Skinner 2004.

Methods	Study design: prospective cohort study. Analyses methods for cohorts: longitudinal dietary intake based on 9 sets of 3‐day dietary data from children aged 2‐8 years. Changes in energy intake over time and gender differences in energy intake tested with GLM repeated measures ANOVA. How were missing data handled? Lost to follow‐up at 3 years: 23 (reasons: travel time required for interviews); at 3.5 and 8 years: 5 (reasons: n = 4: family moved, discontinued participation; n = 1: consistently incomplete data provided by mother). No analysis performed comparing children who completed study to children who did not. Number of study contacts: 11 (2.0, 2.3, 2.7, 3.0, 3.5, 4.0, 4.5, 5.0, 6.0, 7.0 and 8.0 years). Period of follow‐up (total period of observation): 8 years. Periods of recruitment: May‐September 1992. Sample size justification adequately described? No. Sampling method: purposively selected sample of 98 infants aged 2 months recruited from 2 metropolitan areas in Tennessee. Current analysis based on data from 62 children from original cohort, 2 infants who were selected as replacements prior to 1 year of age for cohort and 6 children aged 2 years who participated in a similar infant study from the same laboratory. Study objective: to identify longitudinal variables related to children's BMI at 8 years. Study population: healthy white children aged 2‐8 years in urban area of Tennessee, USA.
Participants	Baseline characteristics (reported for 1 overall group) Age (eligible for inclusion in years): overall 2.0; boys 2.0; girls 2.0. Sex: 47.1% girls. Ethnicity: white. Education: most parents had a college degree. Income: most were from middle or upper socioeconomic families. Pubertal stage: NA. Parental BMI (kg/m2): mother BMI: overall 25.4 (SD 4.6); boys 26 (SD 4.6); girls 24.8 (SD 4.6); father BMI: overall 26.5 (SD 3.7); boys 27.5 (SD 3.9); girls 25.3 (SD 3.0). Child total energy (kJ) (mean of dietary assessments at ages 24, 28 or 32 months): overall 5870 (SD 1474); boys 6061 (SD 1649); girls 5655 (SD 1214). Child total fat (g) (mean of dietary assessments at ages 24, 28 and 32 months): overall 50 (SD 16); boys 51 (SD 17); girls 48 (SD 15). Child total protein (g) (mean of dietary assessments at ages 24, 28 and 32 months): overall 49 (SD 17); boys 49 (SD 17); girls 49 (SD 16). Child total CHO (g) (mean of dietary assessments at ages 24, 28 and 32 months): overall 197 (SD 50); boys 206 (SD 55); girls 186 (SD 40). Child physical activity: NR. Child physical inactivity or screen time or both (hours/day): overall 2.85 (SD 1.21); boys 2.9 (SD 1.3); girls 2.8 (SD 1.1). Child CVD risk (excluding fatness): NR. Child body fatness:BMI (kg/m2): overall 16.4 (SD NR); boys 16.5 (SD NR); girls 16.2 (SD NR); weight (kg): overall 12.6 (SD 1.5); boys 13.0 (SD 1.5); girls 12.1 (SD 1.4). Included criteria: children who participated in the original birth cohort aged 2‐8 years with available follow‐up data. Excluded criteria: NR. Brief description of participants: children aged 2‐8 years. Total number completed in cohort study: 70 (37 boys, 33 girls). Total number enrolled in cohort study: 98 (+2 prior to 1 year; +6 at age 2 years).
Interventions	Description of exposure for cohorts Time span: 6 years. Dietary assessment methods used: interviews conducted by 2 dieticians. Conducted 24‐hour dietary recall + assessment of 2‐day food record. Dietary assessment included 3 non‐consecutive days (2 week days and 1 weekend day). Frequency: single 24‐hour recall and 2‐day DR at 9 time points: 2 years (baseline), 2.3, 2.7, 3.0, 3.5, 4.0, 4.5, 5.0, 6.0, 7.0 and 8.0 years). Dietary intakes from each time interval were averaged to provide 9 representative daily intakes. See Table 9; Table 10; Table 11; Table 12; Table 13; Table 14; Table 15; Table 16; Table 17; Table 18 for details of total fat intake exposure per outcome.
Outcomes	BMI BMI (kg/m2). Body fat Body fat (%). Body fat (g). Body fat (kg). Sum of skinfolds Sum of 4 skinfolds (biceps, triceps, subscapular, supra‐iliac skinfolds) (mm).
Identification	Sponsorship source: Gerber Products Company and Tennessee Agricultural Experiment Station. Country: USA. Setting: Urban households, Tennessee. Comments: NA. Author's name: JD Skinner. Institution: Nutrition Department, University of Tennessee, Knoxville, TN, USA. Email: skinner@utk.edu. Declaration of Interests: no. Study ID: Skinner 2004. Type of record: journal article.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Were adequate outcome data for cohorts available? All outcomes	High risk	Relatively high number of dropouts (36.7% over 6 years; 62/98 children recruited for study were analysed). Baseline data between children who completed and children who did not were not compared.
Was there matching of less‐exposed and more‐exposed participants for prognostic factors associated with outcome or were relevant statistical adjustments done? All outcomes	Low risk	Age, gender, ethnicity and SES were matched while parental BMI, BMI at baseline, adiposity rebound age and physical inactivity were adjusted in linear regression models.
Did the exposures between groups differ in components other than only total fat? All outcomes	Low risk
Can we be confident in the assessment of outcomes? All outcomes	Low risk	Standard methods performed for measurements of weight, height and DEXA (by trained personnel).
Can we be confident in the assessment of exposure? All outcomes	Low risk	Repeated 3‐day DR completed by mothers who were taught to describe and estimate portion sizes of child's food and beverage intake. Dietician reviewed food records with mother.
Can we be confident in the assessment of presence or absence of prognostic factors? All outcomes	Unclear risk	Information on physical inactivity self‐reported and data collection method not well described.
Was selection of less‐exposed and more‐exposed groups from the same population? All outcomes	Low risk	All children selected for 1 cohort study.