The paradox of high apolipoprotein A-I levels independently predicting incident type-2 diabetes among Turks

Abstract

Background Predictive value of apolipoprotein (apo) A-I for incident hypertension, metabolic syndrome (MetS), type 2 diabetes (DM) and coronary heart disease (CHD) needs further exploration. Methods A representative sample of Turkish adults was studied with this purpose prospectively. Sex-specific apoA-I tertiles were examined regarding cardiometabolic risk. Results and conclusions A total of 1044 men and 1067 women (aged 49 ± 12 years at baseline) were followed up over 7.4 years. High serum apoA-I levels were significantly associated in multivariable analysis with female sex, aging, alcohol intake, (inversely) cigarette smoking and, in women, with systolic blood pressure. Risk of diabetes was predicted in logistic regression in both genders by top versus bottom apoA-I tertile (RR 1.98; [95%CI 1.31; 3.0]), additive to age, body mass index (BMI), C-reactive protein (CRP), HDL-cholesterol and lipid lowering drugs. By adding sex hormone-binding globulin to the model in a subset of the sample, the association between high apoA-I and incident diabetes was attenuated only in women. ApoA-I tertiles tended to be positively associated also with hypertension and CHD only in women but this did not reach significance. High compared with low serum apoA-I levels nearly double the risk for incident diabetes, additively to age, BMI, CRP, HDL-cholesterol among Turks. Systemic inflammation concomitant with prevailing MetS might turn apoA-I into proinflammatory particles.