Exposure to light at night (LAN) and risk of obesity: A systematic review and meta-analysis of observational studies

Background: There is emerging evidence of the association between light at night (LAN) exposure and weight gain. Objective: We aim to conduct a systematic review and meta-analysis of observational studies on the association between LAN exposure and risk of obesity in human subjects. Methods: Peer-reviewed observational studies were systematically searched from MEDLINE (EBSCO), Academic Search Complete (EBSCO), CINAHL Plus (EBSCO) and PubMed up to December 24, 2019. Random-effects models were developed to estimate the associations between LAN exposure and weight-related outcomes of overweight and obesity as measured by body mass index (BMI), waist circumference, waist-hip-ratio and waistto-height-ratio. The I-2 statistic was used to assess the degree of heterogeneity across studies. The National Toxicology Program's Office of Health Assessment and Translation (OHAT) risk of bias rating tool and the Grading of Recommendations Assessment, Development and Evaluation (GRADE) guideline were respectively employed to assess the risk of bias and to appraise the quality of the generated evidence. Results: A total of 12 studies (three with longitudinal and nine of cross-sectional design) published between 2003 and 2019 were included for systematic review, while seven of them fulfilling the inclusion/exclusion criteria were included in the meta-analysis. A higher LAN exposure was significantly associated with 13% higher odds of overweight (BMI 25 kg/m(2)) (Summary Odds Ratio; SOR: 1.13, 95% CI: 1.10-1.16) with low heterogeneity (I-2 = 27.27%), and 22% higher odds of obesity (BMI 30 kg/m(2)) (SOR: 1.22, 95% CI: 1.07-1.38) with substantial heterogeneity (I-2 = 85.96%). Stratifying analyses by the levels of measurement of LAN exposures (macro-, meso- and micro-levels) and time of LAN measurement (including before and while sleeping) consistently produced robust estimates, with higher exposure to LAN being positively associated with poorer weight outcomes. Assessment of risk of bias identified substantial detection bias for exposure, with over half of the pooled studies employing subjective LAN measures. The overall evidence of the association between LAN exposure and risk of obesity was rated as 'moderate' as per the GRADE guideline. Conclusions: Exposure to LAN was reported to be a significant risk factor for overweight and obesity. Prospectively designed future studies with objectively measured multi-level LAN exposures and weight outcomes are required.