Table 1—

Characteristics of study participants at baseline according to whether type 2 diabetes developed after the 4-year follow-up and multivariate model of the incidence of type 2 diabetes

TotalType 2 diabetes status after follow-up
Multivariate model*
Not type 2 diabetesType 2 diabetesMultiple-adjusted OR (95% CI)P
n8,5767,698878
Walk to work
    0–10 min1,698 (19.8)1,508 (19.6)190 (21.6)1.00
    11–20 min4,491 (52.4)4,028 (52.3)463 (52.7)0.86 (0.70–1.06)0.158
    ≥21 min2,387 (27.8)2,162 (28.1)225 (25.6)0.73 (0.58–0.92)0.007
Age (years)47.8 ± 4.247.7 ± 4.148.7 ± 4.1
Age categories
    40–442,163 (25.2)1,993 (25.9)170 (19.4)1.00
    45–493,055 (35.6)2,776 (36.1)279 (31.8)1.02 (0.81–1.27)0.899
    50–553,358 (39.2)2,929 (38.0)429 (48.9)1.37 (1.11–1.70)0.003
BMI (kg/m2)23.4 ± 2.923.3 ± 2.824.6 ± 3.21.11 (1.08–1.14)<0.001
FPG (mg/dl)97.6 ± 9.596.3 ± 8.5109.1 ± 9.8
FPG categories (mg/dl)
    <1005,348 (62.4)5,199 (67.5)149 (17.0)1.00
    100–1092,235 (26.1)1,969 (25.6)266 (30.3)4.51 (3.65–5.56)<0.001
    110–125993 (11.6)530 (6.9)463 (52.7)28.01 (22.65–34.63)<0.001
Daily alcohol consumption (g ethanol)25.2 ± 21.925.1 ± 21.925.8 ± 22.3
Drinking habit categories (g ethanol)
    Nondrinkers1,334 (15.6)1,192 (15.5)142 (16.2)1.00
    Tertile 1 (0.1–16.3)1,657 (19.3)1,493 (19.4)164 (18.7)0.94 (0.71–1.23)0.636
    Tertile 2 (16.4–42.6)3,002 (35.0)2,711 (35.2)291 (33.1)0.75 (0.59–0.95)0.018
    Tertile 3 (42.7–115.0)2,583 (30.1)2,302 (29.9)281 (32.0)0.83 (0.65–1.06)0.129
Smoking habit
    Lifelong nonsmokers1,841 (21.5)1,662 (21.6)179 (20.4)1.00
    Past smokers1,924 (22.4)1,710 (22.2)214 (24.4)1.07 (0.84–1.36)0.600
    Current smokers4,811 (56.1)4,326 (56.2)485 (55.2)1.36 (1.11–1.68)0.004
Regular leisure-time physical exercise
    No7,028 (81.9)6,296 (81.8)732 (83.4)1.00
    Yes1,548 (18.1)1,402 (18.2)146 (16.6)0.90 (0.72–1.11)0.317
Parental history of diabetes
    No7,472 (87.1)6,778 (88.0)694 (79.0)1.00
    Yes1,104 (12.9)920 (12.0)184 (21.0)1.65 (1.34–2.02)<0.001
  • Data are means ± SD, n (%), or multiple-adjusted ORs (95% CI).

  • *

    * In multiple logistic regression analysis, age, fasting plasma glucose level, and daily alcohol consumption showed a nonlinear association with the incidence of type 2 diabetes. Therefore, we fit the model by using these variables categorized for easy understanding.

  • As only 5.5% of men reported a walk of ≥31 min, we combined them into the 21- to 30-min walk category.