Obesity represents a severe public health challenge globally. The prevalence of obesity came to a high level recently, exceeding 13% globally, and contributed to a decline in both quality of life and life expectancy (1–3). The Nutrition and Chronic Disease Status of Chinese Residents (2020) estimated that 16.4% of the Chinese adult residents were obese, and obesity prevalence was increasing (4). Obesity has been attributed to behavioral, genetic, socioeconomic, and environmental factors. Furthermore, air pollution has been considered as one of the main environmental causes affecting obesity (5).

Ambient fine particulate matter (PM_2.5) has emerged as a major air pollution globally. A recent cross-sectional research of 2660 children suggested that PM_2.5 was positively related to a high obesity risk (6). Meanwhile, meteorological factors, such as sunlight, are also regarded as novel potential environmental risk factors for obesity. Previous studies on sunlight exposure have shown that sunlight exposure decreased the risk of obesity (7). Recent in vitro and animal experiments have indicated that PM_2.5 and limited sunlight exposure have several physiological effects in common, including systematic inflammation, insulin resistance, and stimulation of the differentiation of pre-adipocytes via reduction of serum vitamin D (8). These effects are all potentially linked with the pathogenesis of obesity. Therefore, it is far more likely that sunlight could modified the contribution of PM_2.5 to obesity. Nevertheless, evidence for the potential impact of PM_2.5 on obesity under different sunlight conditions is limited. Moreover, for countries with relatively high levels of PM_2.5 like China, it is necessary to assess the sunlight effect on the relationship of PM_2.5 with obesity.

Therefore, the research aimed to examine the relationship of PM_2.5 and sunlight exposure with obesity risk among the general population in China using a national representative sample.

Participants Characteristics are presented in Table 1 . Among 47,204 participants, 5,940 (12.6%) and 12,629 (26.8%) were obese and abdominal obese, respectively. Over half of the subjects were women (55.0%), and the mean age was 49.6 ± 15.2 years. Compared with non-obesity participants, those with obesity were more likely to be older, have less education, smoke, drink alcohol frequently.

Estimated PM_2.5 levels ranged from 13.20 μg/m³ to 72.13 μg/m³ for five-year exposure, and the overall mean ambient PM_2.5 concentration in the study population reached 46.62 μg/m³ (SD of 15.51 μg/m³). The overall mean five-year sunshine duration was 6.93 hours (SD of 1.61 hours) per day. The PM_2.5 and sunshine duration levels by each province were diverse ( Table S1 ).

The relationship of PM_2.5 exposure with obesity risk and abdominal obesity risk was significant after adjusting for potential confounding factors ( Table 2 ). A five-year average PM_2.5 level increase of 10 μg/m³ was positively related to obesity risk [OR 1.12 (95% CI, 1.09-1.14)]. A positively significant association was also observed for abdominal obesity [OR 1.10 (95% CI, 1.07-1.13)].

We observed an inverse J-shaped relationship of PM_2.5 level with risk of obesity/abdominal obesity as the sunshine duration increased ( Table 2 ; P value for interaction <0.001), with the lowest risk at 7.18-8.37 h/d. The highest effect estimates of PM_2.5 for obesity (OR 1.56; 95%CI, 1.28-1.91) and abdominal obesity (OR 1.66; 95% CI, 1.34-2.07) were observed in the bottom quartile of sunlight exposure (3.21-5.34 h/d).

Sensitivity analyzes also revealed a positively significant relationship of categorized PM_2.5 exposure with obesity risk (P value for trend <0.001), and the OR increased gradually across the categories, except for abdominal obesity in the highest exposure group ( Supplemental Figure 1 ). The findings were stable after adding other potential confounders to the multivariable models ( Table S2 ). The association of the one-year mean PM_2.5 with the risk of obesity and abdominal obesity showed no substantial change of the risk estimates as to the five-year mean PM_2.5 ( Table S3 ). The association assessed by logistics models was also similar to the association assessed by GAMs ( Table S4 ). The subgroup analyses suggested that the obesity risk attributed to PM_2.5 attenuated as the sunshine duration increased, especially in south areas ( Table S5 ).

We presented the first study to investigate the effect modification of sunlight on the impact of PM_2.5 on obesity risk. An inverse J-shaped relationship of PM_2.5 with obesity and abdominal obesity risk was observed as the sunshine duration increased, with the lowest risk at the middle sunshine duration group (7.18-8.37 h/d) and the highest risk at the bottom quartile (3.21-5.34 h/d).

Most previous studies revealed a positively significant relationship of PM_2.5 level with obesity (6, 19–24). However, several cross-sectional surveys revealed no association between PM_2.5 and obesity (25–27), which could be explained by the relatively low level of PM_2.5 exposure. For instance, in the Framingham Heart Study (27), the annual mean PM_2.5 level reached 10.6 μg/m³, which closely came to the standard from World Health Organization (yearly mean PM_2.5: 10μg/m³). Additionally, several studies reporting no association between PM_2.5 and obesity mainly focused on subjects living in the area with an abundant sun exposure (25, 26). Our study extended previous observation between PM_2.5 and obesity to regions with relatively high level of PM_2.5 exposure. Besides, previous evidence revealed that the risk of abdominal obesity was higher with a high PM_2.5 concentration in rural areas (19), and short-term exposure has been linked to abdominal obesity (23). Our study further identified a positively significant relationship of long-term effect of PM_2.5 with abdominal obesity risk.

We observed that the relationship of PM_2.5 with obesity was influenced by the sunshine duration. Previous studies suggested that sunlight exposure might promote the synthesis of vitamin D and nitric oxide, which in turn could lead to a reduced risk of adiposity accumulation (8, 28). In addition, sunlight exposure may stimulate browning recruitment in white adipose tissue via over-expression of cyclooxygenase-2 (COX-2) (29) and facilitated systemic energy expenditure through a myogenic factor 5 (myf-5) independent pathway (30). A cross-sectional study found that the elevated January sunshine duration was associated with an decreased risk of obesity (7). Furthermore, recent studies using latitude (31) and altitude as substitutes for sunlight exposure indicated that decreased latitude (32) or increased altitude (33) (substituting for elevated sunlight exposure) was related to a reduced risk of obesity. However, excess exposure to ultraviolet radiation could overwhelm the cutaneous antioxidant capacity, leading to inflammation and oxidative stress (8, 34). Therefore, sunshine duration exceeding the appropriate range may associate with an elevated risk of obesity, which is consistent with our results.

Experimental studies suggest possible mechanisms for the effect modification of sunlight on the PM_2.5 in the pathogenesis of obesity. In vitro and in vivo experiments revealed that both of PM_2.5 and sunlight exposure induced the reactive oxygen species (ROS) generation and the expression of COX-2 gene (29, 35, 36), while with various pathways. PM_2.5 could induce generation of the highly reactive hydroxyl radical through catalyzing Fenton’s reaction (37, 38), and could activate COX-2 expression through ROS-nuclear factor kappa B (NF-κB) pathway (39). Meanwhile, sunlight exposure could stimulate ROS production by activating the catalase and promoting nitric oxide synthase synthesis (35), and could activate the expression of COX-2 via protein-tyrosine phosphorylation (40). Both excessive ROS production and COX-2 activation led to the progress of systematic inflammation, insulin resistance, and increased oxidative stress (41–43), contributing to an elevated risk of obesity (44, 45).

The study has strengths that deserve mention. The major strength is that the participants were enrolled from multiple centers with relatively high PM_2.5 levels and the striking latitude gradient. This study has limitations as well. First, our study population was derived from pre-existing cross-sectional study. Causal inferences on effect of PM_2.5 and sunlight on the risk of obesity or abdominal obesity could not be made because this study did not capture the obesity status prior to exposure. Second, we didn’t assess the effects of gaseous pollutants except for NO₂. However, NO₂ was one of the major predictors of health effects and was highly correlated with other gaseous pollutants (46). Third, our assessment of PM_2.5 and sunlight exposure were ascertained based on the nominal levels but not measured levels, which is common in big data surveys. Fourth, we did not collect the information on the workplace, which leads to the effect evaluation of environmental exposure only based on home address. Fifth, we did not take the effect of potential discrepancy in PM_2.5 components as a variable into consideration. Furthermore, the possibility of residual confounding could not be excluded.

In conclusion, the present research reveals that the relationship between PM_2.5 and obesity or abdominal obesity risk varies by sunshine duration, and stronger relationship were observed in short-sunshine-time regions compared to medium- and long-sunshine-time regions. An improved understanding of this interaction effect may offer important insights for lowering obesity risk attributed to environmental factors.

The data analyzed in this study is subject to the following licenses/restrictions: Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data is not available. Requests to access these datasets should be directed to zhanglx@bjmu.edu.cn.

The studies involving human participants were reviewed and approved by the ethics committee at Peking University First Hospital. The patients/participants provided their written informed consent to participate in this study.

RC and LZ contributed to the conception and design of the study. RC and CY contributed to the literature review. RC, CY, PL, JW, and LZ contributed to the data collection and data quality control. ZL, WW, YueW, and CL provided the air pollution exposure data. YanW collected the meteorological data. RC, CY, JW, and LZ cleaned, analysed, and visualized the data. RC, CY, JW, and LZ supervised the analysis and generation of results, and directly accessed and verified the data. RC wrote the manuscript. LZ and CY reviewed and edited the manuscript. All authors contributed to data interpretation, and reviewed and approved the final version manuscript. All authors had full access to all the data in the study and had final responsibility for the decision to submit for publication.

This study was supported by grants from the National Natural Science Foundation of China (91846101, 82003529, 81771938, 81900665, 82090021), Beijing Nova Programme Interdisciplinary Cooperation Project (Z191100001119008), National Key R&D Program of the Ministry of Science and Technology of China (2019YFC2005000), Chinese Scientific and Technical Innovation Project 2030 (2018AAA0102100), the University of Michigan Health System-Peking University Health Science Center Joint Institute for Translational and Clinical Research (BMU2018JI012, BMU2019JI005, 71017Y2027), CAMS Innovation Fund for Medical Sciences (2019-I2M-5-046), and PKU-Baidu Fund (2019BD017, 2020BD032).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.