Physical inactivity is one of the leading causes of morbidity and premature mortality.1, 2 In England, two-thirds of adults do not meet recommended levels of physical activity.3 The mass adoption of private motorised transport has contributed to declines in active travel worldwide. In England and Wales, 23·7 million individuals regularly commute to a workplace, 67% by car.4 Individuals who commute to work actively achieve higher levels of total physical activity than do car users, independent of leisure physical activity.5, 6, 7, 8 Active commuting has been identified by the UK National Institute for Health and Clinical Excellence (now National Institute of Health and Care Excellence) as a feasible way to increase population physical activity without requiring unacceptable financial or time expenditure.9
Middle age is an important life-stage for the development of obesity. In the 2012 Health Survey for England, BMI was found to increase steadily into middle age, before declining beyond age 75 years.10 In the same survey, 44% of adults aged 55–64 years did not meet recommended physical activity levels.11 Using survey data from the UK Household Longitudinal Study, Laverty and colleagues12 reported that after adjustment for socioeconomic and demographic factors, compared with 16–29 year olds, adults aged 50–65 years were 55% less likely to commute by public transport, 45% less likely to commute on foot, and 30% less likely to commute by bicycle.
Previous research has shown that, compared with car commuters, walkers and cyclists have lower BMI,12, 13, 14, 15 lower percentage body fat,13 and lower waist circumference.15 Walkers and cyclists are less likely to report diagnoses of diabetes and hypertension,13, 14, 15 and report higher levels of physical wellbeing than do car commuters.16 Compared with continued car use, switching from sedentary to active commuting was found to predict a decrease in self-reported BMI.17 However, many individuals live too far from their workplace for walking or cycling to be feasible. Cross-sectional studies have shown that individuals who commute by public transport have significantly lower BMI, percentage body fat, and self-reported diagnoses with diabetes and hypertension, and that associations are similar in size to those seen for walking and cycling.12, 13 This finding suggests that incidental physical activity involved in public transport journeys has an important role. Increasing evidence consistently suggests that mass transit contributes to overall physical activity18, 19 and is negatively associated with obesity. 20, 21
Research in context
Evidence before this study
Previous studies have found evidence of an association between active commuting and a range of obesity-related and cardiovascular disease-related health outcomes in the UK. Walking, cycling, and taking public transport to work have been shown to predict self-reported and objectively measured BMI, objectively measured percentage body fat, self-reported doctor-diagnosed diabetes, and self-reported hypertension. However, previous studies have been hampered by insufficiently detailed information about method of commuting exposures (making it largely impossible to capture mixed-mode journeys, for example), self-reported outcome measures, limited scope to control for confounding by health behaviours especially diet, and small sample sizes for meaningful subgroup analyses.
Added value of this study
This study is the first to use UK Biobank data to investigate the relation between commuting behaviour and BMI, based on objectively measured bodyweight and height; and body fat. UK Biobank is a large, observational study of middle-aged individuals in the UK. This dataset offers the opportunity to do a definitive cross-sectional, observational investigation into the relation between active commuting and obesity in terms of sample size, gradation of exposure, robust biological measures, and use of a wide range of covariates. The age span of the participants covers a key life-stage for the development of physical inactivity, obesity, and related cardiovascular risk. UK Biobank allows a substantially more detailed and fine-grained categorisation of method of commute exposure by allowing the derivation of mixed-mode categories. This study also uses objectively measured height, bodyweight, and percentage body fat, so is not subject to bias introduced when respondents are asked to report their own measurements. The comprehensive range of socioeconomic, demographic, behavioural, occupational, and health data available in UK Biobank allowed us to do comprehensive adjustment for confounding. Additionally, we were able to further isolate independent associations between active commuting and obesity in sensitivity analyses controlling for dietary energy intake. To our knowledge, this is the first study of active commuting and obesity to incorporate this dimension.
Implications of all the available evidence
The results of this study add to the growing evidence suggesting that the promotion and facilitation of active commuting should form part of the policy response for population-level obesity prevention.
However, important gaps in the scientific literature remain. First, it is unclear whether active commuting has a graded relationship with obesity. Operationalising the exposure using multiple mixed-mode categories provides better observational evidence of causality, but existing studies are hampered by inadequately detailed measurement. Second, the use of objectively measured obesity is important to reduce the bias associated with self-reporting. A third common limitation of the use of secondary datasets to investigate active commuting and bodyweight is the failure to rule out confounding by energy intake. The extent to which associations between active commuting and obesity are moderated by demographic and socioeconomic factors is also underexplored in the existing scientific literature.
This study is the first to use data from UK Biobank22, 23 to investigate the relation between active commuting and obesity. A large and comprehensive dataset, UK Biobank offers the opportunity to address key limitations in the existing literature. The aim of this study is to investigate the relation between active commuting and objectively measured indicators of obesity in a middle-aged UK population sample. Our hypotheses are that an independent association exists between commuting mode and two measures of obesity (BMI and percentage body fat); adjustment for socioeconomic, demographic, health, and occupational and leisure factors will not fully attenuate the central association; dietary energy intake will have a small but significant confounding effect; among walkers and cyclists, distance of commute will be independently, inversely associated with BMI and percentage body fat.