Table 4

Multivariable model for WC in women (includes all sedentary and physical activity categories that correlated with WC, as well as all established main confounders)

Model (R2=0.14)βSER2p Value
Age (years)0.390.090.027<0.001
Genetic obesity risk score (range 16.1–39.8)2.20.830.0100.01
TV viewing (h/day)1.810.440.022<0.001
Computer use (h/day)1.000.330.012<0.01
Other relaxation (h/day)−0.020.470.00010.96
Listening to radio/music (h/day)−0.050.620.00010.93
Leisure-time exercise (MET h/week)−0.110.030.019<0.01
Strenuous gardening and related housework (h/month)−0.010.040.00010.90
Occupational physical workload (scale 1–6)0.610.340.0040.07
Energy intake (100 kcal/day)−0.0070.020.00020.72
Energy intake*energy intake0.000040.000030.0020.19
Diet score (scale, range 1–27)0.480.700.00060.49
Diet score*diet score−0.020.020.00080.45
Sleep duration (scale, range 1–10)−2.531.380.0050.07
Sleep duration*sleep duration0.200.130.0030.13
Socioeconomic status (scale, 1–3)−0.930.560.0040.09
Alcohol (drinks/day)0.140.580.00010.80
Smoking (no/yes)0.021.250.00010.98
  • β=increase in WC caused by a one unit increase in the measured parameter.

  • R2=per cent that model or measured parameter accounts for the variance of WC.

  • MET index=leisure-time exercise+active commuting.

  • *The increase in β with one unit increase in energy intake, diet score and sleep duration is expressed in their quadratic terms, since these variables had a significant U-shaped relation with BMI and WC.

  • BMI, body mass index; MET, metabolic equivalent; diet score, adherence to the recommended diet; WC, waist circumference.