Is intake of dietary fat associated with adiposity in children?

Moderate evidence from prospective cohort studies suggests that increased intake of dietary fat is associated with greater adiposity in children. However, there were no studies conducted under isocaloric conditions.

The Dietary Guidelines Advisory Committee (DGAC) conducted a full Nutrition Evidence Library (NEL) search to evaluate the association between dietary fat intake and adiposity in children. Results of this review were supplemented by the findings of prospective studies included in an earlier evidence review conducted by the American Dietetic Association (ADA). This conclusion was based on 28 peer-reviewed articles which addressed the research question, 21 studies from the earlier ADA review; and seven studies from the subsequent NEL review. This included four randomized controlled trials (RCTs) (Caballero, 2003; Hakanen, 2006; Lauer, 1995; Niinikoski, 2007); and 24 longitudinal studies (21 from the ADA review and three from the NEL review) (Alexy, 2004; Johnson, 2008b; Karaolis-Danckert, 2007; Alexy, 1999; Berkey, 2000; Bogaert, 2003; Boulton, 1995; Carruth, 2001; Davison, 2001; Eck, 1992; Francis, 2003; Gazzaniga, 1993; Klesges, 1995; Lee, 2001; Maffeis, 1998; Magarey, 2001; Newby, 2003; Robertson, 1999; Rolland-Cachera, 1995; Scaglioni, 2000; Shea, 1993; Skinner, 2003; Skinner, 2004; Wang, 2003). Fourteen of the studies were conducted in the US.

Of the 24 longitudinal studies, 15 found a positive association between total fat intake or intake of high-fat foods and adiposity in all or a sub-sample of the population studied (Carruth, 2001; Davison, 2001; Eck, 1992; Francis, 2003; Gazzaniga, 1993; Johnson, 2008a; Karaolis-Dankert, 2007; Klesges, 1995; Lee, 2001; Magarey, 2001; Newby, 2003; Robertson, 1999; Skinner, 2003; Skinner, 2004; Wang, 2003). The varied results between studies were a product of using multiple measures of adiposity within the same study, conducting analyses stratified by different variables (e.g., sex, weight status) or dietary fat measured in both absolute terms (total grams) as well as a percent of energy intake. Nine other longitudinal studies found no association between total fat intake and adiposity in children (Alexy, 1999; Alexy, 2004; Berkey, 2000; Bogaert, 2003; Boulton, 1995; Maffeis, 1998; Rolland-Cachera, 1995; Scaglioni, 2000; Shea, 1993). A greater proportion of the studies that found a positive association between dietary fat and adiposity, however, used multiple measures of adiposity, such as skinfold measures and body composition by dual energy X-ray absorptiometry (DEXA), rather than only body mass index (BMI), which provides a poor estimate of actual body fat (Freedman, 2009).

Three of the four RCTs found no association between percent energy from dietary fat and adiposity. In the Special Turku Coronary Risk Factor Intervention Project for Children (STRIP) clinical trial, which tested the effects of a fat-modified diet from seven months of age (Hakanen, 2006) reported less obesity among intervention girls compared with control girls at age 10 years, but no differences for boys; while at age 14 years, Niinikoski et al, (2007) found no difference in obesity between treatment groups, for either males or females. Caballero et al, (2003) reported no change in percent body fat in a three-year school-based nutrition and physical activity intervention among 1,704 Native American children, who were age seven years at baseline. Results showed that percent body fat and BMI did not differ by treatment group at study end. However, children in the intervention group reported lower total energy intake (1,892 vs. 2,157kcal per day) and percent energy from total fat (31.1% vs. 33.6%) compared with the control group, and percent energy from fat was lower in the intervention school lunches compared to the control schools (28.2% vs. 32.0%). Finally for the Dietary Intervention in Children (DISC) trial (Lauer, 1995), which tested the safety and efficacy of lowering dietary intake of fat and cholesterol in children with elevated low-density lipoprotein cholesterol (LDL-C), analyses of growth patterns showed no difference in BMI, height or weight between the lower-fat, lower saturated fat intervention groups vs. controls. It should be noted, however, that in this trial, great effort was taken to assure that energy intake would not decrease and growth would be maintained, since the goal was to show that lipids could be improved without a deleterious effect on growth.

In summary, the combination of evidence from methodologically strong studies in the NEL and ADA reviews supports a conclusion that dietary fat and adiposity in children are positively associated. Methodological differences between studies, however, were significant, especially with respect to dietary assessment procedures, identification of implausible energy intake reports, choice of anthropometrics, and statistical approaches. Despite these methodological differences and limitations, collectively the studies tended to find either a positive association or no significant (NS) association between dietary fat and adiposity with the weight of evidence leaning towards a positive association. Additional prospective studies that assess both the amount and type of fat in relation to changes in childhood adiposity are warranted, however.

Randomized Controlled Trials (4)

Caballero et al, 2003 (positive quality) conducted a randomized, controlled, school-based intervention trial (The Pathways Study) in American Indian third graders to evaluate the effectiveness of a school-based, multi-component intervention for reducing percentage body fat. The study included 1,409 children, with a mean age of mean age 7.6±0.6 years at baseline. Attrition rate was 17%. This study tested a four-part intervention: Change in diet (reduction in energy density in school meals via fat reduction and increased fruits and vegetables), increased physical activity, classroom curriculum, family involvement and the primary outcome variable was percent body fat (%BF). Dietary intake was assessed by direct observation of school lunch at baseline and three-year follow-up, and a 24-hour dietary recall was taken at the three-year follow-up. Physical activity was measured by motion sensor and self-reported questionnaire. Body mass index was calculated using measured height and weight and %BF was measured using bioelectric impedance. Results showed that %BF and BMI did not differ between intervention groups at the end of the study. The 24-hour diet recall showed a significantly lower total energy intake (1,892 vs. 2,157kcal per day) and percent of energy from total fat (31.1% vs. 33.6%) in the intervention group compared with the control group. School lunch observation at follow-up was similar in total energy content (683 vs. 688kcal) in the intervention compared with control schools; however, percent energy from fat was lower in the intervention school lunches compare to the control schools (28.2% vs. 32.0%).

Hakanen et al, 2006 (positive quality) analyzed data from a cohort of subjects from Finland to evaluate the impact of nutrition counseling on the prevalence of overweight. This study was a part of the Special Turku Coronary Risk Factor Intervention Project for Children (STRIP), which is a prospective, randomized trial aimed at reducing the exposure of the intervention children to the known risk factors of atherosclerosis. Children were followed from seven months to 10 years of age. Intervention participants were counseled to consume 30% of energy from fat (30-35% between one and two years), with a ratio of 2:1 for unsaturated fat to saturated fat. Height and weight of the children were measured, BMI was calculated, and weight status was determined. Children were classified as overweight or obese if their weight for height was >20% or at least 40% above the mean weight for height of healthy Finnish children, respectively. Analyses were adjusted for study group, birth weight, age, mother’s BMI, father’s BMI and pubertal status at 10 years. The final sample included 585 children. Results showed that there were continuously fewer overweight girls in the intervention group than in the control group. At the age of 10 years, 10.2% of the intervention girls and 18.8% of the control girls were overweight (P=0.0439), whereas 11.6% of the intervention boys and 12.1% of the control boys were overweight (P=1.00). However, the study group was not a statistically significant predictor of overweight in the model. Mean fat intake levels for the intervention and control groups were not reported in this manuscript.

Lauer et al, 1995 (positive quality) conducted a six-center RCT in the US to assess the effects of lowering dietary intake of total fat, saturated fat and cholesterol on weight in children. The intervention group received counseling on a diet containing 28% energy from total fat, less than 8% energy from saturated fat, up to 9% energy from polyunsaturated fat, and <75mg/100kcal per day of cholesterol (not to exceed 150mg per day). The intervention was given over a one year period, with follow-up occurring through year three. The control group received usual care. Dietary intake was assessed using three 24-hour recalls. Height and weight were measured to calculate BMI, skinfold thickness was measured and waist and hip circumference were measured. The final sample included 334 children in the intervention group and 329 children in the control group (ages eight to 10 years at baseline). Mean percentage of energy from fat decreased in both groups (33% in the control and 28.6% in the intervention at year three), through more so in the intervention groups. Mean percent energy from saturated fat (10% in the intervention and 12% in the control at year three) and cholesterol (95mg in the intervention and 113mg in the control group at year three) decreased in the intervention group, with little change in the control group. There were NS differences in mean weight, BMI or skinfold thickness between the groups.

Niinikoski et al, 2007 (positive quality) analyzed data from a cohort of subjects from Finland to evaluate the effect of low-saturated fat, low-cholesterol dietary counseling on fat intakes, growth, serum cholesterol values and pubertal development in children and adolescents. This study was a part of the Special Turku Coronary Risk Factor Intervention Project for Children (STRIP). Children were followed from seven months to 14 years of age. Intervention participants were counseled to consume 30% of energy form fat (30-35% between one and two years), with a ratio of 2:1 for unsaturated fat to saturated fat. Height and weight of the children were measured and BMI and weight status were determined. The final sample included 585 children. Intervention children had significantly lower intake of total fat and saturated fat compared to control children (P<0.001), but mean intake levels for each study group are not reported. The two study groups showed no difference in growth, BMI, pubertal development or age at menarche.

Cohort Studies (24)

Alexy U et al, 1999 (positive quality) analyzed data from a prospective cohort study (the DONALD study) in Germany in order to evaluate the effects of dietary intake, including fruit juice intake, on prevalence of obesity over a three year period. Children were followed from age three to five years. Dietary fat intake was determined using data collected from three-day weighed diet records, and weight status was determined using measured height and weight. The final sample includes 205 children (105 boys, 100 girls). Subjects were split into groups based on juice consumption, with low juice consumers consuming 33% of energy from fat, and high juice consumer consuming 38% energy from fat. Results showed that children’s BMI correlated positively with energy intake (r=0.18; P<0.05), but not with intake fat (percent of energy intake).

Alexy et al, 2004 (positive quality), analyzed data from a prospective cohort study (the DONALD study) in Germany to evaluate the influence of long-term dietary fat intake on BMI. Children were followed from age two to 18 years. Dietary fat intake was determined using data collected from three-day weighed diet records, and weight status was determined using measured height and weight. The final sample included 228 children (114 boys, 114 girls). A cluster analysis revealed four fat intake patterns: Constant (38% energy from fat), Medium (36% energy from fat), High (40% energy from fat) and Low (32% energy from fat). The clusters did not differ on any of the measured subject characteristics or mean energy intake. However, energy density was lowest in the Low cluster (P<0.0001). The High and Constant clusters consumed more meat/fish/eggs and fats/oils (P<0.0001), while the Low and Medium clusters consumed more fruits/vegetables (P<0.0001). Differences in BMI by fat cluster were not seen at the beginning or end of the study; however, during the study period, mean BMI differed significantly between clusters, with the highest BMI in the low fat intake cluster (0.26(0.70); P<0.05), followed by the Medium cluster (0.11(0.85), High cluster (0.06(0.88) and Constant cluster (-0.30(0.79).

Berkey CS et al, 2000 (positive quality) used data from a prospective cohort study in the US to examine the role of dietary patterns on annual weight changes among preadolescents and adolescents. Subjects were from the Growing Up Today Study, and were nine to 14 years old and were followed for one year. Dietary fat intake was determined using a food-frequency questionnaire (FFQ), and fat intakes were energy-adjusted. Adiposity was assessed using BMI, based on self-reported height and weight. All models controlled for race/ethnic group, baseline BMI, annual change in height, menstrual history in girls, Tanner stage, and age. The follow factors were entered into the models: Total energy intake, fat intake, fiber intake, number of gym classes per week, hours of physical activity and hours of tv or video games. The final sample included 6,149 girls and 4,620 boys. For both girls and boys, dietary fat intake was not predictive of one-year change in BMI.

Bogaert N et al, 2003 (neutral quality) used data from a prospective cohort study in Australia to examine whether measures of energy intake predict excessive weight gain over time in children. Children were between the ages of six and nine years, and were followed for one year. Dietary fat intake was assessed using a three-day food record. Body mass index z-score was calculated using measured height and weight, and body composition was assessed using bioelectrical impedance analysis. The final sample included 41 children (mean age=8.6 years). Mean fat intake was as follows: 34% for boys under age eight years, 32% for girls under age eight years, 38% for boys over age eight years and 34% for girls over age eight years. There was no relationship between dietary fat intake and BMI z-score change from baseline to one year.

Boulton TJC et al, 1995 (neutral quality) analyzed prospective cohort data from Australia to assess the relationship between food energy and nutrient intake on children’s growth. Subjects were participants in the Adelaide Nutrition Study, and were followed from age three months to eight years. Percent energy from fat was determined using seven-day (up to two years), three-day (four years), and four-day (six and eight years) food records. Subjects were split into tertiles of fat intake, <30%, 30-34.9%, and >35% in order to test the association between fat intake and body weight/fatness. Body weight was measured and body fatness was determined by the sum of four skinfold measurements. The final sample included 140 children. Median fat intake was 44% at three months, and declined to 36% at six months, and remained at a similar level until eight years. There were no differences between the fat intake groups in body weight or fatness, except at three months (P<0.05), when those in the median fat intake group were heavier and had a higher percent body fat than those in the high fat group.

Carruth BR et al, 2001 (positive quality) used data from a prospective cohort study in the US to examine the association between pre-school children’s dietary intake and body composition. Subjects were followed from two months to eight years. Fat intake (g) and percent energy from fat was determined using three-day diet records taken six times between age two months and eight years. Body composition was measuring using DEXA. The final sample included 53 children. Mean fat intake over time was 30-33% of energy. Higher mean longitudinal intake (24-60 months) of monounsaturated fat was associated with lower body fat at 70 months (P=0.02). However, higher mean longitudinal intake of total dietary fat was positively associated with body fat at 70 months (P=0.02).

Davison KK et al, 2001 (positive quality) analyzed data from a prospective cohort in the US to assess predictors of change in girls’ BMI between ages five and seven years. Children were participants in the Girls Needs Study. Percent energy from fat was determined using three 24-hour recalls from the girls’ mother and BMI was calculated using measured height and weight. Regression models included physical activity, dietary intake, family income, parent education, BMI at age five years, family risk of overweight, parental change in BMI, parent physical activity, parents’ dietary intake. The final sample included 192 girls. Mean percent energy from fat at five years was 31%. Girls with greater increased in BMI between ages five and seven had a higher percentage of energy from fat at age five years (P<0.02), though level of fat intake is not reported.

Eck LH et al, 1992 (neutral quality) examined familial risk of obesity, dietary intake and weight status using data from a group of children in the US. Children were followed for a one-year period from age four to five years. Children were divided into two groups based on parental weight status; the high-risk group had children with one or two overweight parents, while the low-risk group had children with no overweight parents. Percent energy from fat was determined using a FFQ. Children’s weight was measured at baseline and a one-year follow-up. The final sample included 187 subjects (92 high-risk, 95 low-risk). The high-risk group consumed a mean of 34% energy from fat, which was significantly higher than the low-risk group, who consumed a mean of 32% energy from fat (P=0.0004). The relationship between dietary fat intake and one-year weight change was not tested.

Francis LA et al, 2003 (positive quality) analyzed data from a prospective cohort in the US to assess whether consumption of high fat, energy dense snacks were associated with weight status. Children were participants in the Girls Needs Study, and were followed from age five to nine years. Percent energy from fat was determined using three 24-hour recalls from the girls’ mother, and BMI was calculated using measured height and weight. The model included tv viewing, snacking while watching TV, snacking frequency, fat intake from high energy density (ED) snack foods, increase in BMI from age five to nine years, while controlling for child BMI and family income. Girls were also divided into groups based on parental weight status. The final sample included 173 girls. Fat intake from energy-dense snacks was significantly positively associated with change in BMI between ages five years and nine years (P<0.05), however, fat intake levels are not reported.

Gazzaniga JM and Burns TL, 1993 (positive quality) analyzed prospective cohort data from the US to examine the relationship between diet composition and body fatness. Subjects were from the Muscatine Coronary Risk Factors Project and were identified for two study groups, either non-obese or obese. Dietary intake data was collected using three 24-hour diet recalls. Height and weight were measured and body composition was determined using skinfold thickness measurements. The final sample included 48 children (25 girls, 23 boys). Percentage body fat was positively correlated with intakes of total (P<0.0001), saturated fat (P<0.01), monounsaturated fat (P<0.0001) and polyunsaturated fat (P<0.01). After adjustment for study group, energy intake, resting energy expenditure (REE) and physical activity, the association remained for total, saturated and monounsaturated fat.

Johnson et al, 2008 (positive quality), conducted a longitudinal, observational cohort study in the United Kingdom to test the relationship between dietary energy density and fat mass in early adolescence. Subjects were from Children in Focus, a sub-sample of the Avon Longitudinal Study of Parents and Children (ALSPAC), and children were followed from age five to age nine years. Dietary intake was assessed using three-day diet diaries, and body fat mass at age nine was measured using DEXA. The final sample included 521 children with five and nine-year data available and 682 children with seven and nine-year data available. Pattern score at ages five and seven years was correlated with dietary energy density (r=0.8), fiber density (r= -0.7) and percentage of energy intake as fat (r= 0.5), and a one SD-increase in pattern score was associated with a 0.15kg (95% CI: -0.1 to 0.45kg) and a 0.28kg (95% CI: 0.05 to 0.53kg) higher fat mass at age nine years. The adjusted odds of excess adiposity at age nine years for the highest quintile compared to the lowest quintile of dietary pattern score were 2.52 (95% CI: 1.13 to 6.08) at five years of age and 4.18 (95% CI: 2.07 to 9.38) at seven years of age.

Karaolis-Danckert et al, 2007 (positive quality) analyzed data from a prospective cohort study (the DONALD study) in Germany to examine the interaction between rapid weight gain and nutrition in infancy and early childhood and their effect on percent body fat (%BF) trajectories. Children were followed from age two to five years. Dietary fat intake was determined using data collected from three-day weighed diet records. Weight status was determined using measured height and weight and rapid growers were identified as those children with an increase in weight SDS of 0.67 between birth and 24 months of age. The final sample included 249 children (51.4% girls). There was no relationship between a consistently high fat intake at 12 and 18-24 months and rapid weight gain at two years, but rapid growers who had a consistently high fat intake (>35% energy) at both 12 and 18-24 months, did not show the expected physiologic decrease in %BF between two to five years that was seen in rapid growers with an inconsistent or consistently low fat intake at these time points (0.73±0.26 percent per year; P=0.006). Conversely, normal growers with a consistently high fat intake at both 12 and 18-24 months had a significantly greater decrease in %BF between two to five years then when fat intakes were inconsistent or consistently low.

Klesges RC et al, 1995 (positive quality) analyzed data from a prospective cohort study in the US to determine the dietary predictors of weight change in children. Subjects were followed for a two-year period from age three to five years. Percent energy from fat was determined using a FFQ and BMI was calculated using measured height and weight. Analyses were adjusted for the following factors: baseline BMI, gender, age, family risk of overweight, gender by family risk interaction, baseline percent energy as fat, aerobic activity, change in percent energy from fat, change in leisure activity time. The final sample included 146 subjects. Baseline and change in fat intake were related to increases in BMI. A 5% higher percent calorie intake at baseline predicted a 0.168 increase in BMI; a 5% recent increase in fat intake predicted a 0.201 increase in BMI.

Lee Y et al, 2001 (positive quality) analyzed data from a prospective cohort in the US to compare the diet quality and weight status of girls consuming diets meeting the recommendation of the American Academy of Pediatrics for dietary fat with those consuming >30% energy from fat. Children were participants in the Girls Needs Study, and were followed from age five to seven years. Percent energy from fat was determined using three 24-hour recalls from the girls’ mother. Body mass index was calculated using measured height and weight and body fatness was measured using triceps and subscapular skinfold thickness measures. Subjects were divided into two groups: High-fat was >30% fat (55% of subjects) and low-fat was <30% fat (45% of subjects). The final sample included 192 girls. Girls on high-fat and low-fat diets did not differ in BMI at either five or seven years of age, but change in BMI between five and seven years was greater for girls consuming high-fat diets, even when controlling for BMI at five years (P<0.05). Change in the sum of skinfold thickness measured was also higher in the high-fat group compared to the low-fat group, such that girls consuming a higher fat diet at age five years had greater body fatness increases between age five and seven years (P<0.05).

Maffeis C et al, 1998 (positive quality) analyzed prospective cohort data from Italy to assess the relationship between diet, body composition, and adiposity in children. Subjects were followed for a four-year period, starting when children were eight years of age. Dietary intake data was collected through an interview with mothers and children regarding their usual weekly meal and snack intakes and food intake at school was assessed by reviewing the school’s menu with the children and asking which meals were consumed. Weight and height were measured and BMI was calculated. Analyses were controlled for age, gender, energy intake, percent energy intake as fat, protein, carbohydrate, parents’ BMI, TV viewing time, and time spent on vigorous physical activity. The final sample included 112 children (mean age at baseline=8.6 years). Children’s mean percent energy intake from fat was 32% at baseline. Mutivariate analyses showed that dietary fat intake was not associated with children’s BMI at age 12 years.

Magarey AM, et al 2001 (positive quality) analyzed data from a prospective cohort study from Australia to investigate the relationship between food energy and macronutrient intake and body fatness. Children were participants in the Adelaide Nutrition Study and were followed from age two to age 15 years. Fat intake (g) and percent energy from fat were determined using three-day (ages two, four and six years) and four-day (ages eight, 11, 13, 15 years) food records. Body mass index was calculated using measured height and weight and measures of subscapular skinfold thickness were taken. Analyses were adjusted for total energy intake, sex and parental body fat. The final sample included 243 subjects. Mean percent energy from fat decreased from 38% at two years, to 35% at ages four to 13 years, to 33-34% at 15 years. At six years, fat intake was positively associated with BMI-SDS at eight years (P<0.05). Fat intake at two years was positively associated with subscapular skinfold thickness at 15 years (P<0.05). Fat intake was NS associated with triceps skinfold thickness at any time point, nor was it associated with BMI or subscapular skinfold thickness at any other time-point.

Newby PK et al, 2003 (positive quality) analyzed prospective cohort data from the US to examine the relationship between dietary composition and weight change in children. Children were participants in the North Dakota Women, Infants and Children program (WIC), and were followed for two visits ranging from six to 12 months apart. Dietary intake data was collected using a semiquantitative FFQ, from which percent energy from fat was determine, and food groups were determined based on the North Dakota WIC program (referred to as ND food groups). Height and weight were measured. Analyses were adjusted for baseline weight, change in height during the time interval, age, sex, total energy intake, sociodemographic covariates (birth weight, maternal education, race/ethnicity, residence, poverty level). The final sample included 1,379 children (mean age; three years). Mean fat intake among this cohort of children was 34% of energy. Results showed a 0.07kg greater weight change per year for each additional serving of ND fat foods (95% CI 0.03-0.11kg; P=0.003). When all food groups were considered in a single model, each additional serving of ND fat foods was associated with a 0.05kg greater weight change per year (95% CI 0.01-0.09kg; P=0.03). However, total fat intake (percent energy from fat) was NS related to weight change. The authors concluded that high fat foods, but not total dietary fat per se, significantly predicted weight gain over a one-year period.

Robertson SM et al, 1999 (neutral quality) analyzed prospective cohort data from the US to compare diet between a group of children who had increased their sum of seven skinfold measures by 1.5 SD or more since the previous year and those who had not. Subjects were three or four year old children enrolled in a four-year longitudinal study. Food intake was assessed through observation during the school day by trained observers, and reported by parents for the time periods before and after school, and percent energy from fat was determined. Body fatness was determined by taking the sum of seven skinfold thickness measures. Children were grouped by change in adipose tissue over the previous year, those who were 1.5 SD above the mean for the previous year’s value were study subjects and control subjects were matched to the study subjects on gender, ethnicity and age. The final sample included 48 subjects (15 study subjects, 33 controls; mean age; three years at baseline). Mean dietary fat intake was 37% (75g) for study subjects and 34% (58g) for control subjects. A significant difference between study subjects and control subjects was found for fat grams (P=0.02), while the difference between groups for percent energy for fat was borderline significant (P=0.06).

Rolland-Cachera MF et al, 1995 (positive quality) analyzed prospective cohort data from France to test the relationship between early nutrient intake and adiposity development. Children were followed from two to eight years of age. Dietary interviews were conducted to ascertain percent energy from fat and adiposity was assessed via BMI, which was calculated using measured height and weight, and skinfold thickness measurements. The final sample included 112 children. The relationship between dietary fat intake at age two was not associated with adiposity at age eight years.

Scaglioni S et al, 2000 (positive quality) analyzed prospective cohort data from Italy to examine the influence of macronutrient intake in early on the development of overweight in children. Subjects were assessed at birth, one year and five years of age. Dietary intake data was collected via FFQ. Body weight was measured and BMI was calculated to determine children’s weight status. Adjustments were made for infant’s gender, weight and length at birth and one year of age and parental age. Dietary fat intake at age one years was not associated with children’s weight status at age five years and was approximately 35% of energy.

Shea S et al, 1993 (positive quality) analyzed prospective cohort data from the US to determine whether a moderately reduced fat diet affects growth in preschool children. Children were aged three to four years at baseline, and were followed for a mean of two years. Dietary intake was assessed using four 24-hour recalls and three FFQs administered to mothers over the one year baseline period. Height and weight were measured and BMI was calculated. Adjustments were made for age at the first 24-hour recall, sex, race/ethnicity and total energy intake. The final sample included 215 children (105 boys, 110 girls). Mean total fat intake was 32.5% based on the 24-hour recall data and 33.4% based on the FFQ data. Children were split into quintiles of dietary fat intake, I (27%), II (30%), III (32%), IV (35%) and V (38%). There were no differences in height, weight, or BMI across quintiles of total fat or saturated fat intake and there were no differences in growth parameters between children consuming <30% energy as fat compared to those consuming >30% energy from fat.

Skinner JD et al, 2003 (positive quality) used data from a prospective cohort study in the US to examine the association between pre-school children’s dietary intake and body composition. Subjects were followed from two months to eight years. Fat intake (grams) and percent energy from fat was determined using three-day diet records taken six times between age two months and eight years. Body composition was measuring using DEXA. The following variables were included in the model: mother’s percent body fat, mother’s BMI, father’s BMI, gender, sedentary activity, dietary intake (calcium, energy, protein, carbohydrate, fat, saturated fat, polyunsaturated fat, monounsaturated fat). The final sample included 52 children. Mean intake was 59g for boys, 54g for girls, or 32% of energy. Polyunsaturated fat was inversely associated with children’s body fat (P<0.02). However, total dietary fat and saturated fat intake were positively associated with body fat (P<0.01 and P<0.003).

Skinner JD et al, 2004 (positive quality) used data from a prospective cohort study in the US to examine the association between pre-school children’s dietary intake and body composition. Subjects were followed from two months to eight years. Fat intake (grams) and percent energy from fat was determined using three-day diet records taken six times between age two months and eight years. Body mass index was calculated using measured height and weight. The final sample included 70 children (37 boys, 33 girls). The following variables were included in the model: Gender, birth weight, breastfeeding duration, age that cereal was introduced, BMI at age two years, estimated AR, longitudinal energy and macronutrient intake, perception of child as a picky eater at age six years, number of foods liked at age eight years, screen time, and dietary variety score. The final sample included 70 children. Average longitudinal percent energy from fat was 32%. Dietary fat (longitudinal intake in grams and longitudinal percent energy from fat) was positively related to BMI at eight years (P<0.01).

Wang Y et al, 2003 (positive quality) analyzed prospective cohort data from China to examine the influence of dietary intake on adiposity in children. Children ages six to 13 years at baseline were participants in the China Health and Nutrition Survey and were followed over a two-year period. Dietary intake data were collected for three consecutive days via 24-hour recalls and was reported by mothers for children under 10 years. Measures of weight, height, triceps skinfold thickness and arm circumference were taken. The final sample included 95 children (51 boys, 44 girls; mean age at baseline; nine years). Subjects were overweight at baseline and follow-up consumed significantly more energy from fat (24%) compared to those who were not overweight (19%). Overweight children who had a high-fat diet (>30% energy from fat) were more likely to remain overweight at follow-up (P<0.01).