Background: It was suggested that inflammation may mediate or modify biological effects of adiponectin. Few studies examined the association between circulating adiponectin levels and type 2 diabetes (T2DM) while controlling for variables related to inflammation. In addition, East Asians were reported to have lower adiponectin levels but higher diabetes prevalence at a given degree of obesity than Caucasians, raising some possibility that the adiponectin-diabetes association may differ by race. Therefore, we prospectively investigated the associations with a number of covariates including C-reactive protein and smoking status in a cohort of Japanese workers aged 35-66 years.
Methods: Serum adiponectin concentration and other covariates were obtained in 2002 for 3008 civil servants free of T2DM at baseline in urban/suburban Japan. T2DM incidence was defined as the year when annually assessed fasting blood glucose level first exceeded 126 mg/dL or self-reported initiation of medication through 2007. T2DM incidence was examined in relation to the adiponectin quintile.
Results: Age- and sex-adjusted homeostasis model assessment insulin resistance was inversely associated with adiponectin quintiles at baseline. During six years of follow-up, 164 individuals developed T2DM. In a fully adjusted model, hazard ratios (95% confidence intervals) of T2DM in Q2 to Q5 compared with that in Q1 were 0.62 (0.41-0.94), 0.44 (0.25-0.77), 0.40 (0.20-0.78) and 0.85 (0.48-1.49), respectively.
Conclusions: Low adiponectin was related to increased incidence of T2DM independent of baseline levels of blood glucose, insulin and C-reactive protein as well as other confounding variables in middle-aged Japanese. Whether high adiponectin is linearly associated with decreased T2DM risk needs further investigation.
Copyright © 2012 John Wiley & Sons, Ltd.