Table 3

Adjusted associations of BMI and activity (occupational/ leisure) at 36, 43, and 53 years with kOA in men
36 years 43 years 53 years
OR p-value OR p-value OR p-value
Manual occupation 0.98 (0.55, 1.76) 0.955 1.27 (0.74, 2.17) 0.386 2.03 (1.19, 3.49) 0.010
Non-manual occupation Referent Referent Referent
BMI (per z-score) 1.22 (0.95, 1.56) 0.117 1.47 (1.14, 1.89) 0.003 1.44 (1.15, 1.82) 0.002
 p-value for test of interaction 0.350 0.232 0.758
Lifting unlikely 1.15 (0.62, 2.17) 0.654 0.71 (0.41, 1.24) 0.226 0.57 (0.33-0.99) 0.044
Lifting somewhat likely Referent Referent Referent
Lifting highly likely 1.18 (0.57, 2.47) 0.659 0.62 (0.30, 1.27) 0.190 0.83 (0.43-1.60) 0.572
BMI (per z-score) 1.20 (0.93, 1.54) 0.153 1.41 (1.09, 1.81) 0.008 1.40 (1.10-1.77) 0.006
 p-value for test of interaction 0.133 0.011 0.745
Kneeling unlikely 1.17 (0.63, 2.15) 0.621 0.96 (0.54, 1.70) 0.887 0.60 (0.34-1.05) 0.072
Kneeling somewhat likely Referent Referent Referent
Kneeling highly likely 0.93 (0.42, 2.07) 0.868 0.58 (0.26, 1.29) 0.180 0.76 (0.37-1.54) 0.446
BMI (per z-score) 1.19 (0.93, 1.53) 0.163 1.41 (1.09, 1.82) 0.008 1.39 (1.10-1.76) 0.006
 p-value for test of interaction 0.557 0.348 0.807
Sitting unlikely 0.92 (0.47, 1.77) 0.792 0.88 (0.49, 1.58) 0.670 1.00 (0.56-1.78) 0.999
Sitting somewhat likely Referent Referent Referent
Sitting highly likely 1.13 (0.61, 2.06) 0.700 0.69 (0.39, 1.24) 0.226 0.60 (0.34-1.07) 0.085
BMI (per z-score) 1.19 (0.93, 1.53) 0.164 1.41 (1.09, 1.81) 0.008 1.42 (1.12-1.78) 0.004
 p-value for test of interaction 0.171 0.278 0.874
Inactive 1.47 (0.76, 2.86) 0.255 1.07 (0.62, 1.85) 0.800 1.77 (0.90-3.48) 0.095
Less active Referent Referent Referent
Most active 1.59 (0.85, 2.96) 0.145 0.78 (0.42, 1.47) 0.447 1.44 (0.71-2.88) 0.309
BMI (per z-score) 1.31 (1.03, 1.66) 0.027 1.46 (1.16, 1.84) 0.001 1.42 (1.14-1.77) 0.002
 p-value for test of interaction 0.307 0.284 0.476

OR Odds Ratio, BMI Body Mass Index.

Note: separate models were examined for each activity exposure and BMI and coefficients reported are from models without a BMIxActivity interaction, adjusting for potential confounders - education at 26y, income at 53y, childhood social class, family history of arthritis, history of knee injury and health status at 53y. Where there was evidence for an interaction, Figures 1A and 1B show the full stratified odds ratios for BMI and activity in these models.

Martin et al.

Martin et al. BMC Musculoskeletal Disorders 2013 14:219   doi:10.1186/1471-2474-14-219

Open Data