Factors related to adiposity among children aged 3 to 7 years


Author: Robertson, Shay M; Cullen, Karen W Baranowski, Janice Baranowski, Tom et al
Source: American Dietetic Association. Journal 99, no. 8 (Aug 1999): p. 938-943
ISSN: 0002-8223
Number: 04526985
Copyright: Copyright American Dietetic Association 1999

Objective To compare diet and physical activity between a group of children aged 4 to 7 years who had increased their sum of 7 skinfolds by 1.5 standard deviations or more since the previous year and those who had not.

Subjects/design A longitudinal design was used wherein children had their body composition assessed at 4 annual intervals. Fifteen study subjects were identified from a larger study on the development of cardiovascular risk factors in children. Three matched control subjects were identified for all but 4 study subjects. Children were included if they were 3 or 4 years old at the time of the first of 4 annual clinics to collect data. Children were volunteers from a mid-sized city.

Measures Seven skinfold sites were assessed, each 3 times.

An average was taken of the mean at each site. Diet and physical activity were assessed using observational methods for up to 4 days for each of 3 years between the 4 annual clinics (held in the summers of 1986, 1987, 1988, and 1989).

Results Children consumed significantly (P=.02) more fat grams and suggestively higher percentages of energy from fat (P=.06), total energy (P=.08), and percentage energy from protein (P=.10). No differences were detected for percent energy from carbohydrate, physical activity, or height.

Conclusions Programs to prevent childhood obesity might have success by targeting dietary fat consumption among children as young as 4 years old, but further research is needed. JAm Diet Assoc. 1999;99:938-943.

Obesity is a growing public health concern among children in the United States (1). Childhood obesity may be influenced by distinct growth periods during early childhood. The first period occurs during the first year of life with rapid growth in size but a stable number of adipose cells. Fat cell size declines over the next 1 to 2 years and remains stable for several years (2,3). Length increases as the body undergoes an apparent slimming process. The second phase, termed adiposity rebound (4), is characterized by rapid growth in body fat, which usually begins about age 6 years, and includes increases in both cell size and cell number (2,3). Linear increases in body fat and percent body fat occur from about 2 to 14 years, with a substantial increase in the variability of these measures at about 5 or 6 years, especially among minority girls (5). Children who experience adiposity rebound at an earlier age are more likely to have higher adiposity at age 14 years (2,3), and obese adolescents have a relative risk of 5.3 to 6.7 of remaining obese as young adults (6).

Little is known about the influences on adiposity during the period leading up to rebound (7). One possible contributing factor is a low level of physical activity, but substantial error in measurement of physical activity attenuating relationships with adiposity has been noted (8). Changes in triceps skinfold from the age of 3 years to about 6 or 7 years were related to physical activity as measured by the single plane accelerometer type activity monitor (9). Increases in weight from the age of 3 to 5 years were marginally related (P<.10) to physical activity as measured by observation for limited periods during the day (10); and with 1 additional year of data, baseline activity was related to body mass index (BMI), but change in leisure activity was only marginally related (11). Children with high physical activity as measured by self-report had the same weights and BMIs as children reporting low activity, but lower percent arm fat (from a circumference measure) and a later time of adiposity rebound (6.1 years vs 5.4 years) (12). Activity-related energy expenditure using the doubly labeled water method was not related to body fat among 5-year-old Pima Indians or among majority group (Euro-American) children (13), nor was it related to change in fat mass among majority group children aged 5 to 8 years (14), nor to fat mass among children aged 4 to 7 years (15). Number of minutes engaged in activity using a heart rate method was not related to percent body fat, but number of minutes engaged in sedentary activity were (16). Among a large sample of slightly older girls (9- to 10year-olds), activity measured from a questionnaire was marginally (P=.07) related to BMI and to sum of skinfold thicknesses (P=.09) among African-American girls, but not EuroAmerican girls (17). Among boys and girls aged 9 to 11 years, total energy expenditure, resting energy expenditure, and physical activity were substantially higher among obese children than nonobese children, but total energy expenditure per kilogram body weight was lower (18). Regression analyses controlling for dietary variables revealed unexpected positive relationships between physical activity and percent body fat (18). Thus, the evidence for a relationship between physical activity and body fat among children is mixed, with measures that reveal no relationship considered to be more accurate.

Another possible contributing factor is diet. Among 3- or 4year-olds, percent of energy from fat, but not total energy were higher in the high-risk group that gained weight over a 1-year period (10). In the same study, baseline percent of energy from fat and change in percent of energy from fat predicted change in BMI over a 2-year period (11). Children with low activity levels who had greater percent arm fat ate a diet with a higher percent fat (12). Among 4- to 7-year-olds, energy intake from dietary fat was a significant predictor of fat mass among boys, but not girls, even after controlling for total energy intake (15). Among 9- and 10-year-old African-American girls, percent of energy from saturated fat was a significant predictor of BMI, but not of sum of skinfold thicknesses, whereas among EuroAmerican girls percent of energy from total fat was a predictor of BMI and a sum of skinfold thicknesses (17). Among 9- and 10-year-old boys and girls, percent of energy from dietary fat significantly increased across 3 body fat categories, especially after controlling for gender and total energy intake, whereas percent of energy from carbohydrate significantly declined across the 3 categories (19). Similarly, among 9- to 11-year-old boys and girls, percent of energy from fat was significantly higher among obese children and was significantly correlated with percent body fat, whereas percent of energy from carbohydrate was significantly lower among obese children and significantly negatively correlated with percent body fat (18). Thus, dietary fat appears to lead to higher body weight or fat among children at or near the time of adiposity rebound, but some controversy exists concerning the effect of physical activity. Most of this research, however, was cross-sectional in design and used mostly self-report measures, thereby imposing limits on possible inferences.

This study assessed differences in dietary intake and physical activity between 2 groups of preschool children in a longitudinal design using observational measures of diet and physical activity. Children experiencing adiposity take-off, defined as children whose adiposity increased 1.5 standard deviations or more above the mean from the previous year, were compared with a matched group of children whose adiposity did not increase for both the year of the adiposity take-off and the year before.

METHODS

Participants

Participants were 3- or 4-year-old children who were enrolled in a 4-year longitudinal investigation of the development of cardiovascular disease risk factors and related behaviors at the Texas site of the Studies of Child Activity and Nutrition (SCAN), an 8-site study funded by the National Heart, Lung, and Blood Institute (Bethesda, Md) (20). The study was approved by the Institutional Review Board of the University of Texas Medical Branch, and informed consent was obtained from the primary guardian of each child. Participants were recruited by announcements, presentations, home visits, social networking, and telephone calls to parents. Exclusionary criteria included mental retardation or other developmental disabilities, history of a chronic illness affecting diet or exercise habits in an immediate family member, no English-speaking parent, and no parent residing in the household. Only one child per family was enrolled in the study. The original data set contained 310 participants. Thirty-five subjects were excluded from analysis because of missing information about age, gender, and/or ethnicity; 9 were excluded because they were the second child listed for a family; and 141 participants were omitted because 2 or more years (out of 4) of data collection were missing. The final sample included 125 subjects: 107 subjects with all 4 years of data and 18 subjects with 3 years of data. The total sample with at least 3 data points across the 4 years of the study included 60 boys and 65 girls; 41 white, 33 Mexican-American, and 51 Africa99, no. 8 (Aug 1999): p. 938-943n-American children. A subsample of these participants who experienced substantial gain in their sum of 7 skinfolds between annual clinics was selected as study subjects with matched control subjects. There were no statistically significant differences between participants included in the longitudinal subsample vs the others on these demographic characteristics.

(Graph Omitted)

Captioned as: FIG 1.

(Graph Omitted)

Captioned as: FIG 2.

Design

This study used a longitudinal design with no intervention. Children and their parents living at home were invited to an annual clinic for 4 consecutive summers. Anthropometric measures were obtained during the annual clinics. Between annual clinics, each child was followed up by paid observers who met the child at 7 AM, or when the child usually awoke, and continued until dinner was finished or 7 Pm. Observations were conducted between annual clinics, with an attempt to space observations through the seasons of the year. Observational measures included level of physical activity, food intake, and other related variables not included in these analyses.

Measurement

Sum of 7 skinfolds Since adiposity in childhood is primarily subcutaneous, 7 skinfold sites (triceps, biceps, subscapula, abdominal, supraiac, thigh, and calf) adequately measure adiposity (21). Three measurements were taken at each of the 7 skinfold sites using standard procedures (22,23) and averaged. These site averages were summed each year for each child.

Identifying cases of adiposity take-off Graphical displays revealed that the sum of 7 skinfolds of most children grouped closely together over the 4 years, suggesting no increases or slight decreases in adiposity for most children (see Figure 1). A small subset of children, however, appeared to experience a take-off, that is, large increases in subcutaneous adipose tissue at varying years of the study, and never returned to the group of clustered values (see Figure 2). The children who experienced take-off were the ones we studied for purposes of these analyses. Take-off was defined as SSS greater than or equal to 1.5 standard deviations above the mean for the previous year's value. One and a half standard deviations is the 12th percentile (from a normal distribution), which is more conservative but similar to obesity commonly defined as greater than or equal to the 85th percentile of skinfold measurements. Only the first period of take-off per participant was used in the analyses. Participants who were not experiencing take-off were matched to participants who were experiencing take-off on age, ethnicity, and gender for the same year of data collection, and 3 were randomly selected as control subjects per case. No control subjects were identified for 4 study subjects (all were MexicanAmerican). One subject was lost to the analysis of data for the year before take-off because no observation data were collected that year for him. The demographic characteristics of the children experiencing take-off and the control subjects are in Table 1.

Dietary intake Food intake data were collected by 11 trained observers who worked in pairs and alternated observing the child every 2 hours. Each child was observed at approximately 3-month intervals to reflect seasonal variation of dietary intake. Any dietary intake before arrival of the team was recorded from parental report. The observers followed the children wherever they went, except when they used the bathroom or went to their personal bedroom and during any time that they were in a school classroom (where no meaningful physical activity or dietary intake was likely to occur). A trained registered dietitian called the parent the following day and asked what the child ate after the observers left. The dietary information was transferred to a standardized 24-hour recall form and sent to the Nutrition Coordinating Center at the University of Minnesota for nutrient analysis (24). The nutrient variables of interest for these analyses were energy, total fat, carbohydrate and protein intake, and percent of energy from the macronutrients, For each subject, an average intake across days observed was used to represent average intake of nutrients for that year.

Physical activity The same observers were trained to record physical activity in a variety of settings, including home, day care, and school, using the Children's Activity Rating Scale (25). The 5 levels with representative activities were

1=stationary, no movement (eg, lying, sitting);

2=stationary, with movement (eg, standing/coloring, standing/ ball activity);

3=translocation, slow/easy walk;

4=translocation, medium/moderate walk; and

5=translocation, fast, very fast, or strenuous walk.

Children's physical activity was observed for up to 4 days each year (at the same times as diet observation). Continuous minute-by-minute ratings of physical activity levels were recorded into portable computers. After a prompt from the computer at the start of each minute, an initial rating was recorded of the child's activity at that moment and any subsequent activity level changes were recorded during that minute and averaged to obtain an activity score for the minute. Each activity level could be coded only once each minute with a maximum of 5 levels within a given minute.

For this analysis, the time interval from 3 Pm to 6 PM was selected to represent level of physical activity. This interval generally reflects when children are voluntarily active or inactive (26), rather than under the influence of schools or other programs. Participants with less than 60 minutes of observation within this identified time frame were excluded, resulting in the loss of 1 control day in the analysis in the third year. The average minutes of observation from 3 to 6 PM for the study subjects were 90.3 (+/-29.6) minutes in year 1; 120.5 (+/-35.5) minutes in year 2; and 118.8 (+/-40.7) minutes in year 3. The corresponding values for control subjects were 98.1 (+/-38.9) minutes in year 1; 102.5 (+/-30.3) minutes in year 2; and 109.3 (+/-43.6) minutes in year 3.

(Table Omitted)

Captioned as: Table 1

The first index of activity was the average of Children's Activity Rating Scale categories across minutes. A second index was the percent of minutes within the 3-hour window that activity was recorded at levels 3, 4, or 5. In this index, the minute was included in the numerator if it had a 3, 4, or a 5 recorded, even if a level 1 or 2 was recorded in a minute. The denominator was the number of minutes in the interval for which some observation of activity (at any level) was recorded. This index reflects the percent of time there was any moderate to vigorous movement of the participant (eg, walking or running). A third index was the percent of minutes that activity was recorded only at level 4 or 5. The denominator was the same as used for activity index 2. This index reflected the percent of time that movement was only vigorous (eg, running or climbing stairs fast), suggesting activity with potential for aerobic benefit.

Analyses

Intraclass correlations (the ratio of within-person to between-- person variability) for a single value and adjusted for the mean of multiple values (an index of reliability) were calculated across multiple days within a year (27) for the dietary intakes and physical activity. The internal consistency form of reliability of the SSS was assessed with Cronbach's ot. The hypotheses were tested using 2 methods: one-way analysis of variance (ANOVA) between study subjects and control subjects and one-way mixed model ANOVA, because the control subjects were matched to the study subjects on gender, ethnicity, and age. In the latter analyses, a random effect term corresponding to the matched group (1 study subject, 3 control subjects) was added to the ANOVA models to account for the correlation among subjects due to matching. The mixed model ANOVAs were conducted using PROC MIXED in SAS (SAS Institute Inc, Cary, NC), which allows for missing data and unbalanced designs; thus, all study subjects, including those without matching control subjects, were included in the analyses. The hypotheses were tested at an a .05, but due to small sample size, probabilities between .05 and.10, were considered suggestive. All data were analyzed using the Statistical Package for Social Sciences (SPSS Inc, Chicago, Ill).

RESULTS

Participant Characteristics

Sixteen study subjects were identified using the 1.5 standard deviation criterion for take-off; however, one child had experienced take-off on 2 occasions, of which only the first was used in analysis. Of the possible annual periods of adiposity take-off, 1 study subject was identified in the first period, 5 in the second, and 10 in the third. Study subjects had an average (standard deviation) of 2.9 (+/-0.8) days of assessment in year 1, 3.2 (+/-0.9) days in year 2, and 2.5 (+/-0.6) days in year 3. Control subjects had an average of 2.8 (+/-0.9) days of assessment in the first year, 3.0 (+/-0.8) days in the second year, and 2.4 (+/-0.8) days in the third year.

Reliability Coefficients

The intraclass correlations across multiple assessments of skinfolds at 1 site within a year were uniformly high, ranging from .986 to .997 for the first year's clinic; .964 to .999 for the second year's clinic; .991 to .999 for the third year's clinic; and .990 to .999 for the fourth year's clinic. Cronbach's alpha applied to the 7 skinfold measurements by year revealed values ranging from .93 to .97, indicating substantial internal consistency among the skinfolds. Thus, we can have high confidence in the consistency of the skinfold assessments, which lends confidence to the assessment of take-offs.

The intraclass correlations for reliability (adjusted to reflect multiple assessments) for dietary intake and physical activity were moderate for the first year of observation (range from .518 to .695), but got progressively lower over the 3 years with a more severe decline for physical activity. This suggests increasing day-to-day variability in diet and physical activity across years of assessment.

ANOVA

Significant differences between study subjects and control subjects were found for fat grams (P=.02), and suggestive differences were found for percentage of energy from fat (P=.06), total energy (P=.08), and percent energy from protein (P=.10). Percent of energy from carbohydrate, physical activity, and height, however, were not significantly different (Table 2). Virtually identical results (to 2 decimal places) were obtained with the procedure that controlled for matching and thereby are not included here.

(Table Omitted)

Captioned as: Table 2

Analyses were conducted to assess the role of diet and/or physical activity in the year before take-off. For the single study subject from the first year, no data were available for the year before take-off; thus, this subject was eliminated from analyses of differences in the prior year. Significant differences were found for total energy (P=.01) and total fat grams (P=.02), but not for fat as a percent of energy, physical activity, or height for the year before take-off (Table 3).

DISCUSSION

Children experiencing an adiposity take-off (an increase of 1.5 standard deviation in the sum of 7 skinfolds in a year) from 3 to 7 years old were determined to consume more energy from fat and, suggestively, more total energy, higher percent of energy from fat, and higher percent of energy from protein but they did not engage in less physical activity than, nor did they differ in height from, the control subjects. In regard to dietary fat, these findings are consistent with most studies in the literature (10,11,15,17-19,28). The higher consumption of protein as a percent of energy is consistent with one other longitudinal study (4), but not 2 cross-sectional studies (18,19). During the year of take-off, both samples were under the average energy allowance (about 1,800 kcal/day) (29), but substantially above the Recommended Dietary Allowance (29) for protein (24 g/day). The relatively small sample in this study suggests that the effect of dietary fat is strong, but cannot rule out differences between groups in other macronutrients because power may not have been adequate to detect weaker effects. The relatively high intraindividual variability in assessment also limited the ability of these measures to detect such a relationship. The number of days of dietary assessment necessary to reliably estimate macronutrients among children this age has not been determined.

None of the 3 physical activity indexes displayed differences between study subjects and control subjects. This finding is consistent with the results of 3 studies (14-16) and inconsistent with the results of one study (9); one study found a relationship with one ethnic group, but not another (17); and another study found a positive relationship between physical activity and percent body fat (18). Inadequacies in measurement (8) have made it difficult to detect differences in physical activity between obese and nonobese youth. Our study may not have had sufficient days of assessment to overcome intraindividual variability (30). In addition, the limited time interval (3 to 6 Pm), although clearly reflecting volitional physical activity, may not adequately reflect overall physical activity. Alternatively, physical activity levels may be uniformly so low among all children that factors other than physical activity determine who becomes obese. Further research should investigate more comprehensive approaches to assessing activity.

Total energy and total grams of fat were statistically different between groups in the year before the adiposity take-off, indicating that children with high adiposity had been consuming higher fat and energy-dense foods for some time. This suggests that the children were chronically overfed (10) or the children had higher preferences for high-fat foods at early ages (31) and it took at least an additional year for the excess consumption to manifest in greater adiposity. These possibilities require further research. During the year before take-off, both samples were under the average energy allowance (1,800 kcal/day), but substantially above the Recommended Dietary Allowance for protein (24 g/day) (29).

Most cases of take-off occurring in the third year of observation (children aged 5 to 7 years) is consistent with the timing of adiposity rebound (4). The general decline in the SSS across the 4 annual assessments (Figure 1) among children not experiencing adiposity take-off is consistent with the slimming effect just before adiposity rebound (4). Visual inspection of other graphs of the full sample (not presented here) suggest that the pattern of slimming may vary by ethnic and gender groups. The role of diet and physical activity in adiposity takeoff among boys and girls in different ethnic groups is worthy of additional research with larger samples.

Among adults, dietary fat has been thought to contribute to obesity (32) through several possible mechanisms, including lack of auto-regulation of fat intake (33), increased consumption of high-fat foods because of better taste (32), and low satiety from fat (34). Future research must address the possible contribution of each of these factors to the influence of dietary fat on adiposity among children. It would also be important to document how young this overconsumption of fat leading to obesity occurs. Furthermore, it would be important to separate the possible effects of eating at restaurants from the effects of meals and snacks served at home. The relative contribution of psychological variables (eg, child preferences, outcome expectancies) and parent practices (eg, food purchasing, preparation, and socialization practices) to children's consumption of high-fat foods are important to more clearly target future interventions.

(Table Omitted)

Captioned as: Table 3

Several strengths of this research include the prospective design, the non-self-report nature of the behavioral variables, and the assessment of multiple skinfold sites to enhance reliability of estimation of body fat. Future research should be designed to study the phenomenon of adiposity take-off and include a larger sample and more days of behavioral assessment per year. Our findings must be interpreted cautiously.

APPLICATIONS

Programs to prevent obesity might have success by targeting dietary fat consumption among children as young as 4 years old, but further research is needed.

Footnote:

This research was conducted for a master's thesis by

S. M. Robertson with Texas Woman's University, Denton. The data for this study were collected under National Institutes of Health (NIH) grant NL-35131 to

T Baranowski. The authors were supported during the writing of this paper by NIH grants NL-4 7618 (TW), CA61596 (Gimme 5), and CA-73503 (UBS) to

T. Baranowski.

Footnote:

The authors appreciate the contributions of Linda Cashman, MS, RD, Doris Wright, PhD, and Karen Moreland, MS, RD, faculty in the Department of Nutrition at Texas Woman's University, Denton.

Reference:

References

Reference:

1. Troiano RP, Flegal KM, Kuczmarski RJ, Campbell SM, Johnson CL. Overweight prevalence and trends for children and adolescents. Arch Pediatr Adolesc Med. 1995;149:1085-1091.

2. Hagar A, Sjostrom L, Arvidsson B, Bjorntorp P, Smith U. Body fat and adipose tissue cellularity in infants: a longitudinal study. Metabolism. 1977;26:607-614.

Reference:

3. Knittle JL, Timmers K, Ginsberg-Fellner F, Brown RE, Katz DP. The growth of adipose tissue in children and adolescents. Crosssectional and longitudinal studies of adipose cell number and size. J Clin Invest. 1979;63:239-246.

Reference:

4. Rolland-Cachera MF, Deheeger M, Bellisle F, Sempe M, Guilloud-- Bataille M, Patois E. Adiposity rebound in children: a simple indicator for predicting obesity. Am J Clin Nutr. 1984;39:129-135.

5. Ellis KJ, Abrams SA, Wong WW. Body composition of a young multiethnic female population. Am J Clin Nutr. 1997;65:724-731.

6. Kotari K, Nishida M, Yamashita S, Funahashi T, Fujioka S, Tokunaga K, Ishikawa K, Tami S, Matsuzawa Y. Two decades of annual medical examinations in Japanese obese children-do obese children grow into obese adults? Int J Obes. 1997;21:912-921.

7. Dietz WH. Periods of risk in childhood for the development of adult obesity-what do we need to learn? JNutr. 1997;127(suppl):1884S1886&

Reference:

8. Bar-Or 0, Baranowski T. Physical activity, adiposity and obesity among adolescents. Pediatr Exerc Sci. 1994;6:348-360.

9. Moore LL, Nguyen U-SDT, Rothman KJ, Cupples LA, Ellison RC. Preschool physical activity level and change in body fatness in young children. The Framingham Children's Study. Am J Epidemiol. 1995;142:982-988.

Reference:

10. Eck LH, Klesges RC, Hanson CL, Slaw-son D. Children at familial risk for obesity: an examination of dietary intake, physical activity and weight status. Int J Obes. 1992;16:71-78.

11. Klesges RC, Klesges LM, Eck LH, Shelton ML. A longitudinal analysis of accelerated weight gain in preschool children. Pediatrics. 1995;95:126-130.

Reference:

12. Deheeger M, Rolland-Cachera M, Fontvieille A. Physical activity and body composition in 10-year-old French children: linkages with nutritional intake? Int J Obes. 1997;21:372-379.

13. Salbe AD, Fontvielle AM, Harper IT, Ravussin E. Low levels of physical activity in 5-year-old children. JPediatr. 1997;131:423-429. 14. Goran MI, Shewchuk R, Gower BA, Nagy TR, Carpenter WH, Johnson RK. Longitudinal changes in fatness in white children: no effect of childhood energy expenditure.Am JClin Nutr. 1998;67:309316.

Reference:

15. Maffeis C, Zaffariello M, Schutz Y. Relationship between physical inactivity and adiposity in prepubescent boys. JPediatr. 1997;131:288292.

16. Nguyen VT, Larson DE, Johnson RK, Goran MI. Fat intake and adiposity in children of lean and obese parents. Am J Clin Nutr. 1996; 63:507-513.

Reference:

17. Obarzanek E, Schreiber GB, Crawford PB, Goldman SR, Barrier PM, Frederick MM, Lakatos E. Energy intake and physical activity in relation to indexes of body fat: the National Heart, Lung, and Blood Institute Growth and Health Study. Am J Clin Nutr. 1994;60:15-22. 18. Gazzaniga JM, Burns TL. Relationship between diet composition and body fatness, with adjustment for resting energy expenditure and physical activity, in pre-adolescent children. Am J Clin Nutr. 1993; 58:21-28.

19. Tucker LA, Seljaas GT, Hager RL. Body fat percentage of children varies according to their diet composition. JAm Diet Assoc. 1997;97: 981-986.

Reference:

20. Baranowski T, Stone R, Klesges RC, Basch C, Ellison RC, Iannottii R, Kotchen JM, Nader PR, Strong WB. Studies of child activity and nutrition (SCAN): longitudinal research on CVD risk factors and CVH in your children. Cardiovascular Risk Factors. 1993;2:4-16.

21. Power C, Lake J, Cole T. Measurement and long-term health risks of child and adolescent fatness. Int J Obes. 1997;21:507-526.

22. Lohman TG, Roche AF, Martorell R. Anthropometric Standardization Reference Manual. Champaign, Ill: Human Kinetics Books; 1988.

Reference:

23. Greaves KA, Puhl J, Baranowski T, Gruben D, Seale D. Ethnic differences in anthropometric characteristics of young children and their parents. Hum Biol. 1989;61:459-477.

24. Feskanich D, Sielaff B, Chong K, Buzzard M. Computerized collection and analysis of dietary intake information. Computer Meth Programs Biomed. 1989;30:47-57.

25. Puhl J, Greaves K, Hoyt M, Baranowski T. Children's Activity Rating Scale (CARS): Description and calibration. Res Q Exerc Sport. 1990;61:26-36.

Reference:

26. Pate RR, Baranowski T, Dowda M, Trost S. Tracking of physical activity in young children. Med Sci Sports Exerc. 1996;28:92-96. 27. McGraw K, Wong S. Forming inferences about some intraclass correlation coefficients. Psychol Methods. 1996;1:30-46.

28. Muecke L, Simons-Morton B, Huang I, Parcel G. Is childhood obesity associated with high-fat foods and low physical activity. J Sch Health. 1992;62:19-23.

Reference:

29. Food and Nutrition Board. Recommended Dietary Allowances. 10th ed. Washington, DC: National Academy Press; 1989.

30. DuRant RH, Baranowski T, Puhl J, Rhodes T, Davis H, Greaves KA, Thompson WO. Evaluation of the Children's Activity Rating Scale (CARS) in Young Children. Med Sci Sports Exerc. 1993;25:1415-1421. 31. Fisher JO, Birch LL. Fat preferences and fat consumption of 3- to 5-year-old children are related to parental adiposity. JAm Diet Assoc. 1995;95:759-764.

Reference:

32. Ravussin E, Tataranni PA. Dietary fat and human obesity. JAm Diet Assoc. 1997;97(7 suppl):S42-S46.

33. Flatt J, Ravussin E, Acheson K. Effects of dietary fat on postprandial substrate oxidation and on carbohydrate and fat balances. J Clin Invest. 1985;76:1019-1024.

34. Blundell JE, MacDiarmid JI. Fat as a risk factor for overconsumption: satiation, satiety, and patterns of eating. JAm Diet Assoc. 1997; 97(suppl):S63-S69.

Author Affiliation:

S. M. Robertson is a dietitian at the Bay Area Rehabilitation Center in Baytown, Tex. C. de Moor is an assistant professor of biostatistics in the Departments of Behavioral Science and Mathematics, K W Cullen is an assistant professor, J Baranowski is a project director, T. Baranowski is a professor, and S. Hu is a research associate in the Department of Behavioral Science at the University of Texas M. D. Anderson Cancer Center in Houston. At the time of the study, J Baranowski and

T Baranowski were with the Department of Preventive Medicine and Community Health of the University of Texas Medical Branch, Galveston.

Author Affiliation: