Cross-Sectional and Prospective Associations between Abdominal ...

0013-7227/02/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 87(1):77– 83 Copyright © 2002 by The Endocrine Society

Cross-Sectional and Prospective Associations between Abdominal Adiposity and Proinsulin Concentration ANTHONY J. G. HANLEY, GAIL MCKEOWN-EYSSEN, STEWART B. HARRIS, ROBERT A. HEGELE, THOMAS M. S. WOLEVER, JEREMY KWAN, AND BERNARD ZINMAN Department of Public Health Sciences (A.J.G.H., G.M.-E.), University of Toronto, Ontario M5S 1A8; Samuel Lunenfeld Research Institute (A.J.G.H., B.Z.), Mt. Sinai Hospital, Toronto, Ontario M5S 1X5; Centre for Studies in Family Medicine (S.B.H.), University of Western Ontario, London, Ontario N6G 4X8; Robarts Research Institute (R.A.H.), London, Ontario N6A 5K8; Department of Nutritional Sciences (T.M.S.W.), University of Toronto, Ontario M5S 1A8; Banting and Best Diabetes Centre (J.K.), University of Toronto, Ontario M5G 2C4; and Division of Endocrinology and Metabolism (B.Z.), Mt. Sinai Hospital and the University Health Network, Toronto, Ontario M5G 1X5 The objective of this study was to investigate the associations of total and abdominal obesity with variation in proinsulin concentration in a Native Canadian population experiencing an epidemic of type 2 diabetes mellitus (DM). Between 1993 and 1995, 728 members of a Native Canadian community participated in a population-based survey to determine the prevalence and risk factors for type 2 DM. Samples for glucose, C-peptide, and proinsulin were drawn after an overnight fast, and a 75-g oral glucose tolerance test was administered. Type 2 DM and impaired glucose tolerance (IGT) were diagnosed using World Health Organization criteria. Height, weight, waist circumference, and percent body fat were measured. In 1998, 95 individuals who, at baseline, had IGT or normal glucose tolerance with an elevated 2-h glucose level (>7.0 mM) participated in a follow-up evaluation using the same protocol. After adjustment for age, sex, C-peptide concentration, per cent body fat, and waist circumference, proinsulin was found to be significantly elevated in diabetic subjects, relative to

subjects with both impaired and normal glucose tolerance (both P < 0.0001); and the concentration in those with IGT was higher, compared with normals (P < 0.0001). Among nondiabetic subjects, proinsulin showed significant univariate associations with percent body fat, body mass index, and waist circumference (r ⴝ 0.34, 0.45, 0.41, respectively, all P < 0.0001). After adjustment for body fat and other covariates, waist circumference remained significantly associated with proinsulin concentration in nondiabetic subjects (r ⴝ 0.20, P < 0.0001). In prospective analysis, adjusted for covariates (including baseline IGT and follow-up glucose tolerance status), baseline waist circumference was positively associated with both follow-up and change in proinsulin concentration (r ⴝ 0.27, P ⴝ 0.01; r ⴝ 0.24, P ⴝ 0.03, respectively). These data highlight the detrimental effects of abdominal obesity on ␤-cell function, and support the hypothesis that ␤-cell dysfunction occurs early in the natural history of glucose intolerance. (J Clin Endocrinol Metab 87: 77– 83, 2002)

B

ETA-CELL DYSFUNCTION is a central feature in the natural history of type 2 diabetes mellitus (DM) (1, 2). An understanding of the etiology of ␤-cell dysfunction, however, is complicated by the fact that gold standard techniques for its measurement, including the hyperglycemic clamp procedure (1), are excessively costly, time-consuming, and invasive for use in large, population-based studies (1). Fasting proinsulin concentration is emerging as a possible surrogate indicator of ␤-cell dysfunction (3, 4). Proinsulin and its split products circulate in high concentrations in subjects with diabetes and gestational diabetes (5–14). In addition, a number of investigations have reported that elevated absolute and relative concentrations of proinsulin and its split products are prospectively associated with risk of diabetes (15–20). Finally, recent studies have shown that proinsulin and the proinsulin-to-insulin ratio (PI/I) correlate significantly with acute insulin response from the frequently sampled iv glucose tolerance test (21, 22). This body of literature

provides support for the hypothesis that proinsulin can be used as a sensitive marker of early ␤-cell dysfunction (23). The possible role of obesity as an independent determinant of variation in proinsulin concentration has received only limited attention. In nondiabetic subjects, body mass index (BMI) and waist to hip ratio (WHR) have been reported to be positively correlated with proinsulin and, less consistently, inversely related to PI/I in a limited number of studies (9, 11, 21, 24 –27). These papers presented the results of univariate analyses only, however, and the effect of confounding by other variables (including glucose concentrations and insulin secretion) on these relationships is unknown. To our knowledge, the relationship between proinsulin and percent body fat has not been explored, nor has the association with waist circumference alone, which has been shown to be superior to WHR as a surrogate measure of intraabdominal fat (IAF) (28, 29). Abdominal obesity is an important determinant of elevated circulating concentrations of FFA (30 –32), and evidence is accumulating for a detrimental effect of chronically elevated FFA on ␤-cell function (33, 34). A detailed examination of the associations of proinsulin concentration with measures of adiposity would be of value, in light of this literature. We have previously documented high rates of

Abbreviations: BIA, Bioelectrical impedance analysis; BMI, body mass index; CI, confidence interval; DM, diabetes mellitus; GTS, glucose tolerance status; IAF, intraabdominal fat; IGT, impaired glucose tolerance; NGT, normal glucose tolerance; OGTT, oral glucose tolerance test; PI/I, proinsulin-to-insulin ratio; WHO, World Health Organization; WHR, waist to hip ratio.

77

78

J Clin Endocrinol Metab, January 2002, 87(1):77– 83

obesity and glucose intolerance in a population-based study of Native Canadians in northern Ontario (35–37). In the present paper, we employ cross-sectional and prospective data from this study to determine whether measures of total and intraabdominal adiposity are independently related to proinsulin concentration. Subjects and Methods The community of Sandy Lake, Ontario, is located roughly 2000 km northwest of Toronto, in the Boreal Forest region of central Canada. Approximately 1600 people live in this isolated village, which is accessible only by air for most of the year. Historically, the inhabitants of this area lived in small nomadic groups and led a hunting-and-gathering subsistence typical of other North American subarctic populations. Their lives were physically active, and their diet was high in protein from wild meat and fish, with seasonal supplementation from berries and roots. The lifestyle of the people of this region has changed dramatically over the past several decades, with a marked decrease in physical activity and an alteration in diet to one characterized by excess consumption of saturated fat and processed foods (38). This population is consequently undergoing an epidemiologic transition, with a rapid increase in morbidity related to chronic diseases, such as obesity and type 2 DM (39, 40).

Baseline prevalence survey The methodology of the Sandy Lake Health and Diabetes Project prevalence study has been presented in detail in previous publications (35, 36). Briefly, between July 1993 and December 1995, 728 of 1018 (72%) eligible residents of Sandy Lake, 10 –79 yr old, participated in a population-based, cross-sectional survey to determine the prevalence of type 2 DM and its associated risk factors. Signed informed consent was obtained from all participants, and the study was approved by the Sandy Lake First Nation Band Council and University of Toronto Ethics Review Committee. The current study is based on data from 701 individuals for whom specimens were available for proinsulin determination. Participants provided fasting blood samples for glucose, insulin, and proinsulin after an 8 –12 h overnight fast. A 75-g oral glucose tolerance test (OGTT) was administered, and a second sample for glucose was drawn after 120 min. Individuals were excluded from the OGTT if they had physician-diagnosed diabetes and were: 1) currently receiving treatment with insulin or oral hypoglycemic agents; or 2) had a fasting blood glucose concentration exceeding 11.1 mm. Women who were pregnant at the time of initial contact received their OGTT 3 months postpartum. Diabetes and impaired glucose tolerance (IGT) were diagnosed according to World Health Organization (WHO) criteria (41). Insulin was measured using an RIA (Pharmacia, Inc., Uppsala, Sweden) that has a lower detection limit of 22 pm and an interassay coefficient of variation of 7.2– 8.8%. Whereas this assay displays low cross-reactivity with C- peptide (⬍0.18%), the cross-reactivity with proinsulin is very high (100%); and thus, reported values refer to concentrations of total immunoreactive insulin. Glucose concentration was determined using the glucose oxidase method. C-peptide level was measured using an RIA (Diagnostic Products, Los Angeles, CA) that has a minimal detection limit of 43 pm, and cross-reactivities of 0% with insulin and less than 13% with proinsulin. Proinsulin concentration was determined using a human proinsulin RIA [Linco Research, Inc., St. Charles, MO (42)], that has a laboratory sensitivity of 3.5 pm and a coefficient of variation of 6.2–21.0%. This assay displays 46% cross-reactivity with des 31,32 proinsulin, the major form of circulating split proinsulin (43); and thus, reported values refer to total proinsulin-like materials. Cross-reactivity with C-peptide, des 64,65 proinsulin, and insulin is low (⬍0.1%). Proinsulin was measured in serum specimens that had been stored at ⫺70 C for between 3–5 yr at the Core Lab of the Banting and Best Diabetes Center, University of Toronto. Anthropometric measurements were performed with the volunteer wearing either light athletic clothing or undergarments and a hospital gown and no shoes. Each measurement was performed twice, and the average was used in the analysis. Height was measured to the nearest 0.1 cm using a wall-mounted stadiometer. Weight was measured to the nearest 0.1 kg using a hospital balance beam scale. BMI was defined as weight/height2 (kg/m2). The waist was measured to the nearest 0.5 cm

Hanley et al. • Abdominal Adiposity and Proinsulin

at the point of narrowing (as viewed from behind) between the umbilicus and xiphoid process; the hips were measured to the nearest 0.5 cm at the maximum extension of the buttocks. WHR was calculated as the ratio of these two circumferences. Percent body fat was estimated by bioelectrical impedance analysis (BIA) using the Tanita TBF-201 Body Fat Analyzer (Tanita/Stellar Innovations, Inc., Tokyo, Japan). High reproducibility of percent fat estimates using this machine in a sample from this population has been documented (intraclass correlation coefficient ⫽ 0.99) (44), and the instrument has been validated by others against dual-energy x-ray absorptiometry in both diabetic and nondiabetic populations (45– 47).

Prospective study Subjects in the baseline survey who were found to be at high risk for subsequent diabetes, including those with IGT (n ⫽ 74) (36) or normal glucose tolerance (NGT) with a 2-h post challenge glucose level greater than or equal to 7.0 mm (n ⫽ 51), were invited to participate in a follow-up visit during the summer of 1998 to determine current glucose tolerance and risk factor status. Of the 125 individuals in this follow-up cohort, 3 (3 IGT, 0 NGT) had died, 11 (9 IGT, 2 NGT) were no longer living in the community, 2 (1 IGT, 1 NGT) were too sick to participate, and 14 (6 IGT, 8 NGT) refused to attend. Thus, 95 (76%) members of this high-risk cohort participated in the follow-up examination. Nonparticipants did not differ significantly from participants in age, gender, anthropometric, or metabolic variables (data not shown). Metabolic and anthropometric variables and glucose tolerance status (GTS) were determined using the same protocol employed during the 1993–1995 prevalence survey.

Statistical analyses The analytical convention in this body of literature is to adjust proinsulin level for insulin secretion, usually by employing PI/I. This is problematic, for two reasons. First, peripheral insulin levels do not adequately represent insulin secretion, given that insulin undergoes a large and variable hepatic extraction as well as peripheral clearance that varies under different physiological circumstances (48). Second, Kronmal (49) and Allison et al. (50) have pointed out that the use of ratio variables in correlation and regression can result in spurious findings. We have therefore opted to avoid the use of ratios and to conduct our analysis of proinsulin levels after adjustment for C-peptide, which is cosecreted with insulin in an equimolar ratio, is not extracted by the liver, and has a constant peripheral clearance (48). All analyses were carried out using SAS Institute, Inc. (Cary, NC) version 6.12 (51), and two-sided P values less than 0.05 were considered statistically significant. The distributions of continuous variables were assessed for normality using the Shapiro and Wilk statistic and plots provided by SAS Institute, Inc. PROC UNIVARIATE, and the natural log transformation of skewed variables was used in subsequent parametric multivariate analysis. Spearman correlation analysis was employed to examine the associations between anthropometric factors and fasting proinsulin concentration among nondiabetic subjects (NGT or IGT). Scatterplots were examined to scan for departures from linearity. In addition to BMI and WHR, we used percent body fat and waist circumference as measures of total and abdominal adiposity, respectively. As mentioned above, percent body fat measured by BIA has shown good criterion validity against dual-energy x-ray absorptiometry (45– 47), and waist circumference has been demonstrated to be superior to WHR as a measure of IAF (28, 29). Both unadjusted and adjusted (for sex, age, C-peptide, fasting glucose, and 2-h glucose) models are presented. Proinsulin concentrations across gender-specific quartiles of waist circumference were estimated using analysis of covariance, adjusting for age, C-peptide, fasting glucose, 2-h glucose, and percent body fat. In the analysis of the prospective data, the effect of both baseline levels and changes in per cent body fat and waist circumference on both follow-up and change in proinsulin concentrations were evaluated using Spearman correlation analyses. Coefficients were adjusted for age, sex, C-peptide, baseline, and follow-up diabetes status and the baseline values of change variables (per cent body fat, waist circumference, proinsulin), which were included to adjust for the correlation between the initial and follow-up levels. Scatterplots were examined to screen for the presence of outliers.


J Clin Endocrinol Metab, January 2002, 87(1):77– 83 79

Results

Table 1 presents characteristics of participants in the Sandy Lake Health and Diabetes Project, by GTS. There was an increase in the mean of age, proinsulin, fasting and 2-h glucose, waist circumference, and WHR, with worsening GTS category. BMI and immunoreactive insulin and C-peptide concentrations were higher among subjects with IGT and DM, compared with those among NGT. After adjustment for age, sex, C-peptide, percent body fat, and waist circumference, the proinsulin concentration was significantly elevated in diabetic subjects [least-square mean ⫽ 22.2 pm; 95% confidence interval (CI), 20.5–24.0 pm], relative to both subjects with IGT (14.9 pm; 95% CI, 13.5–16.4) and NGT (10.9 pm; 95% CI, 10.5–11.4) (both P ⬍ 0.0001); and the concentration in those with IGT was significantly higher, compared with subjects with NGT (P ⬍ 0.0001). Cross-sectional associations between adiposity and proinsulin concentration

variate models indicate significant associations of per cent body fat, BMI, waist circumference, and WHR with proinsulin concentration (Table 2, column A: r ⫽ 0.34, 0.45, 0.41, 0.23, respectively, all P ⬍ 0.0001). After adjustment for age, sex, C-peptide, and fasting and 2-h glucose concentrations, these associations were attenuated but remained highly significant (Table 2, column C: r ⫽ 0.17– 0.27, all P ⬍ 0.0001). To examine the association between abdominal adiposity and proinsulin concentration, independent of total body adiposity, we further adjusted the waist circumference-proinsulin and WHR-proinsulin correlation coefficients for per cent body fat (Table 2, column D). Whereas both indices of abdominal adiposity remained significantly associated with proinsulin, the magnitude of the association with waist circumference was notably larger. Fig. 1 presents least-square mean proinsulin concentrations by gender and quartiles of waist circumference in nondiabetic subjects. In both males and females, there is a clear trend of increasing proinsulin concentration with

Table 2 presents results of unadjusted and adjusted analyses examining associations between adiposity variables and proinsulin concentration among nondiabetic subjects. UniTABLE 1. Characteristics of participants in the Sandy Lake Health and Diabetes Project, by glucose tolerance status Glucose tolerance status

Variablea

NGT

IGT

b

n (Males/female) 232/273 Age (years) 25.5 (12.9) c Total IRI (pM) 109.3 (70.0) Proinsulin (pM) 12.0 (7.2) C-peptide (pM) 599.5 (372.5) Proinsulin/IRI 0.12 (0.05) Proinsulin/C-peptide 0.02 (0.01) Glucose (mM) 5.4 (0.5) 2-h glucose (mM) 5.2 (1.3) 25.5 (5.6) BMI (kg/m2) Percent body fat (%) 32.8 (12.9) Waist circumference 88.5 (13.8) (cm) WHR 0.90 (0.07)

DM

16/58 39.1 (17.7) 176.9 (116.9) 19.0 (12.4) 841.2 (442.7) 0.12 (0.01) 0.02 (0.01) 5.8 (0.7) 8.9 (0.8) 29.7 (5.2) 43.0 (9.8) 97.6 (10.8)

48/74 44.4 (15.2) 178.6 (169.0) 28.1 (15.7) 853.4 (517.3) 0.19 (0.07) 0.04 (0.03) 11.4 (5.0) 16.1 (5.7) 30.2 (4.8) 39.5 (11.2) 101.6 (11.3)

0.94 (0.06)

0.94 (0.07)

a Values for biochemical variables are fasting concentrations (mean ⫾ SD), unless otherwise indicated. b Sample sizes vary because of occasional missing values. c Total immunoreactive insulin.

FIG. 1. Proinsulin concentrations (pM), by gender and quartiles of waist circumference, among nondiabetic subjects. Waist quartiles for males: ⱕ78.9, 79 –90.8, 90.9 –101.6, ⬎101.6; for females: ⱕ76.0, 77.1– 87.5, 87.6 –97.0, ⬎97.0. Values are least-square means (with 95% CI) from analysis of covariance, adjusted for log age, log C-peptide, fasting glucose, 2-h glucose, and per cent body fat. Males (solid bars), ptrend ⬍ 0.0001; females (open bars), ptrend ⫽ 0.0041.

TABLE 2. Unadjusted and adjusted cross-sectional associations between fasting proinsulin concentration and measures of adiposity among nondiabetic subjects (n ⫽ 579)a Fasting proinsulin concentration

A Unadjusted

Adiposity variable

Percent body fat BMI Waist circumference WHR a b

B Adjusted for age, sex, C-peptide

C Adjusted for age, sex, C-peptide, fasting glucose, 2-h glucose

D Adjusted for age, sex, C-peptide, fasting glucose, 2-h glucose, % body fat

r

P-value

r

P-value

r

P-value

r

0.34 0.45 0.41 0.23

⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001

0.22 0.28 0.31 0.20

⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001

0.19 0.24 0.27 0.17

⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001

b

Spearman correlation analysis. Not estimable, given that this model is adjusted for total body adiposity (% body fat).

P-value

b

0.20 0.11

⬍0.0001 0.0074

80



increasing levels of waist circumference after adjustment for age, insulin secretion, fasting and 2-h glucose, and percent body fat (males, ptrend ⬍0.0001; females, ptrend ⫽ 0.0041). Prospective associations between adiposity and proinsulin concentration

The associations of baseline levels and change in adiposity with follow-up and change in proinsulin concentration are TABLE 3. Prospective associations of baseline levels and change in anthropometric measurements with follow-up and change in fasting proinsulin concentration (total n ⫽ 90)a Proinsulin concentration Independent variable

Percent body fat Baseline Change Waist circumference Baseline Change

Follow-up

Change

Partial r

P-value

Partial r

P-value

0.13b

0.24

0.21c 0.10d

0.06 0.39

0.27b

0.01

0.24c 0.20e

0.03 0.08

a Spearman correlation analyses; sample sizes vary slightly because of occasional missing values. b Analyses adjusted for age, sex, follow-up C-peptide concentration, baseline and follow-up diabetes status, and baseline proinsulin. c Analyses adjusted for age, sex, change in C-peptide concentration, baseline and follow-up diabetes status, and baseline proinsulin. d Analyses adjusted for age, sex, change in C-peptide concentration, baseline and follow-up diabetes status, baseline proinsulin, and baseline percent body fat. e Analyses adjusted for age, sex, change in C-peptide concentration, baseline and follow-up diabetes status, and baseline proinsulin, and baseline waist circumference.

FIG. 2. Association between baseline waist circumference and change in proinsulin concentration in the prospective study. Data are least-square mean values (with SE) for change in fasting proinsulin concentration, adjusted for age, change in C-peptide concentration, baseline IGT, follow-up GTS, and baseline proinsulin concentration using analysis of covariance (pooled models also adjusted for sex). Solid bars, Baseline waist circumference less than the gender-specific median (males ⬍99.55 cm, females ⬍96.00 cm); open bars, baseline waist circumference equal to or greater than the gender-specific median.

presented in Table 3. After adjustment for age, sex, change in C-peptide concentration, baseline proinsulin, and baseline IGT, and follow-up diabetes status, baseline waist circumference was a significant predictor of both follow-up and change in proinsulin concentration (partial r ⫽ 0.27 and 0.24, respectively, both P ⬍ 0.05). Subjects whose baseline waist circumference was equal to or above the gender-specific median had a significantly larger change in proinsulin concentration, compared with those whose baseline waist circumference was below the median (Fig. 2; both sexes, 10.5 vs. ⫺1.8 pm, P ⫽ 0.003 after adjustment for age, sex, change in C-peptide, baseline IGT and follow-up diabetes status, and baseline proinsulin concentration). The pattern was similar for both males and females, although the difference was statistically significant only among females. In addition, subjects who experienced a change in their waist circumference that was equal to or greater than the gender-specific median had a significantly larger change in proinsulin concentration, compared with those whose waist circumference change was below the median (Fig. 3, both sexes, 8.37 vs. 0.55 pm, P ⫽ 0.04 after adjustment for age, sex, change in C-peptide, baseline IGT and follow-up diabetes status, and baseline waist circumference and baseline proinsulin concentration). Discussion

In this paper, we have demonstrated that measures of adiposity were positively and independently associated with proinsulin concentrations, and waist circumference was significantly associated with proinsulin after further adjustment for percent body fat. Further, our prospective analyses dem-

FIG. 3. Association between change in waist circumference and change in proinsulin concentration in the prospective study. Data are least-square mean values (with SE) for change in fasting proinsulin concentration, adjusted for age, change in C-peptide concentration, baseline IGT, follow-up GTS, baseline waist circumference, and baseline proinsulin concentration using analysis of covariance (pooled models also adjusted for sex). Solid bars, Change in waist circumference less than the gender-specific median (males ⬍4.63 cm, females ⬍3.25 cm); open bars, change in waist circumference equal to or greater than the gender-specific median.


onstrated that baseline levels and change, over time, in waist circumference were significant predictors of changes in proinsulin concentration. It is important to point out that elevated absolute concentrations of proinsulin are a feature of the increased ␤-cell secretion that occurs as a compensatory mechanism for insulin resistance. On the other hand, disproportionate elevations in proinsulin (i.e. elevated levels after adjustment for insulin secretion) reflect ␤-cell dysfunction (3, 4, 23). It is conceivable, therefore, that the significant associations between waist circumference and C-peptideadjusted proinsulin concentrations reported here reflect an independent detrimental effect of abdominal adiposity on ␤-cell function. Univariate results from previous papers have indicated that proinsulin concentrations were elevated in obese subjects (5, 6, 8, 12, 52, 53) and that proinsulin was positively correlated with BMI and WHR (9, 21, 24 –27). The results of the present paper extend these findings by demonstrating independent associations with percent body fat and waist circumference in nondiabetic subjects after adjustment for age, sex, insulin secretion, and glucose concentrations. After additional adjustment for total body adiposity, waist circumference was significantly related to proinsulin concentration and further, waist circumference was prospectively associated with changes in proinsulin concentration. These strong and generally consistent positive associations between waist circumference and proinsulin concentration suggest a distinctive, unfavorable role for IAF in the natural history of ␤-cell dysfunction. The possibility of a direct detrimental effect of intraabdominal adipose tissue on the health of the ␤-cell, even before the development of diabetes, is supported by clinical studies demonstrating strong correlations between directly-measured IAF and insulin resistance (54), and by in vitro studies which have shown higher lipolytic activity of mesenteric and omental adipose tissue (55). As is reviewed by Ferrannini (1), it is conceivable that this IAF-related insulin resistance might exacerbate existing ␤-cell dysfunction or injury, which may be present relatively early in diabetes pathogenesis. Kahn et al. (11) have reported a significant unadjusted association between directly-measured IAF and proinsulin in Japanese-American subjects. Recent work by Unger’s group has suggested an alternative mechanism whereby excess adiposity might lead to increased proinsulin concentrations (56). The adipocytes of abdominally obese individuals release high levels of FFAs (56), which, in the short-term, causes ␤-cell hyperplasia and hyperinsulinaemia, but with chronic exposure (and subsequent increase in FFA levels), leads to functional and morphologic changes in ␤-cells and consequent diabetes (56). This notion has been supported with the documentation of substantial fat deposition in islets of obese rats and the demonstration of FFA-induced loss of glucose-stimulated insulin secretion (57). Increased FFAs also induce nitric oxide synthase, and Shimabukuro et al. (58, 59) have shown that elevated FFAs in rat ␤-cells cause increases in both nitric oxide levels and ceramide-mediated ␤-cell apoptosis (programmed cell death). The consequent reduction in the number of ␤-cells may result in elevated proinsulin concentration, in that the rate of secretion by remaining cells is increased, thereby decreasing the intracellular stores and forcing the


release of incompletely processed materials (60, 61). The recent demonstration of lipid deposition in human islets provides indirect evidence that this mechanism might also operate in humans (62, 63). This interpretation is contrary to conclusions drawn from studies of the association between obesity or insulin resistance and PI/I. Reported relationships between measures of obesity and PI/I have been inconsistent (5, 6, 10, 26, 27), although two studies (5, 10) have reported significant inverse associations between BMI and PI/I in subjects with NGT. These results, taken together with data from clinical studies using more detailed measures of insulin resistance, have been interpreted to suggest that nondiabetic ␤-cells are able to respond to the stress of insulin resistance (represented in the population-based studies by elevated adiposity) by increasing their secretion of true insulin without parallel or disproportionate increases in proinsulin secretion (21, 64). Though this hypothesis is clearly tenable, the results of studies using PI/I are not directly comparable with those from the present analysis, in that we elected not to use the PI/I. Rather, we used proinsulin concentration as our primary outcome variable, and we adjusted our models for C-peptide, a reliable surrogate measure of insulin secretion (48 –50). This approach is supported by the results of a recent paper by Vauhkonen et al. 65), who reported that the hepatic extraction of specific insulin was lower among offspring of diabetic patients with an insulin-resistant phenotype (IR group) compared with controls as well as offspring of patients with a deficient insulin-secretion phenotype (IS group). This resulted in a significantly lower PI/I in the IR group, compared with both the IS and control groups. However, there were no significant differences between the groups when proinsulinto-C-peptide levels were compared. In conclusion, this paper lends further support to the observation of elevated concentrations of proinsulin among subjects with IGT and type 2 DM, and it extends the phenomenon to Native Canadians from the central subarctic. We have further documented a possible detrimental role for intraabdominal adiposity in ␤-cell function. These observations require confirmation in future studies. In particular, further analyses employing multivariate adjustment using C-peptide would be of interest. Acknowledgments We acknowledge the following groups and individuals, whose cooperation was essential in the design and implementation of this project: the chief, council and community members of Sandy Lake First Nation; the Sandy Lake community surveyors (Louisa Kakegamic, Tina Noon, Madeliene Kakegamic, Elda Anishinabie, Mary Jane Williams, Areta Bekintis, Connie Kakegamic, Annette Rae, and Mary Mamakeesick); the Sandy Lake nurses; the staff of the University of Toronto Sioux Lookout Program; the Department of Clinical Epidemiology of the Samuel Lunenfeld Research Institute; Dr. Alexander Logan; and Annette Barnie. Received July 18, 2001. Accepted September 24, 2001. Address all correspondence and requests for reprints to: Anthony Hanley, Ph.D., Division of Epidemiology and Biostatistics, Samuel Lunenfeld Research Institute, Mt. Sinai Hospital, 600 University Avenue, Room 843, Toronto, Ontario, M5G 1X5, Canada. E-mail: [email protected]. This work was supported by grants from the NIH (91-DK-01 and 1-R21-DK-44597-01) and the Ontario Ministry of Health (04307). A.J.G.H.

82


was supported by Health Canada through a National Health Research and Development Programme Research Training Award, and is currently supported through a Post-Doctoral training award of the Canadian Institutes of Health Research. S.B.H. is a Career Scientist with the Ontario Ministry of Health. R.A.H. is a Career Investigator of the Heart and Stroke Foundation of Ontario (no. 2729).

References 1. Ferrannini E 1998 Insulin resistance versus insulin deficiency in non-insulin dependent diabetes mellitus: problems and prospects. Endocr Rev 19:477– 490 2. Gerich JE 1998 The genetic basis of type 2 diabetes mellitus: impaired insulin secretion versus impaired insulin sensitivity. Endocr Rev 19:491–503 3. Rhodes CJ, Alarcon C 1994 What ␤-cell defect could lead to hyperproinsulinemia in NIDDM? Some clues from recent advances made in understanding the proinsulin processing mechanism. Diabetes 43:511–517 4. Porte Jr D, Kahn SE 1989 Hyperproinsulinaemia and amyloid in NIDDM: clues to the etiology of islet beta cell function? Diabetes 38:1333–1336 5. Saad MF, Kahn SE, Nelson RG, Pettitt DJ, Knowler WC, Schwartz MW, Kowalyk S, Bennett PH, Porte Jr D 1990 Disproportionately elevated proinsulin in Pima Indians with non-insulin-dependent diabetes mellitus. J Clin Endocrinol Metab 70:1247–1253 6. Shiraishi I, Iwamoto Y, Kuzuya T, Matsuda A, Kamakura S 1991 Hyperinsulinaemia in obesity is not accompanied by an increase in serum proinsulin/ insulin ratio in groups of human subjects with and without glucose intolerance. Diabetologia 34:737–741 7. Williams DRR, Byrne C, Clark PMS, Cox L, Day NE, Rayman G, Wang T, Hales CN 1991 Raised proinsulin concentration as early indicator of beta-cell dysfunction. BMJ 303:95–96 8. Haffner SM, Bowsher RR, Mykkanen L, Hazuda HP, Mitchell BD, Valdez RA, Gingerich R, Monterossa A, Stern MP 1994 Proinsulin and specific insulin concentrations in high and low risk populations for NIDDM. Diabetes 43: 1490 –1493 9. Birkeland KI, Torjesen PA, Eriksson J, Vaaler S, Groop L 1994 Hyperproinsulinaemia of type II diabetes is not present before the development of hyperglycemia. Diabetes Care 17:1307–1310 10. Haffner SM, Stern MP, Miettinen H, Gingrich R, Bowsher RR 1995 Higher proinsulin and specific insulin are both associated with a parental history of diabetes in nondiabetic Mexican American subjects. Diabetes 44:1156 –1160 11. Kahn SE, Leonetti DL, Prigeon RL, Boyko EJ, Bergstrom RW, Fujimoto WY 1995 Relationship of proinsulin and insulin with non-insulin-dependent diabetes and coronary heart disease in Japanese American men: impact of obesity—Clinical Research Center Study. J Clin Endocrinol Metab 80:1399 –1406 12. Iochida LC, Marcopito LF, Franco LJ, Ferreira SR, Iunes M, Dalbosco IS, Russo EM 1996 Proinsulin and insulin levels according to glucose tolerance among Japanese-Brazilians, age 40 –79 years. Diabetes Res Clin Pract Suppl 34:S31–S35 13. Reaven GM, Chen Y-DI, Hollenbeck CB, Sheu WHH, Ostrega D, Polonsky KS 1993 Plasma insulin, C-peptide, and proinsulin concentrations in obese and non-obese individuals with varying degrees of glucose intolerance. J Clin Endocrinol Metab 76:44 – 48 14. Dornhorst A, Davies M, Anyaoku V, Hampton SM, Elkeles RS, Beard RW, Johnston DG 1991 Abnormalities in fasting circulating proinsulin concentration in mild gestational diabetes. Clin Endocrinol (Oxf) 34:211–213 15. Haffner SM, Gonzalez C, Mykkannen L, Stern M 1997 Total immunoreactive proinsulin, immunoreactive insulin, and specific insulin in relation to conversion to type 2 diabetes: the Mexico City Diabetes Study. Diabetologia 40:830 – 837 16. Kahn SE, Leonetti DL, Prigeon RL, Boyko EJ, Bergstom RW, Fujimoto WY 1995 Proinsulin as a marker for the development of type 2 diabetes in JapaneseAmerican men. Diabetes 44:173–179 17. Inoue I, Takahashi K, Katayama S, Harada Y, Negishi K, Ishii J, Shibazaki S, Nagai M, Kawazu S 1996 A higher proinsulin response to glucose loading predicts deteriorating fasting plasma glucose and worsening to diabetes in subjects with impaired glucose tolerance. Diabet Med 13:330 –336 18. Nijpels G, Popp-Snijders C, Kostense PJ, Bouter LM, Heine RJ 1996 Fasting proinsulin and 2-h post-load glucose levels predict the conversion to type 2 diabetes in subjects with impaired glucose tolerance: the Hoorn study. Diabetologia 39:113–118 19. Mykkanen L, Haffner SM, Kuusisto J, Pyorala K, Hales CN, Laakso M 1995 Serum proinsulin levels are disproportionately in early prediabetic subjects. Diabetologia 35:1176 –1182 20. Wareham, NJ, Byrne CD, Williams R, Day NE, Hales CN 1999 Fasting proinsulin concentrations predict the development of type 2 diabetes. Diabetes Care 22:262–270 21. Mykkanen L, Haffner SM, Hales CN, Ronnemaa T, Laakso M 1997 The relation of proinsulin, insulin and proinsulin-to-insulin ratio to insulin sensitivity and acute insulin response in normoglycaemic subjects. Diabetes 46: 1990 –1995 22. Mykkanen L, Zaccaro DJ, Hales CN, Festa A, Haffner SM 1999 The relation of proinsulin and insulin to insulin sensitivity and acute insulin response in


23. 24.

25. 26.

27.

28.

29. 30. 31. 32. 33.

34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

subjects with newly diagnosed type II diabetes: the Insulin Resistance Atherosclerosis Study. Diabetologia 42:1060 –1066 Kahn SE, Halban PA 1997 Release of incompletely processed proinsulin is the cause of the disproportionate proinsulinemia of NIDDM. Diabetes 46:1725– 1732 Nagi DK, Hendra TJ, Ryle AJ, Cooper TM, Temple RC, Clark PM, Schneider AE, Hales CN, Yudkin JS 1990 The relationship of concentrations of insulin, intact proinsulin, and 32–33 split proinsulin with cardiovascular risk factors in type 2 (non-insulin-dependent) diabetic subjects. Diabetologia 33:532–537 Haffner SM, Mykkanen L, Stern MP, Valdez RA, Heisserman JA, Bowsher RR 1993 Relationship of proinsulin and insulin to cardiovascular risk factors in nondiabetic subjects. Diabetes 42:1279 –1302 Mohamed-Ali V, Gould MM, Gillies S, Goubet S, Yudkin JS, Haines AP 1995 Association of proinsulin-like molecules with lipids and fibrinogen in nondiabetic subjects— evidence a modulating effect of insulin. Diabetoligia 38: 1110 –1116 Yudkin JS, Denver AE, Mohamed-Ali V, Ramaiya KL, Nagi DK, Goubet S, McLarty DG, Swai A 1997 The relationship of concentrations of insulin and proinsulin-like molecules with coronary heart disease prevalence and incidence: a study of two ethnic groups. Diabetes Care 20:1093–1100 Pouliot MC, Despres JP, Lemieux S, Moorjani S, Bouchard C, Tremblay A, Nadeau A, Lupien PJ 1994 Waist circumference and abdominal sagittal diameter: best simple anthropometric indexes of abdominal visceral adipose tissue accumulation and related cardiovascular risk in men and women. Am J Cardiol 73:460 – 468 Lemieux S, Prud’homme D, Bouchard C, Tremblay A, Despres J-P 1996 A single threshold value of waist girth identifies normal weight and overweight subjects with excess visceral adipose tissue. Am J Clin Nutr 64:685– 693 Jensen MD, Haymond MW, Rizza RA, Cryer PE, Miles JM 1989 Influence of body fat distribution on free fatty acid metabolism in obesity. J Clin Invest 83:1168 –1173 Roust LR, Jensen MD 1993 Postprandial free fatty acid kinetics are abnormal in upper body obesity. Diabetes 42:1567–1573 Herranz L, Zapata A, Grande C, Megia A, Pallardo LF 1998 Body fat distribution, insulin mediated suppression of non-esterified fatty acids and plasma triglycerides in obese subjects. Horm Metab Res 30:141–145 Carpentier A, Mittelman SD, Bergman RN, Giacca A, Lewis GF 2000 Prolonged elevation of plasma free fatty acids impairs pancreatic beta-cell function in obese nondiabetic humans but not in individuals with type 2 diabetes. Diabetes 49:399 – 408 Bergman RN, Ader M 2000 Free fatty acids and pathogenesis of type 2 diabetes mellitus. Trends Endocrinol Metab 11:351–356 Hanley AJG, Harris SB, Barnie A, Gittelsohn J, Wolever TMS, Logan A, Zinman B 1995 The Sandy Lake Health and Diabetes Project: design, methods and lessons learned. Chronic Dis Can 16:149 –156 Harris SB, Gittelsohn J, Hanley AJG, Barnie A, Wolever TMS, Gao J, Logan A, Zinman B 1997 The prevalence of NIDDM and associated risk factors in aboriginal Canadians. Diabetes Care 20:185–187 Harris SB, Zinman B, Hanley AJG, Gittelsohn J, Hegele RA, Connelly P, Shah B, Hux JE The impact of diabetes on cardiovascular risk factors and outcomes in a Native Canadian population. Diabetes Res Clin Pract, in press Young TK 1988 Health care and cultural change: the Indian experience in the central subarctic. Toronto: University of Toronto Press Young TK 1988 Are subarctic Indians undergoing the epidemiologic transition? Soc Sci Med 26:659 – 671 Delisle HF, Ekoe J 1993 Prevalence of non-insulin-dependant diabetes mellitus and impaired glucose tolerance in two Algonquin communities in Quebec. Can Med Assoc J 148:41– 47 World Health Organization 1985 Diabetes mellitus: report of a WHO Study Group. Geneva: WHO; Technical Report Series No. 727 Bowsher RR, Wolney JD, Frank BH 1992 A rapid and sensitive radioimmunoassay for the measurement of proinsulin in human serum. Diabetes 41: 1084 –1090 Halban PA 1991 Structural domains and molecular lifestyles of insulin and its precursors in the pancreatic beta cell. Diabetologia 34:767–778 Hanley AJG, Harris SB, Barnie A, Smith J, Logan A, Zinman B 1994 Usefulness of bioelectrical impedance analysis in a population-based study of diabetes among Native Canadians. Int J Obes 18(Suppl 2):O383 Sakamoto Y, Miura J, Yamaguchi Y, Ohno M, Ike Y 1994 Usefulness of bioelectrical impedance analysis for measurement of body fat. Int J Obes 18(Suppl 2):P501 Nunez C, Gallagher D, Visser M, Pi-Sunyer FX, Wang Z, Heymsfield SB 1997 Bioimpedance analysis: evaluation of a leg-to-leg system based on pressure contact foot-pad electrodes. Med Sci Sports Exerc 29:524 –531 Tsui EYL, Gao XJ, Zinman B 1998 Bioelectrical impedance analysis (BIA) using bipolar foot electrodes in the assessment of body composition in type 2 diabetes mellitus. Diabet Med 15:125–128 Van Cauter E, Mestrez F, Sturis J, Polonsky KS 1992 Estimation of insulin secretion rates from C-peptide levels: comparison of individual and standard kinetic parameters for C-peptide clearance. Diabetes 41:368 –377 Kronmal RA 1993 Spurious correlation and the fallacy of the ratio standard revisited. J R Stat Soc 156:379 –392



50. Allison DB, Paultre F, Goran MI, Poehlman ET, Heymsfield SB 1995 Statistical considerations regarding the use of ratios to adjust data. Int J Obes Relat Metab Disord 19:644 – 652 51. SAS Institute 1990 SAS/STAT User’s Guide, Version 6, 4th ed. Cary, NC: SAS Institute Inc 52. Koivisto VA, Yki-Jarvinen H, Harling SG, Pelkonen R 1986 The effect of exogenous hyperinsulinaemia on proinsulin secretion in normal man, obese subjects, and patients with insulinoma. J Clin Endocrinol Metab 63:1117–1120 53. Proudler AJ, Godsland IF, Stevenson JC 1994 Insulin propeptides in conditions associated with insulin resistance in humans and their relevance to insulin measurements. Metabolism 43:446 – 449 54. Banerji MA, Chaiken RL, Gordon D, Kral JG, Lebovitz HE 1995 Does intraabdominal adipose tissue in black men determine whether NIDDM is insulinresistant or insulin-sensitive? Diabetes 44:141–146 55. Abate N 1996 Insulin resistance and obesity: the role of fat distribution pattern. Diabetes Care 19:292–294 56. Unger RH 1995 Lipotoxicity in the pathogenesis of obesity dependent NIDDM. Diabetes 44:863– 870 57. Lee Y, Hirose H, Ohneda M, Johnson JH, McGarry JD, Unger RH 1994 Beta-cell lipotoxicity in the pathogenesis of non-insulin-dependent diabetes mellitus of obese rats: impairment of adipocyte-beta cell relationships. Proc Natl Acad Sci USA 91:10878 –10882 58. Shimabukuro M, Ohneda M, Lee Y, Unger RH 1997 Role of nitric oxide in obesity-induced beta-cell disease. J Clin Invest 100:290 –295 59. Shimabukuro M, Zhou Y-T, Levi M, Unger RH 1998 Fatty acid-induced beta

60.

61.

62. 63.

64.

65.

cell apoptosis: a link between obesity and diabetes. Proc Natl Acad Sci USA 95:2498 –2502 Bollheimer LC, Skelly RH, Chester MW, McGarry JD, Rhodes CJ 1998 Chronic exposure to free fatty acid reduces pancreatic beta-cell insulin content by increasing basal insulin secretion that is not compensated for by a corresponding increase in proinsulin biosynthesis translation. J Clin Invest 101: 1094 –1101 Furukawa H, Carroll RJ, Swift HH, Steiner DF 1999 Long term elevation of free fatty acids leads to delayed processing of proinsulin and prohormone convertases 2 and 3 in the pancreatic beta cell line MIN6. Diabetes 48:1395–1401 Unger RH, Orci L 2001 Diseases of liporegulation: new perspective on obesity and related disorders. FASEB J 15:312–321 Cnop M, Grupping A, Hoorens A, Bouwens L, Pipeleers-Marichal M, Pipeleers D 2000 Endocytosis of low-density lipoprotein by human pancreatic beta cells and uptake in lipid-storing vesicles, which increase with age. Am J Pathol 156:237–244 Wang P-W, Abbasi F, Carantoni M, Chen Y-DI, Azhar S, Reaven GM 1997 Insulin resistance does not change the ratio of proinsulin to insulin in normal volunteers. J Clin Endocrinol Metab 82:3221–3224 Vauhkonen IKJ, Niskanen LK, Mykkanen L, Haffner SM, Uusitupa MIL, Laakso M 2000 Hyperproinsulinaemia is not a characteristic feature in the offspring of patients with different phenotypes of type II diabetes. Eur J Endocrinol 143:251–260

Project Announcement Genetically Modified Animals in Endocrinology As a service to the endocrine community, Endocrine Reviews intends to publish bibliographies of papers describing knockout, transgenic, and mutant animals that may be useful in the study of endocrinology. In the print version of the journal, we will publish subject-limited bibliographies as the individual sections become available. We also intend to create a cumulative database to be made available on the web in a searchable format. At this time, we would like to hear what enhancements would be desirable on this web site. Readers are encouraged to contact the editorial office with bibliographic information about knockout, transgenic, and mutant animals that they would wish to have included in the database; please include the species and the citation for the article in which the original description appeared. In addition, suggestions regarding topics that we should consider adding to our bibliographies would be appreciated. Address your contributions to the database via e-mail or standard mail, using the following addresses: Dr. E. Brad Thompson/Endocrine Reviews, The University of Texas Medical Branch, Room 111C, Basic Science Building, Galveston, TX 77555-0628 USA. [email protected]