Low-level serum arsenic and its effect on insulin and the glucose /insulin ratio of non-diabetic fasting human subjects

Human exposure to arsenic can result in adverse health consequences including cancer, hypertension, vascular disease and diabetes. Alterations in serum insulin and subsequently, the glucose/insulin ratio (G/I) in serum, have been considered for use as predictors for type 2 diabetes mellitus. The overall objective of this study was to examine the hypothesis that the presence of low-level arsenic in the serum decreases insulin and increases the G/I ratio in the serum of fasting, non-diabetic subjects. This was accomplished by evaluating the association between serum insulin and serum arsenic levels and the association between the G/I ratio and serum arsenic while controlling for other potential confounding factors such as: age, race, income, education and BMI, glucose, triglycerides, free fatty acids, leptin and percent fat mass. The study described in this dissertation was carried out by analyzing serum for arsenic and evaluating the associations described above. The source of the serum samples used in this study was serum obtained from subjects originally recruited for three studies carried out previously with different objectives. All clinical variables such as serum glucose, insulin, triglycerides, free fatty acids and leptin were analyzed for the original studies. Arsenic concentrations of the samples used in this dissertation were analyzed at the Children\u27s Nutrition Research Center of Baylor College of Medicine in Houston, Texas. Two hundred thirty eight serum samples were selected from the original studies for arsenic analysis in this study. Significant differences in mean insulin levels were observed for different age groups, race, income, education, and body mass index (BMI). Significant differences were also seen in the means for the G/I ratio for race, income, education and BMI. No significant differences were observed in the means of total serum arsenic concentrations for the different groups. In stratifying the study population on the basis of serum arsenic levels, the mean insulin level was lower for the subjects in the higher group compared to the lower group, but the difference was not statistically significant. Controlling for confounders, the results of the regression analyses suggested an inverse relationship between insulin and arsenic, but was not statistically significant. The results also suggested that the G/I ratio increases with arsenic levels, although this finding was not statistically significant