Table 2—

Models for prediction scores from stepwise proportional hazards regression in the development sample (n = 335)*

Coefficient ± SEχ2Hazard ratio (95% CI)
Model 1
    Log-BMI (kg/m2)2.85 ± 0.5824.217.2 (5.5–53.5)
    Age (years)−0.11 ± 0.0222.50.89 (0.85–0.94)
    Total glucose (mg/dl)/1000.84 ± 0.081032.33 (1.98–2.74)
    Log–fasting C-peptide (ng/ml)0.39 ± 0.194.31.48 (1.02–2.13)
    Total C-peptide (ng/ml)/10−0.79 ± 0.2212.80.45 (0.29–0.70)
    FPIR less than the 10th percentile0.49 ± 0.215.41.63 (1.08–2.47)
Model 2
    Log-BMI (kg/m2)2.41 ± 0.5519.511.2 (3.8–32.8)
    Age (years)−0.09 ± 0.0219.10.91 (0.88–0.95)
    Total glucose (mg/dl)/1000.85 ± 0.081112.34 (2.00–2.74)
    Log–fasting C-peptide (ng/ml)0.41 ± 0.194.81.51 (1.05–2.19)
    Total C-peptide (ng/ml)/10−0.93 ± 0.2217.90.40 (0.26–0.61)
  • *

    * A total of 124 of the participants developed type 1 diabetes;

  • P < 0.001;

  • P < 0.05.