Table 2. Multivariate-adjusted hazard ratios and the 95% confidence intervals for predicting the effect of weight-change slope and RMSE on the incident type 2 diabetes, among the participants of Aichi Workers' Cohort Study, 2002–2014.
| Men | Women | |
|---|---|---|
| Number of incident cases | 330 | 70 |
| Total number of participants | 3317 | 917 |
| Person years | 32695.8 | 8408.3 |
| Crude incidence (1/1000 person-years) | 10.1 | 8.3 |
| Hazard ratio (95% CI) | ||
| Slope | ||
| Model 1a | 2.55 (1.72–3.78) | 5.60 (2.48–12.62) |
| Model 2b | 1.80 (1.17–2.77) | 2.78 (1.07–7.23) |
| RMSE | ||
| Model 1a | 1.09 (1.00–1.19) | 1.03 (0.85–1.24) |
| Model 2b | 1.08 (1.00–1.18) | 1.02 (0.84–1.25) |
Abbreviations: CI, confidence intervals; RMSE: root-mean-square-error.
a
Model 1 included slope (continuous), RMSE (continuous), age (<50 and ⩾50 years old), smoking status (never, former and current), drinking frequency (<1, 1–4 and 5–7 occasions per week), leisure-time physical activity (yes, no), everyday eating of breakfast (yes, no), usual sleep duration (⩾7, <7 h per day), family history of diabetes (yes, no) and FBG level at baseline (<100 mg dl−1, ⩾100 mg dl−1).
b
Model 2 included variables in the model 1 plus baseline overweight (baseline BMI <25, baseline BMI ⩾25).