Circulating tumor necrosis factor-alpha concentrations in a native Canadian population with high rates of type 2 diabetes mellitus.

Recent research suggests that tumor necrosis factor-alpha (TNF alpha) may play an important role in obesity-associated insulin resistance and diabetes. We studied the relationship between TNF alpha and the anthropometric and physiological variables associated with insulin resistance and diabetes in an isolated Native Canadian population with very high rates of type 2 diabetes mellitus (DM). A stratified random sample (n = 80) of participants was selected from a population-based survey designed to determine the prevalence of type 2 DM and its associated risk factors. Fasting blood samples for glucose, insulin, triglyceride, leptin, and TNF alpha were collected; a 75-g oral glucose tolerance test was administered, and a second blood sample was drawn after 120 min. Insulin resistance was estimated using the homeostasis assessment (HOMA) model. Systolic and diastolic blood pressure (BP), height, weight, and waist and hip circumferences were determined, and percent body fat was estimated using biological impedance analysis. The relationship between circulating concentrations of TNF alpha and the other variables was assessed using Spearman correlation coefficients, analysis of covariance, and multiple linear regression. The mean TNF alpha concentration was 5.6 pg/mL (SD = 2.18) and ranged from 2.0-12.9 pg/mL, with no difference between men and women (P = 0.67). There were moderate, but statistically significant, correlations between TNF alpha and fasting insulin, HOMA insulin resistance (HOMA IR) waist circumference, fasting triglyceride, and systolic BP (r = 0.23-0.34; all P < 0.05); in all cases, coefficients for females were stronger than those for males. Individuals with normal glucose tolerance had lower log TNF alpha concentrations than those with impaired glucose tolerance or type 2 DM (both P = 0.03, adjusted for age and sex), although differences were not significant after adjustment for HOMA IR (both P > 0.25). Regression analysis indicated that log HOMA IR and log systolic BP were significant independent contributors to variations in log TNF alpha concentration (model r2 = 0.32). We conclude that in this homogeneous Native Canadian population, circulating TNF alpha concentrations are positively correlated with insulin resistance across a spectrum of glucose tolerance. The data suggest a possible role for TNF alpha in the pathophysiology of insulin resistance.