This is a cross-sectional study. This study is part of a behavior and subhealth study of Yunnan adolescents that aims to explore the association between lifestyle factors and physical and mental subhealth among adolescents in Yunnan, China, through a series of cohort studies [34].

From October to December 2022, a sample questionnaire survey was conducted with first-year (grade 7) middle school students from 11 counties in Yunnan Province, China. Yunnan Province lies along China’s southwestern border with Myanmar, Laos, and Vietnam. Yunnan has 26 ethnic groups and 25 ethnic minority groups. A total of 8635 multiethnic adolescents were surveyed, and 8515 questionnaires were completed, for a 98.6% response rate. There were 8500 valid questionnaires; the effective response rate was 99.8%. The average age of participants was 12.64 (SD 0.63; range 12-15) years; 4184 (49.2%) were males and 4136 (48.8%) were females; and 6358 (74.8%) were from rural areas and 2142 (25.2%) were from urban areas. There were 3856 (45.3%) Han and 4659 (54.7%) ethnic minority participants. Figure 1 presents the flowchart of participants. Tables 1 and 2 present the distributions of the demographic variables.

Cluster random sampling was used to conduct a questionnaire survey among adolescents in 11 counties of Yunnan Province, and 1-2 middle schools were selected from each county. SPSS software (version 23.0; IBM Corp) was used for random sampling. All students in grade 7 from the sampled middle schools were included in the survey.

The survey included demographic variables, eating behaviors, anxiety symptoms, and confounding factors that may influence depressive symptoms in adolescents, such as family changes, hospital experiences, screen time, smoking, drinking behavior, physical activity, family history of depression, learning stress, and the impact of COVID-19. The selection of these confounding factors was based on references and previous research findings of the research group.

A pilot test was conducted on 131 middle school students before the field survey to assess participant understanding of the questionnaire. The pilot test results showed that middle school students were able to fully understand the content of the questionnaire, and the average time taken to complete the questionnaire was 16 minutes. The adolescents gathered in a classroom to complete the survey. Trained investigators were on-site to answer questions from adolescents face-to-face. Adolescents completed the paper version of the questionnaire independently, which takes approximately 15-20 minutes.

We used EpiData (version 3.0; EpiData Association) to build the database, and SPSS software (version 23.0; IBM Inc) and R software (version 4.3.1; R Foundation for Statistical Computing) were used for statistical analysis. In the SPSS software, descriptive statistics, a chi-square test, and a generalized linear model (GLM) analysis were performed. Restricted cubic spline (RCS) analysis was performed using R software. A chi-square test was used to compare the difference in depressive symptoms detection rate among middle school students with different demographic characteristics. A GLM and RCS were used to analyze the association between UPF consumption and depressive symptoms. In the model, the UPF consumption frequency (independent variable) and depressive symptom score (dependent variable) were continuous variables. An association analysis was performed after adjusting for demographic and confounding variables that might have influenced depressive symptoms. The association between UPF consumption and depressive symptoms was analyzed using the RCS, and an RCS correlation plot was fitted. A GLM was used to analyze the association between the interaction term of sex, UPF consumption, and depressive symptoms. The test standard was α=.05.

This study was approved by the Ethics Committee of the School of Medicine, Kunming University (approval 20210222), and was carried out in accordance with the Declaration of Helsinki. Ethical guidelines were followed during all the stages of this study. The survey was conducted anonymously with informed consent from the schools, guardians, and students. All participants joined the study voluntarily, and no compensation was provided to them. All data were kept confidential, deidentified, and anonymous.

The total detection rate of depressive symptoms was 28.3% (2402/8500) among middle school students, 25.5% (1069/4184) among males, and 30.9% (1333/4316) among females; the difference was statistically significant (χ²₁=29.83; P<.001). The detection rate of depressive symptoms in 12-year-olds was 26.7% (987/3698), and that in 13- to 15-year-olds was 29.5% (1415/4802); the difference was statistically significant (χ²₁=7.95; P=.005). The stratified analysis by age (12 years and 13-15 years) showed statistically significant differences in the detection rate of depressive symptoms among 12-year-olds for all demographic variables except parental education level and maternal occupation. Among adolescents aged 13-15 years, there were statistically significant differences in the detection rates of depressive symptoms for other demographic variables except for only child, residence, and parental occupation (P<.05).

The consumption rates (frequency ≥1) of staple processed food, processed meat, processed eggs, preserved fruits, instant foods, salted snacks, sweet snacks, carbonated beverages, juice beverages, energy beverages, milk beverages and other SSB for adolescents are presented in Table 3.

Table 4 presents the results of the GLM analysis of the association between UPF consumption and depressive symptoms in adolescents after adjusting for variables (Model 2). Table S1 in Multimedia Appendix 1 shows the analysis results without adjusting for variables (Model 1). After adjusting for confounding variables, instant food (β=.22, 95% CI 0.13-0.32; P<.001), carbonated beverage (β=.20, 95% CI 0.12-0.29; P<.001), energy beverage (β=.21, 95% CI 0.10-0.32; P<.001), and other SSB (β=.24, 95% CI 0.14-0.34; P<.001) consumption were associated with depressive symptoms for all participants. The sex-stratified analysis showed that instant food (β=.15, 95% CI 0.03-0.27; P=.01), carbonated beverage (β=.15, 95% CI 0.04-0.26; P=.006), energy beverage (β=.22, 95% CI 0.08-0.36; P=.002), milk beverage (β=–0.14, 95% CI –0.26 to –0.02; P=.03), and other SSB (β=.21, 95% CI 0.08-0.35; P=.002) consumption was associated with depressive symptoms in males. In addition, instant food (β=.29, 95% CI 0.15-0.43; P<.001), carbonated beverage (β=.29, 95% CI 0.16-0.43; P<.001), energy beverage (β=.22, 95% CI 0.04-0.40; P=.01), and other SSB (β=.25, 95% CI 0.10-0.41; P=.001) consumption were associated with depressive symptoms in females.

Tables 5 and 6 present the association between the interaction of sex, UPF consumption, and depressive symptoms in adolescents. The results showed that female×instant foods (β=.31, 95% CI 0.18-0.44; P<.001), female×carbonated beverages (β=.27, 95% CI 0.15-0.39; P<.001), female×energy beverages (β=.18, 95% CI 0.01-0.35; P=.04), and female×other SSB (β=.17, 95% CI 0.02-0.32; P=.03) were associated with depressive symptoms, respectively. However, no association was observed between the interaction of these UPFs and male and depressive symptoms.

Figure 2 displays the results of the RCS analysis for the association between UPF consumption and depressive symptoms in adolescents. Overall, UPF consumption was significantly associated with depressive symptoms in adolescents, and females were more sensitive to UPF consumption (exposure) than males.

In this study, the detection rate for depressive symptoms among adolescents of multiple ethnic groups in Yunnan was higher (2402/8500, 28.3%), which is consistent with the results of our previous study on adolescents in 4 cities in China (27.3%) [13]. This detection rate was higher than that in a meta-analysis of middle school students in mainland China (24.3%) and that of middle school students in eastern China (23.3%) [38,39]; however, it was lower than that for Bangladeshi adolescents (36.6%) [40] and Indian adolescents (52.3%) [10]. Our data showed the detection rate of depressive symptoms for adolescents aged 13-15 years (1415/4802, 29.5%) was higher than that for adolescents aged 12 years (987/3698, 26.7%). A school-based survey in the United Kingdom determined that the detection rate of depressive symptoms among 15- to 16-year-olds (32%) was higher than that among 11- to 12-year-olds (24%) [9]. In addition, we found the detection rate of depressive symptoms in females (1333/4316, 30.9%) was significantly higher than that for males (1069/4184, 25.5%), which is consistent with the conclusions of other studies [40]. For example, a Bangladeshi study reported a significantly higher prevalence of depressive symptoms among female adolescents than male adolescents (32.5% vs 25.3%) [40]. The higher prevalence of depressive symptoms among female adolescents is consistent worldwide. Numerous studies have suggested that depressive symptoms increase during adolescence, adolescent females are more susceptible to depressive symptoms, and sex differences appear only during adolescence [41]. A meta-analysis found that sex differences peaked in adolescence (OR 3.02, 95% CI 2.76-3.30 for ages 13-15 years), then declined and remained stable in adulthood [42]. Moreover, the longitudinal study of Australian children also found that between the ages of 4 and 14 years, females were more likely than males to be on a trajectory for increased depressive symptoms [43]. Therefore, schools and families need to pay more attention to depressive symptoms in adolescents, especially female adolescents.

Our study suggests that fast foods, carbonated beverages, energy beverages, and other SSB (except those listed in this study) were positively associated with depressive symptoms in both males and females. Other studies have also determined that UPF consumption is associated with depressive symptoms [25,29,31]. Our survey of Chinese college freshmen demonstrated that carbonated beverage and milk tea consumption were associated with depressive symptoms in males, whereas carbonated beverages, tea beverages, and milk tea consumption were associated with depressive symptoms in females [44]. Higher UPF consumption in Spanish adolescents aged 14-17 years was associated with a higher detection rate of depressive symptoms, and the higher the intake, the more pronounced the depressive symptoms [22]. The consumption of sweets, biscuits, snacks, and fast food among adolescents in England, Wales, and Northern Ireland was significantly positively associated with depressive symptoms [30]. Instant noodles, instant rice, and other instant foods are very popular among Chinese adolescents because they are easy to carry and eat. Our survey indicated that the consumption rate of instant foods such as instant noodles was as high as 61.6% (5234/8500), which has not been previously reported in China. Our previous research found an association between carbonated beverage and processed snack consumption and anxiety symptoms in Yunnan college students [45]. This was the first study to report a correlation between instant food consumption and depressive symptoms among Chinese multiethnic adolescents.

Furthermore, the data from this study showed that milk beverage consumption was negatively associated with depressive symptoms in males. The milk beverages in this study were also SSB; thus, we will further observe and evaluate these associations in a cohort study. The RCS analysis revealed that instant food, carbonated beverages, energy beverages, and other SSB were associated with depressive symptoms in males and females, and the association between UPF consumption and depressive symptoms in females was more sensitive. We further analyzed the interaction between sex and UPF and its association with depressive symptoms. The results indicated that females who interact with instant foods, carbonated beverages, energy beverages, and other SSBs were associated with depressive symptoms, respectively. However, no association was observed between the interaction of males and these 4 types of UPFs and depressive symptoms. The research data further support the view that females who consumed UPF were more likely to experience depressive symptoms. Therefore, when schools develop intervention measures to reduce depression symptoms among adolescents through dietary behavior intervention, they need to take into account sex-specific factors. For female adolescents, it is even more necessary to strengthen health education and behavioral intervention.

Several mechanisms by which UPF consumption is associated with depressive symptoms are possible. First, UPF consumption is associated with sleep dissatisfaction in adolescents [28]. The low sleep quality is a predictor of depressive symptoms in adolescents. Previous studies have suggested that the fourth quartile of UPF (44.3%-81.8% of total calories; prevalence ratio=1.14) was associated with a higher prevalence of poor sleep quality in adolescents [46]. A systematic review and meta-analysis showed that the higher the intake of UPF, the greater the risk of insomnia among adolescents (OR 5, 95% CI 1.21-1.99), and there is a clear dose-response relationship [47]. Second, UPF consumption may lead to disturbances in the intestinal flora and increase stress susceptibility of adolescents through the hypothalamic-pituitary-adrenal axis, which can lead to depressive symptoms occurring [5]. Experimental studies have shown that excessive intake of UPF leads to a reduced abundance of potentially beneficial butyrogenic microbes (Roseburia, Faecalbacterium prausnitzii, Lactobacillus, Bifidobacterium, and Lachnospiraceae), and the abundance of potentially pathogenic or opportunistic microorganisms (Alistiopsis and Pasteurella) increases [48]. Finally, the intake of UPF leads to an increase in chronic inflammatory levels, which causes the occurrence of depressive symptoms [49]. Adolescents with UPF energy consumption ≥30% (tertile 3 of UPF) had a 79% increase in interleukin-8 levels and had higher C-reactive protein levels when compared with adolescents in tertile 1 of UPF [50]. UPF intake may lead to a dysregulation of immune cell homeostasis and inflammation, including decreased levels of anti-inflammatory cytokines and increased levels of proinflammatory cytokines and chemokines [51].

Adolescents have great potential for physical, cognitive, and social development, and adolescence is regarded as the second critical window after infancy [52]. The mental health and growth of adolescents are related to the future of a country and require extensive attention. Adolescent depression is a global concern. Adolescence is a period of rapid emotional and cognitive development and critical life transitions. Notably, the treatment of depression is expensive. A Netherlands study of 11- to 18-year-olds with subclinical depressive symptoms showed that the social costs associated with depressive symptoms amounted to €54 million (US $57 million) per year [53]. Therefore, prevention and early intervention of adolescent depression is a top priority, using strategies that target predisposing factors, causes, and depression symptoms. Modifiable risk factors are critical for reducing depression prevalence among adolescents. Eating behavior interventions have been recommended as an effective measure to prevent mental illness [31]. Therefore, it may be necessary to expand the monitoring of UPF consumption behaviors and depressive symptoms in adolescents, strengthen health education in schools and families, and promote healthy eating environments in schools to reduce UPF consumption and improve depressive symptoms in adolescents.

This study has several strengths. First, the sample comprised adolescents from 11 counties in Yunnan Province, China, with good representativeness and extrapolation of research results. Second, this study was the first to report the association between UPF consumption and depressive symptoms in Chinese multiethnic adolescents. Finally, we used a variety of models to analyze the association between UPF consumption and depressive symptoms and obtained stable results.

This study also has a few limitations. First, information bias was inevitable in the investigation, such as recall and social expectation biases. Second, the current assessment of UPF intake relies on frequency data without quantification in terms of grams or caloric contribution, which is limited for an accurate analysis of the association between the precise intake of UPF and depressive symptoms in adolescents. Finally, cross-sectional data cannot determine causal associations, and bidirectional associations may need to be clarified in further cohort studies.

This study found that the consumption of instant foods, carbonated beverages, and energy beverages was related to depressive symptoms among Chinese multiethnic adolescents. It is important to note that the observed association may have a reverse correlation, where depressive symptoms could lead to an increase in the intake of ultrapalmitic fatty acids among adolescents. This needs to be further confirmed in cohort studies. The results highlighted the importance of reducing UPF consumption among adolescents. These findings should be replicated in cohort studies to determine the causal relationship and the direction. The government, schools, and families should take effective measures to reduce UPF consumption among Chinese multiethnic adolescents to support mental health development in adolescence.