Identificación de individuos con resistencia a la insulina mediante mediciones clínicas y marcadores bioquímicos de rutina. Construcción de un índice de riesgo individual

Abstract

This study aimed to know the markers and routine biochemical measures that are associated to the insulin resistance (IR) and to develop an index of individual IR risk from them. Cross-sectional study made in Primary Health Care population of both genders between 40 and 74 years (n = 2,143). A representative sample was obtained by simple random sampling of 305 patients after excluding the diabetic subjects. Sociodemographic variables, background, examinations, routine analyses as well as fasting insulin levels were obtained. IR was considered if HOMA was higher than 2.9. A step by step logistic regression was done to obtain the best variables to predict IR. A logistic equation, categorical scale and simple additive scale from the beta coefficients was then constructed and was compared with other instruments designed to predict IR. IR prevalence was 25.2%. There were no differences between genders or by age. The four variables that entered the model were fasting plasma glucose, BMI, HDL cholesterol and diastolic blood pressure. The logistic model had good adjustment. The logistic equation was: IR = 1/ 1 + exp (-[-21.011) - [0.119 * fasting plasma glucose] - [0.231 * BMI] - [-0.046 * HDL cholesterol] - [0.048 * diastolic blood pressure]). The scale constructed assesses each subject between -1 and 7 points; cutoff to predict IR was established at 3.5 points, obtaining a sensitivity and specificity similar to the McAuley index and better than the triglycerides/HDL cholesterol ratio, the first model of Stern and the diagnosis of metabolic syndrome according to ATP-III. A very easy-to-use instrument has been obtained to predict IR by means of exploratory measures and routine biochemical measures, which makes it possible to select the patients at the greatest risk in order to intensify preventive actions in them.