TABLE 4

Poisson regression analyses of the ability of candidate indexes of IR to predict incident diabetes with combined analysis using data from the SAHS, the MCDS, and the IRAS

AROC curve
Top 10% vs. bottom 90%
IGT
NGT
IGT
NGT
AROCrankAROCrankRRrankRRrank
-ISI0,12067.3170.8*21.9612.931
-ISIgly_a58.81368.831.47132.882
-SiM64.8571.2*11.7722.763
-ISIgly_b65.4268.741.597.52.746.5
-QUICKI65.4368.651.597.52.746.5
-Stum_nodem56.01767.271.26182.5212
-BFSI63.9667.081.6142.659
-ISI-2h56.61565.2131.26172.5310
-McAuley62.2866.5111.45142.4616
ln(FI)61.8966.5121.50112.4714.5
-Stum_wdem56.01668.461.27162.4017
-ISI65.2466.8101.597.52.746.5
BMI59.21266.891.5952.5311
-Raynaud62.2764.6141.50112.4714.5
FIRI59.71160.8151.597.52.746.5
HOMA-IR59.81059.9161.6232.744
FI58.11459.9171.50112.4713
IGR55.81855.2191.32151.8019
IGR-2h53.51959.3181.14191.9218
  • Ordering of indexes in this table were maintained based on ranking in the unadjusted AROC column in the pooled analysis (Table 3). Unadjusted analysis was by glucose tolerance status. For consistency in the direction of the associations in the tables, we have taken the negative of indices that estimate SI (versus insulin resistance).

  • *

    * AROCs for ISI0,120 and SiM are not significantly different (P = 0.68). BFSI, Bennett’s fasting insulin sensitivity index; FI, fasting insulin; IGR, insulin/glucose ratio; Stum_wdem, Stumvoll index without demographic variables.