Dose-Dependent Association Between Body Mass Index and Mental Health and Changes Over Time

Abstract

Importance: Overweight and obesity affect 340 million adolescents worldwide and constitute a risk factor for poor mental health. Understanding the association between body mass index (BMI) and mental health in adolescents may help to address rising mental health issues; however, existing studies lack comprehensive evaluations spanning diverse countries and periods.

Objective: To estimate the association between BMI and mental health and examine changes over time from 2002 to 2018.

Design, setting, and participants: This was a repeated multicountry cross-sectional study conducted between 2002 and 2018 and utilizing data from the Health Behaviour in School-aged Children (HBSC) survey in Europe and North America. The study population consisted of more than 1 million adolescents aged 11 to 15 years, with all surveyed children included in the analysis. Data were analyzed from October 2022 to March 2023.

Main outcomes and measures: Mental health difficulties were measured by an 8-item scale for psychological concerns, scoring from 0 to 32, where a higher score reflects greater psychosomatic issues. BMI was calculated using weight divided by height squared and adjusted for age and sex. Data were fitted by multilevel generalized additive model. Confounders included sex, living with parents, sibling presence, academic pressure, the experience of being bullied, family affluence, screen time, and physical activity.

Results: Our analysis of 1 036 869 adolescents surveyed from 2002 to 2018, with a mean (SD) age of 13.55 (1.64) years and comprising 527 585 girls (50.9%), revealed a consistent U-shaped association between BMI and mental health. After accounting for confounders, adolescents with low body mass and overweight or obesity had increased psychosomatic symptoms compared to those with healthy weight (unstandardized β, 0.14; 95% CI, 0.08 to 0.19; unstandardized β, 0.27; 95% CI, 0.24 to 0.30; and unstandardized β, 0.62; 95% CI, 0.56 to 0.67, respectively), while adolescents with underweight had fewer symptoms (unstandardized β, -0.18; 95% CI, -0.22 to -0.15). This association was observed across different years, sex, and grade, indicating a broad relevance to adolescent mental health. Compared to 2002, psychosomatic concerns increased significantly in 2006 (unstandardized β, 0.19; 95% CI, 0.11 to 0.26), 2010 (unstandardized β, 0.14; 95% CI, 0.07 to 0.22), 2014 (unstandardized β, 0.48; 95% CI, 0.40 to 0.56), and 2018 (unstandardized β, 0.82; 95% CI, 0.74 to 0.89). Girls reported significantly higher psychosomatic concerns than boys (unstandardized β, 2.27; 95% CI, 2.25 to 2.30). Compared to primary school, psychosomatic concerns rose significantly in middle school (unstandardized β, 1.15; 95% CI, 1.12 to 1.18) and in high school (unstandardized β, 2.12; 95% CI, 2.09 to 2.15).

Conclusions and relevance: Our study revealed a U-shaped association between adolescent BMI and mental health, which was consistent across sex and grades and became stronger over time. These insights emphasize the need for targeted interventions addressing body image and mental health, and call for further research into underlying mechanisms.

Figure 1.. Generalized Additive Models of Psychosomatic Concerns as a Function of Body Mass Index (BMI) z Score by Survey Year

The gray area presents 95% CIs of fitting on the function of BMI z score. The P value for spline of BMI z score was <.001. The P value was extracted from the multilevel generalized additive model, with psychosomatic concerns as the outcome and smooth term of BMI z score as the exposure, controlling for survey year, sex, grade, living with parents, sibling presence, academic pressure, having experienced bullying, and smooth term of survey year, Family Affluence Scale, screen time, and physical activity, as well as the random intercept and random slope for BMI z score at the level of classroom, school, and country. Psychosomatic concern scores were measured using an 8-item instrument assessing feeling low, irritability, nervousness, sleep difficulties, dizziness, headache, stomachache, and backache with frequency reported on a 5-point scale from daily to rarely/never.

Figure 2.. Generalized Additive Models of Psychosomatic Concerns as a Function of Body Mass Index (BMI) z score, by Survey Year and Sex

The gray area presents 95% CIs of fitting on the function of BMI z score. For boys and girls, both P values for spline of BMI z score were <.001. P values were extracted from the multilevel generalized additive model, with psychosomatic concerns as the outcome and smooth term of BMI z score as the exposure, controlling for survey year, grade, living with parents, sibling presence, academic pressure, having experienced bullying, and smooth term of Family Affluence Scale, screen time, and physical activity, as well as the random intercept and random slope for BMI z score at the level of classroom, school, and country. Psychosomatic concern scores were measured using an 8-item instrument assessing feeling low, irritability, nervousness, sleep difficulties, dizziness, headache, stomachache, and backache with frequency reported on a 5-point scale from daily to rarely/never.

Figure 3.. Generalized Additive Models of Psychosomatic Concerns as a Function of Body Mass Index (BMI) z score, by Survey Year and Grade

The gray area presents 95% CIs of fitting on the function of BMI z score. For primary school, middle school, and high school, all P values for spline of BMI z score were <.001. P values were extracted from the multilevel generalized additive model, with psychosomatic concerns as the outcome and smooth term of BMI z score as the exposure, controlling for survey year, sex, living with parents, sibling presence, academic pressure, having experienced bullying, and smooth term of Family Affluence Scale, screen time, and physical activity, as well as the random intercept and random slope for BMI z score at the level of classroom, school, and country. Psychosomatic concern scores were measured using an 8-item instrument assessing feeling low, irritability, nervousness, sleep difficulties, dizziness, headache, stomachache, and backache with frequency reported on a 5-point scale from daily to rarely/never.