrinology Research Center Core Radioimmunoassay Laboratory) after storage at -70°C for up to 2 years, with a lower limit of detection of 3 µU/mL and interassay coefficients of variability of 9% and 8% at low (mean, 21 µU/mL) and high (mean, 82 µU/mL) insulin concentrations, respectively.²⁵

Hyperinsulinemia
Hyperinsulinemia was defined as a fasting insulin concentration at or above the 95th percentile of the distribution among a subset of the population who did not have several characteristics known to influence insulin levels, such as glucose intolerance, obesity, and inadequate duration of fasting. Thus, this definition was applied to subjects who met the following specific criteria: (1) normal glucose tolerance by World Health Organization definition,²⁶ (2) negative history and medication use for diabetes, (3) nonobesity (BMI <25.0), and (4) an overnight fast of at least 12 hours. Among the 3562 men who had insulin measured between 1991 and 1993, 2265 men had no missing values for the variables included in these criteria (most of those excluded did not have 2-hour glucose values). Of these subjects, 1554 men had abnormal glucose tolerance (diabetes or impaired glucose tolerance by World Health Organization criteria), had a history of diabetes, or were taking diabetic medication. Of the remaining 711 men, 190 were obese by this definition and 21 had fasted <12 hours (4 of these men met both criteria). Thus, a total of 504 subjects of the 3562 men with fasting insulin measurements (14.1%) met these criteria. The 95th percentile of the fasting insulin concentration for this group of 504 subjects was 20 µU/mL, and an insulin level at or above this value was used to define hyperinsulinemia for the entire population.

Prevalent CVD
Assessment of prevalent manifestations of clinical and subclinical CVD was based on current as well as previously collected information. A comprehensive hospital-based surveillance system has been in existence since initiation of the study in 1965. Both fatal and nonfatal cardiovascular events are identified by a committee on the basis of information from hospital discharge summaries, autopsy reports, and death certificates. The committee uses standardized criteria to ascertain prevalent and incident clinical events, as described in detail previously.^{19 27}

The definition for CHD included (1) myocardial infarction, (2) coronary insufficiency or angina pectoris with surgical intervention (angioplasty or bypass graft) or with angiographic evidence of coronary artery stenosis >70% identified through surveillance, and (3) reported heart attack or angina resulting in hospitalization or surgical treatment at the current examination. Criteria for infarction included a clinically apparent event based on electrocardiographic and cardiac enzyme evidence identified through surveillance, a silent event detected by electrocardiograms at any examination or through surveillance, or a temporal change in electrocardiograms diagnostic of myocardial infarction. Angina was also examined as a separate clinical manifestation and was considered present if it was identified by hospital surveillance using specific criteria¹⁹ or was diagnosed by a physician at any Honolulu Heart Program examination. Angina defined by Rose criteria was also assessed.²⁸

Definite and probable thromboembolic and hemorrhagic strokes (excluding transient ischemic attacks) were identified by hospital surveillance and met specific criteria as described in detail previously.^{19 29} In general, criteria for stroke included a persistent neurological deficit and evidence from cerebrospinal fluid, nuclear brain scans, and CT.

PVD was identified by using a ratio of blood pressures measured by a Doppler device in the ankle to that in the arm.³⁰ The mean of two measurements for the left and right ankle was used as the numerator, and the mean of two measurements over the right brachial artery was used as the denominator. An ABI<0.9 on the left or right was considered evidence of PVD, since it reflects a relatively low ankle pressure that may indicate lower-extremity atherosclerosis.

Statistical Analysis
All subjects were initially categorized by hyperinsulinemia status using the cutpoint established among nonobese subjects with normal glucose tolerance. Age-adjusted mean levels and prevalence of a number of cardiovascular risk factors were compared for subjects with and without hyperinsulinemia by using general linear models for continuous variables and logistic regression for discrete variables.³¹ For descriptive purposes, age-adjusted mean fasting and 2-hour insulin levels were also compared for subjects with and without each of the CVD manifestations. Because insulin level was the dependent variable in this instance and its distribution was skewed,²⁵ a logarithmic transformation was used in statistical comparisons, and the antilogarithm of log₁₀ mean insulin was calculated to express insulin levels in their original units.

In addition, the age-adjusted prevalence of each cardiovascular manifestation was estimated for subjects with and without hyperinsulinemia by using a marginal prediction method that calculates probabilities of disease from a logistic model for each individual on the basis of their age and then averages these probabilities for the entire population.^{32 33} Logistic regression models were also used to assess the association of hyperinsulinemia with each CVD manifestation. ORs and their 95% CIs were calculated from a series of logistic models adjusting for the following covariates: (1) age only; (2) age and BMI; (3) age, BMI, and indicators of diabetes (diabetic medication use and glucose level); (4) previous variables plus hypertension and triglycerides; and (5) previous variables plus smoking, alcohol intake, physical activity, hematocrit, and heart rate. The same logistic analyses were repeated among nondiabetic subjects. Separate logistic analyses were conducted to determine whether or not the association of hyperinsulinemia with CVD was more pronounced at elevated total cholesterol levels (defined by two cutpoints, 240 mg/dL and 200 mg/dL) than at lower cholesterol levels. Interaction terms were also included in logistic regression models to assess effect modification by cholesterol level.

	Results

Top Abstract Introduction Methods Results Discussion References

 
An elevated fasting insulin level (20 µU/mL) was found in 21.7% of these 3562 men. Subjects with hyperinsulinemia were more obese and less physically active; had higher levels of glucose, triglycerides, and hematocrit; and had higher heart rates and a higher prevalence of diabetes and hypertension than did subjects who were not hyperinsulinemic (P<.001; Table 1). Mean systolic and diastolic blood pressures (data not shown), alcohol consumption, and prevalence of current smoking did not differ significantly between the two groups.

For descriptive purposes, age-adjusted mean fasting and 2-hour insulin levels are presented for subjects with and without five manifestations of CVD (Table 2). Fasting, but not 2-hour, insulin levels were significantly higher in subjects with these prevalent disease manifestations than in those without these conditions. Angina defined by Rose criteria was not associated with insulin level (data not shown). Somewhat larger absolute differences in mean fasting insulin levels were observed for both types of stroke than for other manifestations, although fewer events were observed, particularly for hemorrhagic stroke.

The age-adjusted prevalence of all CVD manifestations was significantly higher in subjects with hyperinsulinemia than in those without (Table 3). As expected, CHD, angina, and subclinical evidence of PVD were more prevalent overall than were thromboembolic and hemorrhagic stroke. Compared with subjects who did not have hyperinsulinemia, those who were hyperinsulinemic had a significantly higher prevalence of CHD (15.2% versus 21.8%), angina (10.6% versus 13.7%), PVD (12.5% versus 16.9%), thromboembolic stroke (2.7% versus 4.6%), and hemorrhagic stroke (0.6% versus 1.6%).

Age-adjusted ORs indicating the association of hyperinsulinemia with prevalent CVD were significantly elevated and ranged between 1.3 and 1.6 for CHD, angina, and PVD and were somewhat higher (1.8 and 2.8) for thromboembolic and hemorrhagic stroke, respectively (Table 4). After additional adjustment for BMI, ORs for CHD and angina were slightly reduced and those for PVD and both types of stroke slightly enhanced. Associations remained significant except for angina. With further adjustment for glucose level and use of diabetic medication, ORs for CHD were further attenuated and were no longer significantly elevated, whereas ORs for PVD and stroke were attenuated slightly but remained significantly elevated. Addition of other covariates to the models further reduced associations with CHD and angina. Additional adjustment for triglycerides, HDL cholesterol, and hypertension attenuated the OR for PVD to a nonsignificant level. However, ORs for thromboembolic and hemorrhagic stroke showed independent associations with hyperinsulinemia, even after adjustment for these variables, and these associations were slightly stronger after further adjustment for smoking, alcohol, fibrinogen, hematocrit, physical activity, and heart rate.

Similar analyses are presented in Table 5 for 2801 nondiabetic subjects (ie, excluding those with diabetic history or medication use). In general, the magnitude of the ORs was smaller and not significantly elevated after adjustment for age in nondiabetic subjects than in the entire cohort. Associations of hyperinsulinemia with CHD, angina, and hemorrhagic stroke were not statistically significant after adjustment for multiple CVD risk factors; results for hemorrhagic stroke were limited by a relatively small number of events. Adjustment for age and BMI yielded ORs that were significantly elevated for PVD and nearly so for thromboembolic stroke. With additional adjustment for glucose level, associations were significant for both PVD and thromboembolic stroke. Adjustment for additional CVD risk factors attenuated the ORs for PVD but had little impact on those for thromboembolic stroke. Associations reached borderline significant levels for thromboembolic stroke in the final two models (P=.055 and P=.069, respectively). When all stroke events among nondiabetic men were considered together, the OR remained significantly elevated (2.05, 95% CI=1.00 to 4.19), even after adjustment for all CVD risk factors.

Age-adjusted ORs were also calculated for all subjects stratified by total cholesterol level at two different cutpoints (Table 6). ORs for CHD were similar in magnitude in men with low and high total cholesterol based on the 240 mg/dL and 200 mg/dL cutpoints. Associations were somewhat more variable for angina, though not strikingly different by cholesterol level. Associations involving PVD were significant for both low- and high-cholesterol groups at the 240 mg/dL cutpoint and were somewhat stronger among subjects with elevated cholesterol levels (ORs=1.4, 95% CI=1.1 to 1.7, and 2.6, 95% CI=1.2 to 5.3, respectively) but were similar in subjects stratified by the 200 mg/dL total cholesterol level. ORs for thromboembolic stroke appeared stronger in men with elevated total cholesterol relative to men with lower levels for both cutpoints (7.5 versus 1.7 for 240 mg/dL and 2.8 versus 1.5 for the 200 mg/dL cutpoint); however, ORs in the 240 mg/dL group were based on only three events among 244 subjects. ORs for hemorrhagic stroke could not be estimated or were less stable in the elevated cholesterol groups, given the relatively infrequent occurrence of these outcomes (n=0 and n=4, respectively) and the low prevalence of cholesterol levels 240 mg/dL in this cohort. Despite the apparent stronger associations of hyperinsulinemia with PVD in subjects with cholesterol levels 240 mg/dL and with stroke in both elevated-cholesterol groups, there was no evidence of statistical interaction when interaction terms were included in logistic models.

	Discussion

Top Abstract Introduction Methods Results Discussion References

 
The prevalence of CHD, angina, PVD, thromboembolic stroke, and hemorrhagic stroke was significantly higher in elderly Japanese-American men who had fasting hyperinsulinemia (defined as 20 µU/mL) than in those who did not. In all subjects, associations of hyperinsulinemia with CHD, angina, and, to a lesser extent, PVD were accounted for by other CVD risk factors, whereas those involving both types of stroke were independent of other CVD risk factors. When multivariate analyses were restricted to nondiabetic subjects, ORs tended to be slightly smaller and statistically significant less frequently. In general, the likelihood of having most of these CVD manifestations was not significantly elevated in nondiabetic men with hyperinsulinemia compared with men without hyperinsulinemia. However, the likelihood of having thromboembolic stroke remained twofold greater in hyperinsulinemic men compared with nonhyperinsulinemic men after adjustment for CVD risk factors, although the OR was no longer significantly elevated (1.99, 95% CI=0.95 to 4.17).

In a recent review of studies assessing the association between insulin and heart disease, Wingard et al³⁴ concluded that insulin does not appear to be a major risk factor for heart disease, based primarily on evidence from prospective epidemiological studies. It was also recently reported that considerable controversy still exists over whether endogenous hyperinsulinemia is an independent risk factor for CVD.³⁵ Other investigators have concluded that hyperinsulinemia and insulin resistance are related to atherosclerotic CVD but that their roles as independent CVD risk factors is less certain and that they are not major risk factors for CVD in the absence of other risk factors.⁸

Results from our study are consistent with several other recent population-based studies that have reported associations between insulin level or hyperinsulinemia and CHD.^{7 36 37 38 39} In most prospective studies of nondiabetic subjects, including the present one, associations of insulin level with heart disease have not been independent in multivariate analyses.³⁴ In some of these studies, no evidence of an association between insulin and heart disease was present.^{5 6 9} Although a recent investigation found these associations to be independent of other CVD risk factors, including HDL cholesterol,⁷ the findings from most studies are consistent with a more indirect role for hyperinsulinemia, one in which elevated insulin levels may lead to adverse changes in CVD risk factors³⁷ or may reflect an "innocent bystander" status, wherein fasting insulin levels serve as an indicator of an underlying insulin-resistant state.^{8 36}

Recent results from the Quebec Cardiovascular Study have suggested that the relation of hyperinsulinemia to ischemic heart disease among nondiabetic men may be independent of body weight, blood pressure, and lipoprotein levels.⁷ The authors of this investigation suggested that hyperinsulinemia may therefore serve as a marker for other metabolic and hemostatic disturbances. Such a role for hyperinsulinemia is also consistent with earlier evidence suggesting that both CVD and non-insulin-dependent diabetes mellitus may share common etiologic factors that may have been present for a number of years prior to clinical onset of disease.^{40 41 42}

Although not ideal, fasting insulin level appears to be a reasonable indicator of insulin resistance, particularly in subjects with normal glucose tolerance.⁴³ It has been suggested that insulin resistance or the compensatory hyperinsulinemia that follows⁴⁴ may be the central underlying defect of "syndrome X," or the "insulin resistance syndrome," which is characterized by glucose intolerance, dyslipidemia, and hypertension.^{45 46} It has also been suggested that hyperinsulinemia is intimately related to the CVD risk factors that are known to cluster within individuals and that the components of this syndrome are involved, to a substantial degree, in several major chronic disease states (CHD, non-insulin-dependent diabetes mellitus, obesity, and essential hypertension).^{47 48 49} Because of the close associations between insulin and constituents of the insulin resistance syndrome, adjustment for glucose intolerance, dyslipidemia, and hypertension would be expected to attenuate relations with CVD outcomes. Any remaining excess risk associated with insulin could reflect mechanisms other than atherosclerosis.

Relatively few studies have examined associations of insulin with PVD and stroke. Previously, subjects with lower-extremity arterial disease were shown to have elevated insulin levels following an oral glucose challenge.^{15 16} Results from the current study were consistent with the findings from the Cardiovascular Health Study, another population-based cohort study of the elderly. In the Cardiovascular Health Study, fasting insulin level was inversely associated with the ABI index, after adjustment for age and sex, but did not exhibit an independent association in multivariate analysis wherein multiple CVD risk factors, including diabetes and HDL cholesterol, were included.¹⁴ Also consistent with findings from this study, insulin resistance with compensatory hyperinsulinemia was significantly related to atherothrombotic stroke in a small case-control study.¹³ In a recent investigation of the current cohort, evidence suggested that unidentified factors other than hypertension might play an increasingly important role in the etiology of stroke with advancing age.⁵⁰ It is possible that insulin resistance or hyperinsulinemia might be one of these factors.

Indicators of asymptomatic atherosclerosis have recently been used to investigate potential associations with insulin levels. Fasting^{51 52} and postload⁵³ insulin levels as well as insulin resistance⁵⁴ have been associated with intimal wall thickness of the carotid artery. In addition, femoral atherosclerosis, defined as the presence of plaques by ultrasonography, was associated with insulin resistance in one study⁵⁵ but not with fasting insulin level or insulin sensitivity index in another.⁵⁶ Further studies that assess associations of insulin levels and insulin resistance with subclinical indicators of atherosclerosis are needed.

Several populations, including the Pima Indians,¹⁰ Hispanics,¹¹ and perhaps Japanese,¹² have paradoxically low rates of CHD despite relatively high proportions of individuals with glucose intolerance. It has been suggested that the relatively low total and LDL cholesterol levels in these populations may protect them from the adverse effects of glucose intolerance, hyperinsulinemia, or insulin resistance.⁵⁷ Results from the present study do not support this hypothesis with regard to CHD but instead are suggestive of an enhanced risk of PVD and thromboembolic stroke in Japanese-American men with elevated cholesterol levels, although evidence of statistical interaction was not significant. In addition, other investigators have suggested that hyperinsulinemia may potentiate the risk of CVD in subgroups who have an adverse risk factor profile characterized by hypertriglyceridemia and abnormal glucose tolerance^{58 59 60} or obesity.⁶¹

Atherogenic and thrombogenic roles for insulin are biologically plausible because insulin may (1) promote smooth muscle cell proliferation in arterial walls,^{48 62 63 64 65} (2) increase intima-media arterial wall thickness,^{51 52} (3) increase arterial wall stiffness independent of its effect on wall thickness,⁶⁶ (4) inhibit fibrinolysis,^{62 67} or (5) act indirectly through adverse effects on lipids^{45 48 68 69} and possibly blood pressure.^{45 48 68 70} It is possible that hyperinsulinemia may have an adverse influence through mechanisms other than atherosclerosis of large vessels. Evidence indicates that hyperinsulinemia may impair fibrinolysis by stimulating plasminogen activator inhibitor-1 synthesis.⁷¹ Consistent with this hypothesis, insulin resistance has been associated with low fibrinolytic activity.⁶⁷ Another potential mechanism includes the association of hyperinsulinemia with cerebral small-vessel disease.⁷² Such potential influences of elevated insulin levels through nonatherosclerotic mechanisms are consistent with population-based autopsy evidence of associations between diabetes and small-vessel disease in the heart.⁷³

Potential limitations of this study include its cross-sectional design, possible bias from selective survival, and a limited number of subjects who developed stroke (particularly hemorrhagic stroke) and survived to participate in the current examination. Causal inferences cannot be made from cross-sectional studies, since the temporal sequence between the presumed exposure (hyperinsulinemia) and disease (CVD) has not been elucidated. In addition, if subjects have had CVD for a number of years, it is possible that associations between risk factors and CVD may be altered by treatment of CVD. Use of medication for hypertension and diabetes was taken into account in multivariate logistic models. Exclusion of diabetic subjects may also have been useful in this regard. It is possible that this limitation was diminished in the case of PVD, since subjects identified with an abnormally low ABI could have been asymptomatic. Subjects who had CVD, survived, and participated in the examination may have had a less-severe CVD event than similar subjects who died, and if hyperinsulinemic subjects were also less likely to survive, measures of association would most likely be underestimated. Such a survival bias could account for a lack of association reported among some elderly populations. Thus, in this study of elderly men who may have a relatively advanced stage of atherosclerosis, hyperinsulinemia was significantly associated with CVD outcomes, and these associations may actually be stronger than those observed if a survival bias were present. Strengths of this study include its large sample size, its population-based design, the availability of a comprehensive surveillance system that used standardized criteria for more than three decades, and the ability to investigate the influence of multiple CVD risk factors in associations of hyperinsulinemia with these CVD manifestations.

In this population-based study of elderly Japanese-American men, hyperinsulinemia was associated with several manifestations of CVD. After adjustment for multiple CVD risk factors, hyperinsulinemia remained independently related to thromboembolic and hemorrhagic stroke. When analyses were restricted to nondiabetic subjects, hyperinsulinemia was not in general significantly associated with CVD, although individuals who had a stroke were at least twice as likely to have hyperinsulinemia as those who were stroke-free, even after adjustment for other CVD risk factors. For PVD and stroke, but not CHD, there was some evidence of enhanced risk associated with hyperinsulinemia among elderly men with elevated cholesterol levels relative to those with normal cholesterol levels. Prospective studies are clearly needed to discern whether hyperinsulinemia acts directly or indirectly to increase the risk of CVD. The relatively stronger associations of hyperinsulinemia with thromboembolic and hemorrhagic stroke may also warrant further investigation. Results from this study are in general consistent with an indirect role for insulin in CVD, wherein insulin may adversely affect other CVD risk factors or may serve as a marker for an enhanced atherogenic or thrombogenic state.

	Selected Abbreviations and Acronyms

 

ABI = ankle/brachial index BMI = body mass index CHD = coronary heart disease CI = confidence interval CVD = cardiovascular disease OR = odds ratio PVD = peripheral vascular disease

	Acknowledgments

 
This study was supported by contract NO1-HC-05102 to the Honolulu Heart Program, Kuakini Medical Center from the National Heart, Lung, and Blood Institute, Bethesda, Md.

	Footnotes

 
Reprint requests to Cecil M. Burchfiel, PhD, Division of Epidemiology and Clinical Applications, National Heart, Lung, and Blood Institute, National Institutes of Health, II Rockledge Centre, 6701 Rockledge Dr MSC 7934, Bethesda, MD 20892-7934.

Received April 7, 1997; accepted November 19, 1997.

	References

Top Abstract Introduction Methods Results Discussion References

 

1. Welborn TA, Wearne K. Coronary heart disease incidence and cardiovascular mortality in Busselton with reference to glucose and insulin concentrations. Diabetes Care. 1979;2:154–160.[Abstract]
2. Pyorala K. Relationship of glucose tolerance and plasma insulin to the incidence of coronary heart disease: results from two population studies in Finland. Diabetes Care. 1979;2:131–141.[Abstract]
3. Ducimetiere P, Eschwege E, Papoz L, Richard JL, Claude JR, Rosselin G. Relationship of plasma insulin levels to the incidence of myocardial infarction and coronary heart disease mortality in a middle-aged population. Diabetologia. 1980;19:205–210.[Medline] [Order article via Infotrieve]
4. Hargreaves AD, Logan RL, Elton RA, Buchanan KD, Oliver MF, Riemersmaa RA. Glucose tolerance, plasma insulin, HDL cholesterol and obesity: 12-year follow-up and development of coronary heart disease in Edinburgh men. Atherosclerosis. 1992;94:61–69.[Medline] [Order article via Infotrieve]
5. Welin L, Eriksson H, Larsson B, Ohlson LO, Svardsudd K, Tibblin G. Hyperinsulinaemia is not a major coronary risk factor in elderly men: the study of men born in 1913. Diabetologia. 1992;35:766–770.[Medline] [Order article via Infotrieve]
6. Ferrara A, Barrett-Connor EL, Edelstein SL. Hyperinsulinemia does not increase the risk of fatal cardiovascular disease in elderly men or women without diabetes: the Rancho Bernardo Study, 1984–1991. Am J Epidemiol. 1994;140:857–869.[Abstract/Free Full Text]
7. Despres J-P, Lamarche B, Mauriege P, Cantin B, Dagenais GR, Moorjani S, Lupien P-J. Hyperinsulinemia as an independent risk factor for ischemic heart disease. N Engl J Med. 1996;334:952–957.[Abstract/Free Full Text]
8. Savage PJ, Saad MF. Insulin and atherosclerosis: villain, accomplice or innocent bystander? Br Heart J. 1993;69:473–475.[Free Full Text]
9. Orchard TJ, Eichner J, Kuller LH, Becker DJ, McCallum LM, Grandits GA. Insulin as a predictor of coronary heart disease: interaction with apolipoprotein E phenotype: a report from the Multiple Risk Factor Intervention Trial. Ann Epidemiol. 1994;4:40–45.[Medline] [Order article via Infotrieve]
10. 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 non-insulin-dependent diabetes. Circulation. 1990;81:987–995.[Abstract/Free Full Text]
11. Stern MP, Bradshaw BS, Eifler CW, Fong DS, Hazuda HP, Rosenthal M. Secular decline in death rates due to ischemic heart disease in Mexican Americans and non-Hispanic whites in Texas, 1970–1980. Circulation. 1987;76:1245–1250.[Abstract/Free Full Text]
12. American Diabetes Association. The pharmacological treatment of hyperglycemia in NIDDM (consensus statement). Diabetes Care. 1996;19:S54–S61.
13. Shinozaki K, Naritomi H, Shimizu T, Suzuki M, Ikebuchi M, Sawada T, Harano Y. Role of insulin resistance associated with compensatory hyperinsulinemia in ischemic stroke. Stroke. 1996;27:37–43.[Abstract/Free Full Text]
14. Newman AB, Siscovick DS, Manolio TA, Polak J, Fried LP, Borhani NO, Wolfson SK. Ankle-arm index as a marker of atherosclerosis in the Cardiovascular Health Study. Circulation. 1993;88:837–845.[Abstract/Free Full Text]
15. Sloan JM, Mackay JS, Sheridan B. Glucose tolerance and insulin response in atherosclerosis. BMJ. 1970;4:586–588.
16. Sorge F, Schwartzkopff W, Neuhaus GA. Insulin response to oral glucose in patients with peripheral vascular disease: hyperinsulinism and its relationships to hypertriglyceridemia and overweight. Diabetes. 1976;25:586–594.[Abstract]
17. Worth RM, Kagan A. Ascertainment of men of Japanese ancestry in Hawaii through World War II Selective Service registration. J Chron Dis. 1970;23:389–397.[Medline] [Order article via Infotrieve]
18. Rhoads GG, Kagan AM, Yano K. Usefulness of community surveillance for the ascertainment of coronary heart disease and stroke. Int J Epidemiol. 1975;4:265–270.[Abstract/Free Full Text]
19. Yano K, Reed DM, McGee DL. Ten-year incidence of coronary heart disease in the Honolulu Heart Program: relationship to biological and life-style characteristics. Am J Epidemiol. 1984;119:653–666.[Abstract/Free Full Text]
20. Burchfiel CM, Sharp DS, Curb JD, Rodriguez BL, Hwang L-J, Marcus EB, Yano K. Physical activity and incidence of diabetes: the Honolulu Heart Program. Am J Epidemiol. 1995;141:360–368.[Abstract/Free Full Text]
21. Ettinger WH, Wahl PW, Kuller LH, Bush TL, Tracy RP, Manolio TA, Borhani NO, Wong ND, O'Leary DH. Lipoprotein lipids in older people: results from the Cardiovascular Health Study. Circulation. 1992;86:858–869.[Abstract/Free Full Text]
22. Fried LP, Borhani NO, Enright P, Furberg CD, Gardin JM, Kronmal RA, Kuller LH, Manolio TA, Mittelmark MB, Newman A, O'Leary DH, Psaty B, Rautaharju P, Tracy RP, Weiler PG. The Cardiovascular Health Study: design and rationale. Ann Epidemiol. 1991;1:263–276.[Medline] [Order article via Infotrieve]
23. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18:499–502.[Abstract]
24. Morgan CR, Lazarov A. Immunoassay of insulin: two antibody systems. Diabetes. 1963;21:115–126.
25. Burchfiel CM, Curb JD, Sharp DS, Rodriguez BL, Arakaki R, Chyou P-H, Yano K. Distribution and correlates of insulin in elderly men: the Honolulu Heart Program. Arterioscler Thromb Vasc Biol. 1995;15:2213–2221.[Abstract/Free Full Text]
26. World Health Organization. Diabetes Mellitus: Report of a WHO Study Group. Geneva, Switzerland: World Health Organization; 1985. WHO technical report series No. 727.
27. Kagan A, Popper JS, Rhoads GG. Factors related to stroke incidence in Hawaii Japanese men: the Honolulu Heart Study. Stroke. 1980;11:14–21.[Abstract/Free Full Text]
28. Cook DG, Shaper AG, Macfarlane PW. Using the WHO (Rose) angina questionnaire in cardiovascular epidemiology. Int J Epidemiol. 1989;18:607–613.[Abstract/Free Full Text]
29. Burchfiel CM, Curb JD, Rodriguez BL, Abbott RD, Chiu D, Yano K. Glucose intolerance and 22-year stroke incidence. Stroke. 1994;25:951–957.[Abstract]
30. Curb JD, Masaki K, Rodriguez BL, Abbott RD, Burchfiel CM, Chen R, Petrovitch H, Sharp D, Yano K. Peripheral artery disease and cardiovascular risk factors in the elderly: the Honolulu Heart Program. Arterioscler Thromb Vasc Biol. 1996;16:1495–1500.[Abstract/Free Full Text]
31. SAS Institute Inc. SAS/STAT User's Guide, Version 6. Cary, NC: SAS Institute Inc; 1990.
32. Lane PW, Nelder JA. Analysis of covariance and standardization as instances of prediction. Biometrics. 1982;38:613–621.[Medline] [Order article via Infotrieve]
33. Wilcosky TC, Chambless LE. A comparison of direct adjustment and regression adjustment on epidemiologic measures. J Chron Dis. 1985;38: 849–856.
34. Wingard DL, Ferrara A, Barrett-Connor EL. Is insulin really a heart disease risk factor? Diabetes Care. 1995;18:1299–1304.[Medline] [Order article via Infotrieve]
35. Stern MP. The insulin resistance syndrome: the controversy is dead, long live the controversy! Diabetologia. 1994;37:956–958.[Medline] [Order article via Infotrieve]
36. Perry IJ, Wannamethee SG, Whincup PH, Shaper AG, Walker MK, Alberti KGMM. Serum insulin and incident coronary heart disease in middle-aged British men. Am J Epidemiol. 1996;144:224–234.[Abstract/Free Full Text]
37. Mykkanen L, Laakso M, Pyorala K. High plasma insulin level associated with coronary heart disease in the elderly. Am J Epidemiol. 1993;137:1190–1202.[Abstract/Free Full Text]
38. Kahn SE, Leonetti DL, Prigeon RL, Boyko EJ, Bergstrom RW, Fujimoto WY. Relationship of proinsulin and insulin with noninsulin-dependent diabetes mellitus and coronary heart disease in Japanese-American men: impact of obesity: clinical research center study. J Clin Endocrinol Metab. 1995;80:1399–1406.[Abstract]
39. Feskens EJM, Kromhout D. Hyperinsulinemia, risk factors, and coronary heart disease: the Zutphen Elderly Study. Arterioscler Thromb. 1994;14:1641–1647.[Abstract/Free Full Text]
40. Jarrett RJ. Epidemiology and public health aspects of non-insulin-dependent diabetes mellitus. Epidemiol Rev. 1989;11:151–171.[Free Full Text]
41. Haffner SM, Stern MP, Hazuda HP, Mitchell BD, Patterson JK. Cardiovascular risk factors in confirmed prediabetic individuals: does the clock for coronary heart disease start ticking before the onset of clinical diabetes? JAMA. 1990;263:2893–2998.[Abstract]
42. Stern MP. Diabetes and cardiovascular disease: the "common soil" hypothesis. Diabetes. 1995;44:369–374.[Abstract]
43. Laakso M. How good a marker is insulin level for insulin resistance? Am J Epidemiol. 1993;137:959–965.[Abstract/Free Full Text]
44. Reaven GM, Laws A. Insulin resistance, compensatory hyperinsulinaemia, and coronary heart disease. Diabetologia. 1994;37:948–952.[Medline] [Order article via Infotrieve]
45. Reaven GM. Banting lecture 1988: role of insulin resistance in human disease. Diabetes. 1988;37:1595–1607.[Abstract]
46. Haffner SM, Valdez RA, Hazuda HP, Mitchell BD, Morales PA, Stern MP. Prospective analysis of the insulin-resistance syndrome (syndrome X). Diabetes. 1992;41:715–722.[Abstract]
47. Reaven GM. Role of insulin resistance in human disease (syndrome X): an expanded definition. Annu Rev Med. 1993;44:121–131.[Medline] [Order article via Infotrieve]
48. Simonson DC. Hyperinsulinemia and its sequelae. Horm Metab Res Suppl. 1990;22:17–25.[Medline] [Order article via Infotrieve]
49. Liese AD, Mayer-Davis EJ, Tyroler HA, Davis CE, Keil U, Duncan BB, Heiss G. Development of the multiple metabolic syndrome in the ARIC cohort: joint contribution of insulin, BMI and WHR. Ann Epidemiol. 1997;7:407–416.[Medline] [Order article via Infotrieve]
50. Curb JD, Abbott RD, MacLean CJ, Rodriguez BL, Burchfiel CM, Sharp DS, Ross GW, Yano K. Age-related changes in stroke risk in men with hypertension and normal blood pressure. Stroke. 1996;27:819–824.[Abstract/Free Full Text]
51. Folsom AR, Eckfeldt JH, Weitzman S, Ma J, Barnes RW, Cram KB, Hutchinson RG. Relation of carotid artery wall thickness to diabetes mellitus, fasting glucose and insulin, body size and physical activity. Stroke. 1994;25:66–73.[Abstract]
52. Howard G, O'Leary DH, Zaccaro D, Haffner S, Rewers M, Hamman R, Selby JV, Saad MF, Savage P, Bergman R. Insulin sensitivity and atherosclerosis. Circulation. 1996;93:1809–1817.[Abstract/Free Full Text]
53. Niskanen L, Rauramaa R, Miettinen H, Haffner SM, Mercuri M, Uusitupa M. Carotid artery intima-media thickness in elderly patients with NIDDM and in nondiabetic subjects. Stroke. 1996;27:1986–1992.[Abstract/Free Full Text]
54. Suzuki M, Shinozaki K, Kanazawa A, Hara Y, Hattori Y, Tsushima M, Harano Y. Insulin resistance as an independent risk factor for carotid wall thickening. Hypertension. 1996;28:593–598.[Abstract/Free Full Text]
55. Laakso M, Sarlund H, Salonen R, Suhonen M, Pyorala K, Salonen JT, Karhapaa P. Asymptomatic atherosclerosis and insulin resistance. Arterioscler Thromb. 1991;11:1068–1076.[Abstract/Free Full Text]
56. Kekalainen P, Sarlund H, Farin P, Kaukanen E, Yang X, Laakso M. Femoral atherosclerosis in middle-aged subjects: association with cardiovascular risk factors and insulin resistance. Am J Epidemiol. 1996;144:742–748.[Abstract/Free Full Text]
57. McKeigue PM, Miller GJ, Marmot MG. Coronary heart disease in South Asians overseas: a review. J Clin Epidemiol. 1989;42:597–609.[Medline] [Order article via Infotrieve]
58. Fontbonne A, Eschwege E, Cambien F, Richard JL, Ducimetiere P, Thibult N, Warnet JM, Claude JR, Rosselin GE. Hypertriglyceridaemia as a risk factor of coronary heart disease mortality in subjects with impaired glucose tolerance or diabetes: results from the 11-year follow-up of the Paris Prospective Study. Diabetologia. 1989;32:300–304.[Medline] [Order article via Infotrieve]
59. Fontbonne A, Charles MA, Thibult N, Richard JL, Claude JR, Warnet JM, Rosselin GE, Eschwege E. Hyperinsulinaemia as a predictor of coronary heart disease mortality in a healthy population: the Paris Prospective Study, 15-year follow-up. Diabetologia. 1991;34:356–361.[Medline] [Order article via Infotrieve]
60. Yarnell JWG, Sweetnam PM, Marks V, Teale JD, Bolton CH. Insulin in ischaemic heart disease: are associations explained by triglyceride concentrations? The Caerphilly prospective study. Br Heart J. 1994;71:293–296.[Abstract/Free Full Text]
61. Fontbonne A, Tchobroutsky G, Eschwege E, Richards JL, Claude JR, Rosselin GE. Coronary heart disease mortality risk: plasma insulin level is a more sensitive marker than hypertension or abnormal glucose tolerance in overweight males: the Paris Prospective Study. Int J Obes. 1988;12:557–565.[Medline] [Order article via Infotrieve]
62. Stolar MW. Atherosclerosis in diabetes: the role of hyperinsulinemia. Metabolism. 1988;37(suppl 1):1–9.
63. Stout RW. Insulin and atheroma: 20-yr perspective. Diabetes Care. 1990;13:631–654.[Abstract]
64. Stout RW, Bierman FL, Ross R. Effect of insulin on the proliferation of cultured primate arterial smooth muscle cells. Circ Res. 1975;36:319–327.[Abstract/Free Full Text]
65. Pfeifle B, Dischuneit H. Effect of insulin on growth of cultured human arterial smooth muscle cells. Diabetologia. 1981;20:155–158.[Medline] [Order article via Infotrieve]
66. Salomaa V, Riley W, Kark JD, Nardo C, Folsom AR. Non-insulin-dependent diabetes mellitus and fasting glucose and insulin concentrations are associated with arterial stiffness indexes: the ARIC Study. Circulation. 1995;91:1432–1443.[Abstract/Free Full Text]
67. Lindahl B, Asplund K, Eliasson M, Evrin P-E. Insulin resistance syndrome and fibrinolytic activity: the Northern Sweden MONICA Study. Int J Epidemiol. 1996;25:291–299.[Abstract/Free Full Text]
68. Zavaroni I, Bonora E, Pagliara M, Dall'Agio E, Luchetti L, Buonanno G, Bonati PA, Bergonzani M, Gnudi L, Passeri M, Reaven G. Risk factors for coronary artery disease in healthy persons with hyperinsulinemia and normal glucose tolerance. N Engl J Med. 1989;320:702–706.[Abstract]
69. Jeppesen J, Hollenbeck CB, Zhou M-Y, Coulston AM, Jones C, Chen Y-DI, Reaven GM. Relation between insulin resistance, hyperinsulinemia, postheparin plasma lipoprotein lipase activity, and postprandial lipemia. Arterioscler Thromb Vasc Biol. 1995;15:320–324.[Abstract/Free Full Text]
70. Daly PA, Landsberg L. Hypertension in obesity and NIDDM: role of insulin and sympathetic nervous system. Diabetes Care. 1991;14:240–248.[Abstract]
71. Vague P, Juhan-Vague I, Aillaud MF, Badier C, Viard R, Alessi MC, Colen D. Correlation between blood fibrinolytic activity, plasminogen activator inhibitor level, plasma insulin level, and relative body weight in normal and obese subjects. Metabolism. 1986;35:250–253.[Medline] [Order article via Infotrieve]
72. Zunker P, Schick, A, Buschmann H-C, Georgiadis D, Nabavi DG, Edelmann M, Ringelstein EB. Hyperinsulinism and cerebral microangiopathy. Stroke. 1996;27:219–223.[Abstract/Free Full Text]
73. Burchfiel CM, Reed DM, Marcus EB, Strong JP, Hayashi T. Association of diabetes mellitus with coronary atherosclerosis and myocardial lesions: an autopsy study from the Honolulu Heart Program. Am J Epidemiol. 1993;137:1328–1340.[Abstract/Free Full Text]