Table 3

The RR of incident type 2 diabetes at the 6-year follow-up in Cox proportional hazards models based on A1C status at baseline

RR (95% CI)PRR (95% CI)P
A1C ≥5.6% (vs. <5.6%) in the entire study population
 Model A*3.44 (2.87–4.13)<0.0014.60 (3.75–5.66)<0.001
 Model B3.17 (2.62–3.84)<0.0014.00 (3.24–4.95)<0.001
 Model C2.41 (1.98–2.93)<0.0013.06 (2.46–3.81)<0.001
A1C ≥5.8% (vs. <5.8%) in subjects with IFG
 Model A*3.15 (2.13–4.64)<0.0016.29 (3.03–13.05)<0.001
 Model B3.57 (2.36–5.41)<0.0015.99 (2.83–12.66)<0.001
 Model C3.47 (2.27–5.29)<0.0015.15 (2.39–11.11)<0.001
  • There was significant interaction between A1C and sex. The interaction between A1C and FPG also was significant.

  • *Age adjusted.

  • †Model A and waist circumference, family history of diabetes, living in urban area, hypertension, smoking, and alcohol intake were adjusted.

  • ‡Model B and triglycerides (log), HDL cholesterol, HOMA-IR (log), HOMA-β (log), and hsCRP (log) were adjusted.