Indices Calculated by the OGTT-DM Risk Stratifier Tool
Abbreviations: OGTT = oral glucose tolerance test; G₀, G₃₀, G₆₀, G₁₂₀ = glucose at fasting, 30 min, 60 min, and 120 min; I₀, I₃₀ = insulin at fasting and 30 min. Gmean = mean of 0, 30, 60, 90, and 120 min glucose values; Imean = mean of 0, 30, 60, 90, and 120 min insulin values. Glucose values are in mg/dL and insulin in mU/L, except for the Stumvoll equation, which uses insulin in pmol/L and glucose in mmol/L (conversion: insulin [mU/L] × 6; glucose [mg/dL] ÷ 18).
1. Matsuda Index (Insulin Sensitivity Index, ISI)
HOMA-IR correlates well with basal hepatic insulin resistance (higher HOMA-IR indicating greater hepatic resistance), but correlates poorly with peripheral resistance as measured by total glucose disposal during the hyperinsulinemic–euglycemic clamp. In contrast, the Matsuda index correlates well with total glucose disposal (lower Matsuda values indicating greater peripheral insulin resistance), but correlates poorly with basal hepatic insulin resistance.7
2. Insulinogenic Index (IGI)
Estimates early-phase insulin secretion (β-cell function) in response to an oral glucose challenge. In SAHS, among individuals with normal glucose tolerance (NGT), those in the 10th percentile of IGI had a 7.69-fold higher risk of type 2 diabetes compared with those in the 90th percentile; risk was even greater in IGT (OR ≈20).8
3. Disposition Index (DI)
Reflects β-cell secretory capacity in the context of insulin sensitivity/resistance. In SAHS, a DI < 3.1 had sensitivity 83% and specificity 77% for predicting type 2 diabetes after 7–8 years of follow-up.9
4. Homeostatic Model Assessment of Insulin Resistance (HOMA-IR)
Correlates well with basal hepatic insulin resistance (higher HOMA-IR indicating greater hepatic resistance), but correlates poorly with peripheral resistance as measured by total glucose disposal during the hyperinsulinemic–euglycemic clamp. The Matsuda index provides a complementary estimate that better reflects peripheral insulin sensitivity.7
5. PG AUC (Weighted) — Weighted Plasma Glucose Area Under the Curve
Approximates integrated glucose exposure during OGTT using a weighted mean of glucose at 0, 30, 60, and 120 minutes. In one study, a glucose AUC cutoff of 290 mg·h/dL showed high discrimination for identifying IFG, IGT, and diabetes (ROC AUC 0.96; sensitivity 90%; specificity 93%).10
Although PG AUC is not incorporated into the risk-stratifier algorithm, it may provide supplementary information when interpreting OGTT-based diabetes risk.
6. Stumvoll First-Phase Insulin Release Index (Stumvoll 1st-phase)
Requires unit conversion: insulin in pmol/L and glucose in mmol/L (insulin [mU/L] × 6; glucose [mg/dL] ÷ 18).
Stumvoll-derived first-phase insulin secretion correlates more strongly with clamp-derived measures of insulin secretion than IGI.11 No validated cutoff has been established to define increased future diabetes risk, and OGTT-derived first-phase insulin secretion should be interpreted cautiously. In the original validation study, limits of agreement between OGTT-derived and clamp-measured first-phase insulin secretion (mean ±2 SD) spanned approximately −600 to +600 pmol/L. The cutoff used in our risk stratifier is empiric and corresponds to the 25th percentile derived from analysis of 117 donor OGTTs at our center.