Résumé
Objective: Assessment of insulin resistance (IR) is essential in non-diabetic patients with obesity. Thus study aims to identify the best determinants of IR and to propose an original approach for routine assessment of IR in obesity. Design and patients: All adult with obesity defined by a body mass index >= 30 kg/m(2), evaluated in the Nutrition Department between January 2010 and January 2015 were included in this cross-sectional study. Patients with diabetes were excluded. IR was diagnosed according to the HOMA-IR. Based on a logistic regression, we determined a composite score of IR. We then tested the variables with a principal component analysis and a hierarchical clustering analysis. Results: A total of 498 patients with obesity were included. IR was associated with grade III obesity (OR = 2.6[1.6-4.4], p < 0.001), HbA1c >= 5.7% (OR = 2.6[ 1.7-4.0], p < 0.001), hypertriglyceridemia > 1.7 mmol/l (OR = 3.0[ 2.0-4.5], p < 0.001) and age (OR = 0.98[ 0.96-0.99], p = 0.002). Exploratory visual analysis using factor map and clustering analysis revealed that lipid and carbohydrates metabolism abnormalities were correlated with insulin resistance but not with excessive fat accumulation and low-grade inflammation. Conclusions: Our results highlight the interest of simple blood tests such as HbA1c and triglyceride determination, which associated with BMI, may be widely available tools for screening IR in obese patients.