Bidirectional association between sleep duration and obesity was found among junior high school students, with higher BMI predicting shorter subsequent sleep and shorter sleep predicting higher subsequent BMI.
Key Findings
Results
Higher BMI Z-score at baseline was significantly negatively associated with sleep duration one year later in the overall sample.
Cross-lagged panel analysis showed BMIZ in 2021 (T1) was negatively correlated with sleep duration in 2022 (T2): β=-0.030, P<0.05
This indicates that students with higher obesity measures tended to sleep less in the following year
Sample included 3,134 junior high school students with mean baseline age of (12.49±0.40) years
52.04% were boys, 56.96% were from urban areas, 78.14% were Han ethnicity
Results
Shorter sleep duration at baseline was significantly negatively associated with higher BMI Z-score one year later in the overall sample.
Sleep duration in 2021 (T1) was negatively correlated with BMIZ in 2022 (T2): β=-0.016, P<0.05
This indicates that students sleeping less tended to have higher obesity measures the following year
The bidirectional relationship was observed across both directions in the same one-year interval (2021–2022)
Analysis used cross-lagged panel modeling to assess temporal directionality
Results
The negative association between prior-year BMI Z-score and subsequent sleep duration was found in boys but not in girls.
In boys, BMIZ in 2021 was negatively associated with sleep duration in 2022: β=-0.040, P<0.05
No significant association between BMIZ and subsequent sleep duration was found in girls
Boys comprised 52.04% of the analytic sample (approximately 1,630 students)
Gender-stratified cross-lagged panel analysis was conducted separately
Results
Among rural students, bidirectional associations between sleep duration and BMIZ were observed, and sedentary time was positively associated with subsequent BMIZ.
BMIZ in 2021 was negatively associated with sleep duration in 2022 among rural students: β=-0.043, P<0.05
Sleep duration in 2021 was negatively associated with BMIZ in 2022 among rural students: β=-0.021, P<0.05
Sedentary duration in 2022 (T2) was positively associated with BMIZ in 2023 (T3) among rural students: β=0.008, P<0.05
Rural students comprised approximately 43.04% of the sample
Results
Among urban students, higher BMI Z-score at baseline was positively associated with greater sedentary duration one year later, but no sleep-obesity associations were detected.
BMIZ in 2021 was positively associated with sedentary duration in 2022 among urban students: β=0.140, P<0.05
No significant association between sleep duration and BMIZ was found in urban students
Urban students comprised 56.96% of the sample
This suggests that obese urban students may increase sedentary behavior over time
Results
Among Han ethnic students, bidirectional sleep-obesity associations were observed across two consecutive time intervals.
BMIZ in 2021 negatively associated with sleep duration in 2022: β=-0.041, P<0.05
Sleep duration in 2021 negatively associated with BMIZ in 2022: β=-0.023, P<0.05
Sleep duration in 2022 (T2) was also negatively associated with BMIZ in 2023 (T3): β=-0.021, P<0.05
Han ethnic students comprised 78.14% of the analytic sample (approximately 2,449 students)
Methods
The study used a longitudinal cohort design with three annual waves of data collection from junior high school students in Hunan Province, China.
Baseline survey conducted in 2021 with 3,197 students; 3,134 were included in the final analysis
Annual follow-up surveys conducted in 2022 and 2023
Data from the Surveillance Project of Common Diseases and Health Influencing Factors in Students in Hunan Province
BMI Z-score (BMIZ), sleep duration, and sedentary duration were collected at each wave
Cross-lagged panel analysis was the primary statistical method, with stratification by gender, place of residence, and ethnicity
What This Means
This research suggests that among Chinese middle school students (average age ~12.5 years), the relationship between sleep and obesity runs in both directions over time. Students who were heavier (higher BMI) tended to sleep less in the following year, and students who slept less tended to become heavier in the following year. This two-way relationship was identified by following over 3,100 students annually from 2021 to 2023 and using a statistical method called cross-lagged panel analysis that can tease apart which factor comes first.
The findings varied by subgroup. The bidirectional sleep-obesity link was found in boys but not girls, and was consistently observed among Han ethnic students across two consecutive years. In rural areas, more sedentary time also predicted higher BMI the following year. Interestingly, in urban areas, heavier students tended to become more sedentary over the next year, but sleep duration was not significantly linked to obesity changes in that group.
This research suggests that interventions targeting childhood obesity may benefit from addressing both sleep and sedentary behavior, and that tailoring programs by gender, urban/rural setting, and ethnicity could improve their effectiveness. Because the association goes both ways — poor sleep contributing to weight gain and higher weight contributing to poorer sleep — breaking this cycle may require addressing both issues simultaneously rather than focusing on just one.
Qin L, Wang F, Luo J, Li K, Wan J, Luo M, et al.. (2026). [Bidirectional association of sleep duration and sedentary time with obesity risk among junior high school students].. Zhonghua liu xing bing xue za zhi = Zhonghua liuxingbingxue zazhi. https://doi.org/10.3760/cma.j.cn112338-20250705-00463