BACKGROUND The objective of this study is to compare, in obese women, the assessment of insulin sensitivity and the presence of hypoglycemia between the usual and low doses of insulin used in the short insulin tolerance test (ITT). METHODS The patients were 12 obese women on whom a randomized, double-masked, auto-controlled clinical trial was carried out. An ITT was performed on each volunteer in duplicate with insulin at 0.1 (usual dose) or 0.05 U/kg (low dose). RESULTS The constant for the rate serum glucose disappearance calculated for the ITT was not significantly different between both tests (4.3 +/- 0.5 vs. 4.4 +/- 1.0%/min, usual dose and low dose of insulin, respectively; p = 0.49). There was a significant correlation between both tests (r = 0.59, r2 = 0.34, p < 0.05). Differences between both tests had estimated limits of agreement of -0.97 to 0.65%/min. Between tests, the coefficient of variation was 16%. No subject developed hypoglycemia with any of the ITT measurements employing usual or low doses of insulin. CONCLUSIONS We recommend a low dose of insulin in the assessment of insulin sensitivity with ITT in obese people. In our study usual and low doses of insulin used in the ITT were safe and had similar results between both doses for assessing insulin sensitivity.