The predictive relationship between baseline insulin and glucose with subclinical carotid atherosclerosis after 5 years in a multi-ethnic cohort.

Background And Aims: We aimed at exploring the relationship between baseline insulin and glucose and the progression of carotid atherosclerosis in a multi-ethnic cohort.

Methods: Males and females (n = 797) of European, Chinese, South Asian and Aboriginal origin were assessed as part of the Multicultural Community Health Assessment Trial (MCHAT) study for socio-demographics, smoking status, fasting insulin and glucose at baseline. IMT, plaque area and total area were assessed after 5 years.

Results: A total of 545 participants returned after 5 years for a follow-up assessment. Average age of the study participants was 47.5 (SD 8.9) years. At baseline, the median and interquartile range for insulin was 62.0 (49.5) pmol/L, and glucose was 5.2 (0.60) mmol/L. Baseline glucose and insulin predicted the 5-year progression of atherosclerosis in our models, after adjusting for covariates. We found significant insulin-ethnicity interactions in the IMT model (p = 0.044) with the slope of the relationship showing that for every percentage change in insulin the Europeans experienced 7.3% more increase in IMT at 5 years than the Aboriginals. In the plaque area and total area models, there were significant glucose-ethnicity interactions (p = 0.009 and p=0.016 respectively), with the slope showing a 101% and 121% increase for plaque area and total area, respectively, in Europeans, at 5 years per percent change in glucose at baseline. Logistic regression found a significant glucose-ethnicity interaction with the presence of plaques (OR = 0.31, p = 0.03) such that compared to the Europeans, the South Asians had a lower odds of developing plaque presence. Similarly, we found glucose-ethnicity interactions in the logistic regression when comparing the Chinese to the Europeans (OR = 0.2, p=0.005), with the Chinese being less likely to develop plaque presence.

Conclusions: Ethnicity modifies the predictive relationship between insulin and glucose with sub-clinical indicators of carotid atherosclerosis but not consistently so.