LAUSR

We read the recently published article by Pan et al. with great interest. This trial reviewed the effect of bike lanes access on childhood obesity and indicated that bike lane access was significantly associated with children and adolescents' physical activities (PAs). However, we feel that some issues need to be clarified and improved. First, although this study included 21 papers in the systematic review, only three trials were enrolled in the meta-analysis. In addition, as a timely study, we noticed that the included studies were published 10 years ago. Limited by the sample size and timeliness of enrolled study, we are concerned about the reliability of the results and the strength of the evidence. Moreover, included studies reported PA by questionnaires, which was prone to recall bias. Second, the characterized outcomes were poor. Authors tried to demonstrate the relationship between bike lanes and obesity, but we found that the three studies included for meta-analysis mainly explored related factors affecting PA, including the provision of recreational spaces and distance. However, recreational spaces were not limited to bike lanes, but also included yard, courtyard, and school playground and so forth. In addition, none of three included studies evaluated obesity, such as BMI. In other words, all three articles lack evidence to support the association between bicycle lane access and childhood obesity. As it is known, obesity is a complex disease that could be influenced by many factors, and PA is just one of them; other factors such as psychology and diet also play an important role in obesity. The influence of bike lanes access on obesity might be negligible. Norah et al. reported that the reason for inactive commuting to school for childhood is mainly the distance and time, not the bike lane access. According to subgroup analysis of Chomitz's study, the association between recreational spaces and PA depended on the intensity of PA. Only when the intensity was moderate, there was a close relationship between bike lanes and PA. When the intensity was vigorous and 60+ min PA, there was no obvious correlation between them. Third, children's growth and development are influenced by gender and age, especially adolescence. The participants for meta-analysis included two group students (aged 9–14 and aged 15–17). The behaviors of children in different adolescence are influenced by many factors, including personal factors, environmental factors, and macrosystem factors. James et al. established a multilevel model and elaborated factors that affected active living in detail. According to this model, PA behaviors such as cycling for transport were determined by many factors, including environment and individual characteristics (e.g., gender and self-efficacy). We also noticed that obesity is related to gender, which was basically consistent with the results presented by the author. Thus, gender differences should be considered, especially when the results of research reports are inconsistent between men and women. To summary, the authors should be applauded for addressing an interesting topic. However, we consider that this conclusion needs careful discussion and further validation. The longitudinal studies are needed to further support the conclusion.