1 INTRODUCTION

Obesity is a global public health challenge.1 In 2016, over 340 million children and adolescents aged 5–19 years worldwide with overweight or obese.2 In China, the prevalence of overweight and obese children and adolescents aged 7–18 years was 20.5% in 2014.3 Obesity in childhood and adolescence is associated with adverse health consequences throughout the life course.4, 5 Many efforts have been made in the past6-8 to curb the global obesity epidemic.

However, since December 2019, the coronavirus disease 2019 (COVID-19) has become a global pandemic. To contain the spread of COVID-19, many governments have undertaken uncompromising measures, including home quarantine, lockdown, and school closures. In Shanghai, China, after experiencing an extended winter vacation until the end of February 2020, all of the primary and secondary school students began to attend classes online in March and did not start returning to school until mid-May. As a direct consequence, children's lifestyles, such as physical activity behaviours, screen behaviours, and dietary behaviours, might have been negatively affected,9-11 and this has been associated with childhood obesity.12

Some studies have investigated the impact of the COVID-19 lockdown on lifestyle behaviours and obesity status among children.13-15 An Italian study13 determined that children with obesity had decreased their exercise time and increased their screen and sleeping time during the COVID-19 lockdown. A Chinese study of youths found that, before and after COVID-19 lockdown, the prevalence of overweight/obesity and obesity increased. In addition, significant decreases were seen in the frequency of engaging in physical activity, while significant increases were observed in the average sedentary time, sleeping time, and screen time.14 A French study that included 6491 children also showed the deleterious effects of confinement caused by the lockdown on physical activity and sedentary behaviours.15

Now the COVID lockdown has been cancelled in most countries, but the COVID-19 pandemic is still ongoing. In the first year of the COVID-19 pandemic, children and adolescents in China have gone through home isolation and online classes, and then gradually returned to normal study life. From the end of 2019 to the end of 2020, that is, before the outbreak of the COVID-19 pandemic and 1 year after the outbreak of the COVID-19 pandemic, what changes have taken place in the obesity and obesity-related lifestyle behaviours of children and adolescents? Therefore, this study aims at analysing the changes in obesity and lifestyle behaviours of Chinese adolescents before and 1 year after the outbreak of the COVID-19 pandemic, providing evidence for the global strategies and policies to respond to the impact of the COVID-19 pandemic on adolescent obesity.

2 METHOD 2.1 Study design and participants

Data were collected from the Surveillance of Students' Common Diseases and Health Influencing Factors conducted in Shanghai, China. The surveillance was conducted in all 16 districts of Shanghai during September and November of 2019 and 2020. According to the level of social and economic development, 16 districts were divided into seven urban areas (district of Huangpu, Xuhui, Jing'an, Changning, Putuo, Hongkou, and Yangpu) and nine suburbs (district of Pudong, Jinshan, Minhang, Fengxian, Songjiang, Qingpu, Jiading, Baoshan, and Chongming).

A multi-stage stratified cluster sampling method was used, and two junior high schools from each district were randomly selected, with a total of 32 junior high schools selected. Two to three classes from each grade (grades 6–9) of each school were randomly selected, and all of the students in the selected classes participated in the survey. Figure 1 shows the flow chart of the participants' selection and data extraction and matching. A total of 6047 junior high school students participated in physical examinations and a questionnaire survey for 2 years was included in this study. This study was approved by the Ethics Committee of the Shanghai Municipal Center for Disease Control and Prevention.

Flow chart of the participants' selection and data extraction and matching from the surveillance of the Students' Common Diseases and Health Influencing Factors in Shanghai, China, 2019–2020

2.2 Physical examination

Each district established a physical examination team composed of professionals, including professional technicians from the centres for disease control and prevention and community health service in each district. The physical examination team entered the schools after a unified training program. Heights and weights were measured by two professional technicians following a uniform procedure: one for the measurement and one for the recording. The standing height was measured to the nearest 0.1 cm using a stadiometer, and weight was measured to the nearest 0.1 kg using a digital scale. The digital scale was Locosc XK3150 manufactured by Ningbo Langke Precision Technology Limited Company, and the stadiometer was TZG manufactured by Wuxi Weigher Factory Limited Company. Boys and girls were measured separately, and students were required to remove their shoes and only wear underwear during the measurements. Each student's physical examination form had a unique identification code. Prevention and control measures for COVID-19 had been added to the 2020 survey, including wearing masks, maintaining distances, and disinfecting the instruments after each student's physical examination.

The definitions of overweight and obesity were based on the body mass index (BMI), which was calculated as weight (in kilograms) divided by the square of height (in meters). The age- and sex-specific BMI cut-offs defined by the World Health Organization (WHO) were chosen to assess the weight status of each individual.16 A Z-score of the BMI greater than one was defined as overweight, and a Z-score greater than two was defined as obesity.16

2.3 Questionnaire survey

The survey was conducted using the “Student Health and Influencing Factors questionnaire” compiled by the project of National Students' Common Diseases and Health Influencing Factors. On the day of the investigation, after a brief explanation, the questionnaires were completed by the students themselves. The investigators collected the questionnaires on the spot and checked the accuracy and completeness of the questionnaires simultaneously. The questionnaires collected the students' gender, age, self-recognized household income, screen behaviours, physical activity behaviours, sleep duration, and dietary behaviours. Household income was obtained through the following questions “What do you think of your family's financial situation?” The options were “very good, good, general, poor, or very poor.” The question wording and details for lifestyle behaviour variables are shown in Table 1. Most of the variables are categorical variables, not continuous variables. This study only analyses dietary behaviours related to adolescent obesity, including the frequency of intake of sugar-sweetened beverages, fresh vegetables, fruits, and breakfast.

TABLE 1. Question wording and details for lifestyle behaviour variables—Surveillance of Students' Common Diseases and Health Influencing Factors, Shanghai, China, 2019–2020 Variable Question Response options Coding for analysis Screen behaviours TV time ≥2 h/day During the past week, how long did you watch TV on average every day? I have not watched it, <1 h, 1–2 (excluding 2) h, 2–3 (excluding 3) h, 3–4 (excluding 4) h, or ≥ 4 h ≥2 h/day versus <2 h/day Computer time ≥2 h/day During the past week, how long did you usually play computer every day? I have not played it, <1 h, 1–2 (excluding 2) h, 2–3 (excluding 3) h, 3–4 (excluding 4) h, or ≥ 4 h ≥2 h/day versus <2 h/day Mobile screen time ≥2 h/day During the past week, how long did you spend on mobile electronic devices on average every day (including mobile phones, tablets, etc.)? I have never used it, or an average of ____ h ___ min/day ≥2 h/day versus <2 h/day Mobile screen time, h/day (P25–P75) Numerical variables Physical activities behaviours MVPA ≥ 60 min/day on all 7 days During the past week, how many days have you been able to do at least 60 min of MVPA every day (accumulatively)? (MVPA refers to exercises that make you breathless or increase your heartbeat. Such as running, basketball, football, swimming, aerobics in the gym, lifting heavy objects, etc.) 0 days, 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, or 7 days 7 days versus <7 days Days of MVPA≥60 min/day Numerical variables Outdoor activities ≥2 h/day During the past week, how much time did you spend outdoors during the day (accumulatively)? <1 h, 1–2 (excluding 2) h, 2–3 (excluding 3) h, ≥3 h, or I do not know ≥2 h/day versus <2 h/day Sleep duration Sleep duration, h/day During the past week, how long did you sleep on average every day? ____h____min Numerical variables Dietary behaviours Sugar-sweetened beverage ≥1 time/day During the past week, how many times did you drink sugar-sweetened beverages? Never drink, <1 time/day, or ≥ 1 time/day ≥1 time/day versus <1 time/day Fresh fruit ≤1 time/day During the past week, how many times did you eat fresh fruits (not including canned fruits)? Never eat, <1 time/day, 1 time/day, or ≥2 times/day ≤1 time/day versus ≥2 times/day Vegetables ≤1 time/day During the past week, how many times did you eat vegetables? (Both raw and cooked) Never eat, <1 time/day, 1 time/day, or ≥2 times/day ≤1 time/day versus ≥2 times/day Did not eat breakfast on all 7 days During the past week, did you eat breakfast every day? Every day, sometimes, or never eat <7 days versus 7 days 2.4 Quality control

All of the investigators took up their posts after the uniform training, and all of the surveying personnel came from professional institutions. All of the instruments were calibrated before the measurements. The physical examinations and questionnaire surveys in each district were supervised by the Shanghai Municipal Center for Disease Control and Prevention. Quality control personnel were established in each district to perform the quality control. On the day of the survey, the data of the physical examinations and questionnaires were checked for whether there existed omissions or errors and whether the handwriting on each survey was clear. In addition, 5% of students were randomly selected every day to retest their heights and weights, and the problems found were corrected at that time.

2.5 Statistical analysis

EpiData3.1 software was used for the data entry. The data were cleaned and merged into a final database. SAS9.3 software was used for the statistical analysis. The normal numerical variables were expressed by mean (SD), and the non-normal numerical variables were expressed by median (P25–P75). The categorical variables were expressed using n (%) or prevalence (95% CI). Paired tests, paired t-tests or Wilcoxon signed-rank test was used to compare changes in the obesity prevalence, BMI, screen behaviours, physical activity behaviours, sleep duration and dietary behaviours from 2019 to 2020. tests, t-tests or Wilcoxon rank-sum test was used to compare the differences in the obesity prevalence, obesity incidence, screen behaviours, physical activity behaviours, sleep duration and dietary behaviours between different subgroups. Because the data are 2-year longitudinal data, generalized estimation equation analysis was used to analyse the influencing factors of the obesity prevalence of junior high school students from 2019 to 2020. The dependent variable was the obesity prevalence of students from 2019 to 2020. The independent variables included gender, age, region, household income, screen behaviours, physical activity behaviours, sleep duration, and dietary behaviours from 2019 to 2020 (see Table 1 for detailed variable information). p < 0.05 indicated that the difference was statistically significant.

3 RESULTS 3.1 Characteristics of participants

Table 2 shows the characteristics of participants in this study in 2019. A total of 6047 junior high school students aged 11–16 years participated in physical examinations and questionnaire surveys for 2 years. Among them, 3093 (51.1%) were boys and 2954 (48.9%) were girls, with an average age of 12.7 years. The average BMI of participants in 2019 was 20.3 kg/m2, of which the boys were 20.8 kg/m2 and the girls were 19.7 kg/m2. The prevalence of obesity was 14.2%, and boys (20.9%) were higher than girls (7.1%).

TABLE 2. Characteristics of 6047 participants in Shanghai, China, 2019 Total Boys Girls Total 6047 3093 (51.1) 2954 (48.9) District Urban 2706 (44.8) 1449 (46.8) 1257 (42.6) Rural 3341 (55.3) 1644 (53.2) 1697 (57.4) Age 11~ 1597 (26.4) 787 (25.4) 810 (27.4) 12~ 1993 (33.0) 1060 (34.3) 933 (31.6) 13~ 2043 (33.8) 1010 (32.7) 1033 (35.0) 14–16 414 (6.9) 236 (7.6) 178 (6.0) Household income Good 2856 (47.3) 1463(47.4) 1393 (47.2) General 2886 (47.8) 1462 (47.3) 1424 (48.3) Poor 298 (5.0) 164 (5.3) 134 (4.5) Height, cm (SD) 158.6 (8.7) 160.3 (9.9) 156.7 (7.0) Weight, kg (SD) 51.4 (13.3) 53.9 (14.5) 48.7 (11.3) BMI, kg/m2 (SD) 20.3 (4.1) 20.8 (4.3) 19.7 (3.9) Overweight prevalence, % (95% CI) 22.2 (21.2–23.3) 25.2 (23.6–26.7) 19.2 (17.8–20.6) Obesity prevalence, % (95%CI) 14.2 (13.3–15.0) 20.9 (19.4–22.3) 7.1 (6.2–8.1) Note: Data are n (%), mean (SD), or prevalence (95% CI). 3.2 Changes in the obesity prevalence and incidence from 2019 to 2020

The obesity prevalence among the 6047 junior high school students in Shanghai, China, rose from 14.2% in 2019 to 15.4% in 2020 (p < 0.01), of which boys increased from 20.9% in 2019 to 23.1% in 2020 (p < 0.01), and girls did not increase (7.1% vs. 7.2%, p = 0.29). In addition, from 2019 to 2020, the prevalence of obesity in urban and rural areas in 12- and 13-year-old groups and adolescents with good and general household income also increased significantly (see Table 3).

TABLE 3. Obesity prevalence and incidence of adolescents in Shanghai, China, 2019–2020 Prevalence of obesity Incidence of obesity N 2019 2020 p* Na n (%) p** Total 6047 856 (14.2) 929 (15.4) <0.01 5191 224 (4.3) / District 0.60 Urban 2706 408 (15.1) 428 (15.8) <0.01 2298 103 (4.5) Rural 3341 448 (13.4) 501 (15.0) <0.01 2893 121 (4.2) Gender <0.01 Boys 3093 645 (20.9) 715 (23.1) <0.01 2448 169 (6.9) Girls 2954 211 (7.1) 214 (7.2) 0.29 2743 55 (2.0) Age 0.12 11~ 1597 257 (16.1) 268 (16.8) 0.21 1340 52 (3.9) 12~ 1993 293 (14.7) 317 (15.9) 0.02 1700 80 (4.7) 13~ 2043 256 (12.5) 296 (14.5) <0.01 1787 84 (4.7) 14–16 414 50 (12.1) 48 (11.6) 0.18 364 8 (2.2) Household income 0.77 Good 2856 394 (13.8) 422 (14.8) 0.01 2462 105 (4.3) General 2886 413 (14.3) 454 (15.7) <0.01 2473 106 (4.3) Poor 298 49 (16.4) 53 (17.8) 0.86 249 13 (5.2)

Table 3 also showed the incidence of obesity in 2019–2020 was 4.3% and that of boys was 6.9%, which was higher than the 2.0% of girls (p < 0.01). There was no statistical difference between the obesity incidence of urban and rural areas, different ages, and different household incomes.

Taking BMI as the analysis variable, the average BMI of participants increased from 20.3 kg/m2 in 2019 to 21.2 kg/m2 in 2020. From 2019 to 2020, the average BMI of adolescents in different districts, different age groups, and different household incomes all have a significant increase (p < 0.01). The increase in BMI of boys was greater than that of girls, and the younger the age groups, the greater the increase in BMI (see Table 4 and Figure 2).

TABLE 4. BMI of adolescents in Shanghai, China, 2019–2020, kg/m2 N 2019 2020 Δ2020–2019a p* Total 6047 20.3 (4.1) 21.2 (4.4) 0.9 (0.2–1.8) / District 0.27 Urban 2706 20.3 (4.2) 21.3 (4.2) 1.0 (0.2–1.8) Rural 3341 20.2 (4.1) 21.2 (4.5) 0.9 (0.2–1.8) Gender <0.01 Boys 3093 20.8 (4.3) 21.9 (4.7) 1.0 (0.2–2.0) Girls 2954 19.7 (3.9) 20.6 (3.9) 0.9 (0.2–1.6) Age <0.01 11 1597 19.6 (3.8) 20.6 (4.0) 1.1 (0.4–1.9) 12 1993 20.2 (4.2) 21.2 (4.3) 1.0 (0.2–1.9) 13 2043 20.7 (4.2) 21.7 (4.5) 0.8 (0.1–1.7) 14–16 414 21.0 (4.4) 21.7 (5.0) 0.7 (0.0–1.6) Household income 0.45 Good 2856 20.2 (4.0) 21.1 (4.2) 0.9 (0.2–1.8) General 2886 20.3 (4.2) 21.4 (4.5) 1.0 (0.2–1.8) Poor 298 20.6 (4.1) 21.6 (4.4) 0.9 (0.2–2.0) Note: Data are mean (SD) or median (P25–P75). Abbreviation: BMI, body mass index.

Changes in BMI of Chinese adolescents with different characteristics from 2019 to 2020. From 2019 to 2020, the average BMI of adolescents in overall, different districts, different age groups, and different household incomes all have a significant increase (p < 0.01). The increase in BMI of boys was greater than that of girls, and the younger the age groups, the greater the increase in BMI

3.3 Changes in lifestyle behaviours from 2019 to 2020 3.3.1 Screen behaviours

Table 5 shows the changes in lifestyle behaviours of adolescents in Shanghai, China, from 2019–2020. The proportion of the 6047 junior high school students who watched TV for ≥2 h/day during the past week dropped from 16.6% in 2019 to 13.3% in 2020 (p < 0.01). The proportion of students who played computers for ≥2 h/day during the past week increased from 8.7% in 2019 to 9.6% in 2020 (p = 0.05). The proportion of students who used mobile electronic devices for ≥2 h/day during the past week increased from 22.2% in 2019 to 28.0% in 2020 (p < 0.01). The 25th to 75th percentile of mobile screen time increased from 0.25 to 1.50 h/day to 0.33–2.00 h/day (p < 0.01). From 2019 to 2020, the changes in the screen behaviours of boys and girls were consistent with the overall trend, except that the proportion of playing computers for ≥2 h/day increased significantly in girls, while no significant increase in boys. In 2019 and 2020, the proportion of boys who watched TV and played computers for ≥2 h/day was higher than that of girls, except for mobile screen time.

TABLE 5. Lifestyle behaviours of adolescents in Shanghai, China, 2019–2020 Variable Total Boys Girls 2019 2020 p* 2019 2020 p* 2019 2020 p*