Do physical activity, commuting mode, cardiorespiratory fitness and sedentary behaviours modify the genetic predisposition to higher BMI? Findings from a UK Biobank study

Abstract

Objective To investigate whether the association between a genetic profile risk score for obesity (GPRS-obesity) (based on 93 SNPs) and body mass index (BMI) was modified by physical activity (PA), cardiorespiratory fitness, commuting mode, walking pace and sedentary behaviours. Methods For the analyses we used cross-sectional baseline data from 310,652 participants in the UK Biobank study. We investigated interaction effects of GPRS-obesity with objectively measured and self-reported PA, cardiorespiratory fitness, commuting mode, walking pace, TV viewing, playing computer games, PC-screen time and total sedentary behaviour on BMI. Body mass index (BMI) was the main outcome measure. Results GPRS-obesity was associated with BMI (beta:0.54 kg.m(-2) per standard deviation (SD) increase in GPRS, [95% CI: 0.53; 0.56]; P = 2.1 x 10(-2)(41)). There was a significant interaction between GPRS-obesity and objectively measured PA (P-[interaction] = 3.3 x 10(-11)): among inactive individuals, BMI was higher by 0.58 kg.m(-2) per SD increase in GPRS-obesity (p = 1.3 x 10(-70)) whereas among active individuals the relevant BMI difference was less (beta:0.33 kg.m(-2), p = 6.4 x 10(-41)). We observed similar patterns for fitness (Unfit beta:0.72 versus Fit (beta:0.36 kg.m(-2), P-[interaction] = 1.4 x 10(-11)), walking pace (Slow beta:0.91 versus Brisk (beta:0.38 kg.m(-2), P-[interaction] = 8.1 x 10(-27)), discretionary sedentary behaviour (High beta:0.64 versus Low (beta:0.48 kg.m(-2), P-[interaction] = 9.1 x 10(-12)), TV viewing (High beta:0.62 versus Low (3:0.47 kg.m(-2), P-[interaction] = 1.7 X 10(-11)), PC-screen time (High beta:0.82 versus Low (beta:0.54 kg.m(2), P-[interaction] = 0.0004) and playing computer games (Often beta:0.69 versus Low (beta:0.52 kg.m (2), P-[interaction] = 8.9 x 10(-10)). No significant interactions were found for commuting mode (car, public transport, active commuters). Conclusions Physical activity, sedentary behaviours and fitness modify the extent to which a set of the most important known adiposity variants affect BMI. This suggests that the adiposity benefits of high PA and low sedentary behaviour may be particularly important in individuals with high genetic risk for obesity.

Objective To investigate whether the association between a genetic profile risk score for obesity (GPRS-obesity) (based on 93 SNPs) and body mass index (BMI) was modified by physical activity (PA), cardiorespiratory fitness, commuting mode, walking pace and sedentary behaviours. Methods For the analyses we used cross-sectional baseline data from 310,652 participants in the UK Biobank study. We investigated interaction effects of GPRS-obesity with objectively measured and self-reported PA, cardiorespiratory fitness, commuting mode, walking pace, TV viewing, playing computer games, PC-screen time and total sedentary behaviour on BMI. Body mass index (BMI) was the main outcome measure. Results GPRS-obesity was associated with BMI (beta:0.54 kg.m(-2) per standard deviation (SD) increase in GPRS, [95% CI: 0.53; 0.56]; P = 2.1 x 10(-2)(41)). There was a significant interaction between GPRS-obesity and objectively measured PA (P-[interaction] = 3.3 x 10(-11)): among inactive individuals, BMI was higher by 0.58 kg.m(-2) per SD increase in GPRS-obesity (p = 1.3 x 10(-70)) whereas among active individuals the relevant BMI difference was less (beta:0.33 kg.m(-2), p = 6.4 x 10(-41)). We observed similar patterns for fitness (Unfit beta:0.72 versus Fit (beta:0.36 kg.m(-2), P-[interaction] = 1.4 x 10(-11)), walking pace (Slow beta:0.91 versus Brisk (beta:0.38 kg.m(-2), P-[interaction] = 8.1 x 10(-27)), discretionary sedentary behaviour (High beta:0.64 versus Low (beta:0.48 kg.m(-2), P-[interaction] = 9.1 x 10(-12)), TV viewing (High beta:0.62 versus Low (3:0.47 kg.m(-2), P-[interaction] = 1.7 X 10(-11)), PC-screen time (High beta:0.82 versus Low (beta:0.54 kg.m(2), P-[interaction] = 0.0004) and playing computer games (Often beta:0.69 versus Low (beta:0.52 kg.m (2), P-[interaction] = 8.9 x 10(-10)). No significant interactions were found for commuting mode (car, public transport, active commuters). Conclusions Physical activity, sedentary behaviours and fitness modify the extent to which a set of the most important known adiposity variants affect BMI. This suggests that the adiposity benefits of high PA and low sedentary behaviour may be particularly important in individuals with high genetic risk for obesity.