SERUM-INSULIN AND ECG ABNORMALITIES SUGGESTING CORONARY HEART-DISEASEIN THE POPULATIONS OF MAURITIUS AND NAURU - CROSS-SECTIONAL AND LONGITUDINAL ASSOCIATIONS

Cross-sectional associations between insulin and ECG abnormalities sug gestive of 'possible' and 'probable' coronary heart disease (CHD) in t he populations of Nauru (n - 568) and Mauritius (n = 3280) have been e xamined in both non-diabetic and diabetic subjects. Additionally, the longitudinal relationship between baseline insulin and incident ECG ab normalities has been explored in non-diabetic Nauruans (n = 177) over 5 years. Age-adjusted mean 2-hour serum insulin was generally higher i n subjects with ECG abnormalities than those with a normal ECG, but th e difference was significant only for non-diabetic Mauritian men (p < 0.01). There was no clear association between prevalence of ECG abnorm alities and quintiles of fasting or 2-hr insulin in Mauritians, and in non-diabetic Nauruans there was a non-significant positive associatio n between prevalence of ECG abnormalities and tertiles of 2-hr insulin . Logistic regression analyses showed a slight positive association be tween 2-hr insulin and ECG abnormalities in non-diabetic Mauritians (p = 0.06 in males, p = 0.09 in fem.ales), and non-diabetic male Nauruan s (p = 0.054) independent of possible confounders. Fasting insulin was not associated in any group. In longitudinal analyses in Nauruans the re were no significant differences in mean baseline fasting or 2-hr se rum insulin between subjects who maintained a normal ECG and those who developed abnormalities consistent with 'possible CHD' (there were no changes suggestive of 'probable CHD'). The incidence of ECG changes s uggesting 'possible CHD' was slightly higher in the upper tertile of b aseline fasting insulin in both sexes and 2-hr insulin in females, but when other factors were accounted for, multiple logistic regression a nalyses did not support this finding. Data from the populations of Mau ritius and Nauru do not support a major role for serum insulin in ECG abnormalities suggestive of CHD.