Insulin sensitivity (Si) assessment in lean and overweight subjects using two different protocols and updated software

Background. The standard frequently-sampled intravenous glucose tolerance test (FSIVGTT) is an alternative procedure to the clamp technique for estimating the insulin sensitivity (Si) parameter. The goal of this study was to compare Si in lean and overweight individuals in addition to assessing intra-individual reproducibility using two different protocols and updated software. Methods. FSIVGTT was carried out in 14 lean (BMI < 25 kg/m<SU2</SU) and 14 overweight (BMI > 25 kg/m<SU2</SU) subjects using two different protocols; full (29 samples) and short (12 samples). For reproducibility assessment four normal subjects (triplicate on three and twice on one) were recruited to undergo the same procedure at 1-week intervals. Data analysis was performed using COMAL and Minmod Millennium software. Results. Mean Si (10<SU-4</SUmin<SU-1</SU[pmol/l]<SU-1</SU) values were significantly different between lean and overweight subjects (p < 0.001) but not between the two protocols using both software packages. For the full and short protocols, Si values were more closely related in lean versus overweight subjects using either COMAL (r = 0.98, p < 0.001), (r = 0.89, p < 0.001) or Minmod Millennium (r = 0.99, p < 0.001), (r = 0.85, p < 0.001) software respectively. The intra-individual reproducibility (%CV) of Si (COMAL) in full versus short protocol was 18.3 +/- 11.1% and 13.7 +/- 1.9% respectively. Reproducibility values for Si (Minmod Millenium) in full versus short protocols were 14.3 +/- 3.8 and 14.9 +/- 1.9% respectively. Conclusions. Si can be assessed accurately by a short protocol FSIVGTT in normal individuals. The short protocol may give less acceptable results for insulin sensitivity in individuals who have normal glucose tolerance but high BMI.