Twisk 1998.

Methods	Study design: prospective cohort study. Analyses methods for cohorts: multiple dietary assessments. Analyses: 1st‐order autoregressive model (fatness at each time point related to exposure at previous time point) estimated by GEEs) with the within‐subject correlations taken into account using GEEs. How were missing data handled? 24% (233/307) lost to follow‐up over 1st 4 years of study. Comparisons between dropouts and remaining participants revealed no selective dropout after 1st year in relation to anthropometric variables, nutrition intake and physical activity. Number of study contacts? 4 (baseline‐13 years, 14 years, 15 years, and 16 years). Period of follow‐up: 3 years. Periods of recruitment: 1977. Sample size justification adequately described ‐ yes/no? The AGAHLS study included 698 children from 2 equally large secondary schools in Amsterdam. Schools selected based on location, i.e. 1 of the schools in a rural area, the other in an urban area, as being representative of the Dutch adolescent population of the 1970s. Sampling method: convenience. Healthy pupils from the 1st and 2nd years of 1 secondary school in Amsterdam. Study objective: to analyse longitudinal relationships between BMI/SSF, and biological and lifestyle risk factors for coronary heart disease. Study population: boys and girls aged 13 years in Amsterdam.
Participants	Baseline characteristics (overall) Age (years): overall 13 (SD 0.7); boys 13 (n = 82); girls 13 (n = 97). Sex (% girls): 54.2%. Ethnicity: NR. Education: above average. Parent income: above average. Pubertal stage: NR. Parental BMI: NR. Child total energy: NR. Child total fat: NR. Child total protein: NR. Child total CHO: NR. Child physical activity: NR. Child physical inactivity or screen time or both: NR. Child CVD risk (excluding fatness): NR. Child body fatness, BMI (kg/m2): overall 17.7 (SD 1.93); boys 17.3 (SD 1.6); girls 18.1 (SD 2.1). Child body fatness, SSF (mm): overall 33.3 (SD 12.8); boys 28.4 (SD 10.9); girls 37.5 (SD 12.8). Included criteria: healthy boys and girls aged 13 years. Excluded criteria: NR. Pretreatment: NA. Brief description of participants: healthy, Dutch school children aged 13 years with above average socioeconomic status who were participants of the Amsterdam Growth Health Longitudinal Study. Total number completed in cohort study: 233 (102 boys, 131 girls) completed 4 annual measurements. Total number enrolled in cohort study: 307 (148 boys, 159 girls).
Interventions	Description of exposure for cohorts Time span: 4 years. Dietary assessment method used: cross‐checked, dietary history interview. Frequency of dietary assessments: 1 assessment at each follow‐up visit (at 14, 15 and 16 years).
Outcomes	Regression data reported in a graph.
Notes	We contacted the authors about the data at ages 14, 15 and 16 years, but had not received this by time of publication, and thus could not classify this study.

AGAHLS: Amsterdam Growth and Health Longitudinal Study; BMI: body mass index; FFQ: Food Frequency Questionnaire; GEE: generalised estimating equation; NA: not available; NR: not reported; SD: standard deviation; SSF: sum of skinfolds.