Baseline characteristics and predictors (OR with 95% CI) of the incidence of diabetes during a 15-year follow-up period
Characteristic | Incident diabetes | Univariate age-adjusted P value | Multivariate logistic regression | ||||
---|---|---|---|---|---|---|---|
No | Yes | OR (95% CI) | P | ||||
n | 6,318 | 281 | — | — | — | ||
Sleep disturbances | |||||||
None (%) | 88.7 | 80.1 | <0.001 | Reference | — | ||
Difficulty falling asleep or regular use of hypnotics (%) | 9.1 | 15.3 | 0.005 | 1.52 (1.05–2.20) | 0.025 | ||
Both types of sleep disturbance (%) | 2.3 | 4.6 | 0.034 | 1.78 (0.96–3.32)* | 0.070 | ||
Age (years) | 44.5 ± 4.1 | 45.9 ± 3.3 | <0.001 | 1.11 (1.07–1.17) | <0.001 | ||
Systolic blood pressure (mmHg) | 125.5 ± 13.7 | 131.8 ± 16.7 | <0.001 | 1.06 (0.94–1.20)† | 0.317 | ||
Diastolic blood pressure (mmHg) | 84.9 ± 9.2 | 89.6 ± 10.8 | <0.001 | — | — | ||
Use of antihypertensive medication (%) | 2.9 | 8.2 | <0.001 | 1.58 (0.95–2.61) | 0.077 | ||
Fasting whole blood glucose (mmol/l) | 4.8 ± 0.5 | 5.1 ± 0.5 | <0.001 | 1.72 (1.49–1.99)† | <0.001 | ||
Heart rate (bpm) | 69.2 ± 10.3 | 72.1 ± 11.4 | <0.001 | 1.13 (0.99–1.30)‡ | 0.068 | ||
BMI (kg/m2) | 24.4 ± 3.0 | 26.9 ± 3.5 | <0.001 | 1.21 (1.17–1.26) | <0.001 | ||
Change in BMI from baseline (kg/m2) | 1.82 ± 1.88 | 1.48 ± 2.60 | 0.015 | 0.99 (0.93–1.05) | 0.986 | ||
Follow-up time (years) | 15.2 ± 2.4 | 14.8 ± 2.4 | 0.561 | 1.03 (0.96–1.09) | 0.433 | ||
Diabetes heredity (%) | 11.5 | 23.5 | <0.001 | 2.05 (1.50–2.80) | <0.001 | ||
Current smoker (%) | 41.0 | 52.0 | <0.001 | 1.87 (1.44–2.42) | <0.001 | ||
Use smokeless tobacco (%) | 2.7 | 2.8 | 0.266 | — | — | ||
Low physical activity (%) | 49.4 | 55.5 | 0.126 | 0.97 (0.75–1.25) | 0.801 | ||
Social class | |||||||
Nonmanual worker (%) | 43.6 | 32.7 | <0.001 | Reference | — | ||
Manual worker (%) | 52.6 | 58.4 | 0.054 | 1.37 (1.04–1.80) | 0.025 | ||
Other classifications (%) | 3.8 | 8.9 | <0.001 | 2.63 (1.58–4.37) | <0.001 |
Data are means ± SD or proportions unless otherwise indicated. Results are from univariate and multivariate logistic regression analysis of predictors. Fasting blood glucose was log transformed in the multivariate analysis due to its skewed distribution. ORs are given for increase per unit. Diastolic blood pressure and use of smokeless tobacco were not included in the multivariate model.
↵* P for trend 0.007;
↵† OR for systolic blood pressure and log fasting blood glucose are given per SD increase, e.g., 13.7 mmHg and 0.107 mmol/l, respectively;
↵‡ OR for heart rate is given per increase of 10 beats/min.