Receiver operating characteristic curve analysis of BMI in assessing obesity among adult Bengalee males in India.

The present study attempted to understand the appropriateness of different body mass index (BMI) cut-off points in assessing obesity. Four hundred thirty adult Bengalee males with mean age 36.48 +/- 12.23 years (mean +/- SD) from West Bengal, India were studied. Height, weight and percent body fat were measured and BMI was derived following standard equation. Receiver operating characteristic (ROC) curve analysis demonstrated low sensitivity and high specificity of international (> or = 30 kg/m2) and proposed (> or = 25 kg/m2) BMI cut-off points in identifying individuals with obesity. Furthermore, these international and proposed BMI cut-off points also found to be in substantial misclassification to assess obesity as much as 32.09% and 17.44% respectively. However, based on ROC curve analysis, a BMI cut-off point of > or = 24 kg/m2 revealed optimal sensitivity (83.66%) and specificity (90.61%) and as well as less misclassification (11.86%) in assessing obesity among the adult Bengalee males. The present study accentuated that the international and proposed BMI cut-off points so far might not be appropriate in assessing obesity and on the other hand, lower BMI cut-off point (> or = 24 kg/m2) has relatively high sensitivity and specificity in assessing obesity as well. Therefore, the present study envisages the requirement of further lowering down of BMI cut-off point in assessing excess body fat in adult Bengalee males.