The aims of the study were to re-examine existing longitudinal data to:

1. Determine differences in beverage consumption from baseline to Year Two follow-up in all subjects and based on body mass index (BMI) Z-scores
2. Examine the relationship between change in milk consumption and change in sugar-sweetened beverage consumption
3. Identify predictors of BMI Z-score at Year Two.
   

• Data from a convenience sample of elementary school children in Grades Three through Six, who had participated in a previous study
• Fall 1992 to spring 1996
• Consisted of only subjects (N=166) who were measured at baseline and again two years later.

Recruitment

• The research protocol was approved by the University of Nebraska-Kearney Institutional Board. Subjects and parents signed informed assent or consent prior to any data collection.  

Design

• Longitudinal cohort.

Statistical Analysis

Means and SD were calculated for all variables. Subjects were categorized into four groups based on BMI Z-score at baseline and Year Two.

1. Normal weight: BMI Z-score <1.0 at baseline and Year Three (N=99)
2. Overweight: BMI Z-score ≥1.0 at both baseline and Year Two (N=48)
3. Gained weight: BMI Z-score <1.0 at baseline and a BMI Z-score ≥ Year Two (N=11)
4. Lost weight: BMI Z-score ≥1.0 at baseline and a BMI Z-score of <1.0 at Year Two (N=6).

No gender differences in consumption of any type of beverage at baseline, Year Two or change from baseline to Year Two were found. Therefore, results are represented for boys and girls combined. Paired T-tests, repeated measures ANOVA, Pearson product and regression analyses used for analyses.

Timing of Measurements

• Beverage consumption: A 24-hour recall, was used to determine total caloric intake and beverage consumption at baseline and Year Two. Diet recalls represented intakes only on school days. Procedure included interviewing the subject twice during the 24-hour period.
• First interview: Occurred at noon following the lunch meal. Subjects were interviewed to obtain all foods eaten from the morning through the noon meal. To aid in the recall, a sample meal tray was available for reference. A breakfast tray was also available for those who participated in the school breakfast.
• Second interview: Occurred the following morning prior to the start of school. Subjects were interviewed to obtain all foods eaten since the previous noon meal. Three-dimensional food models were used to aid in the recall.
• For a random subsample of subjects, parents were called to verify foods and beverages consumed at home during the 24-hour recall period.
• Caloric intake was quantified using the Nutrtionist IV diet analysis program. Reliability was established repeatedly between the research nutritionists that conducted the 24-hour recalls and the interclass coefficient was always >0.95.
• All beverages were reported in fluid ounces. Beverages were classified as follows:
• Milk (skim, 1%, 2%, whole, chocolate, milkshakes)
• 100% juice
• Diet soda
• Sugar-sweetened drinks (regular soda, HI-C, sports drinks, Kool-Aid, fruit-flavored drinks, iced tea, hot chocolate).

Body Mass Index Z-Score

• Subjects were weighed to the nearest 0.1kg and height was measured without shoes to the nearest 0.1cm once at baseline and Year Two. These scores were used to compute age- and gender-specific Z-scores of BMI using the EPIInfo program.
• Subjects were classified as "overweight" if the calculated age and gender-specific BMI Z-score was ≥1.0 and classified as "normal weight" if the calculated age- and gender-specific BMI Z-score was <1.0.

• Initial N: Total eligible sample was about 830 children from Grades Three through Five.
• N=166: 92 girls and 74 boys.
• Attrition (final N): Two subjects had outlier values (>2.5SD) on the change in BMI Z-score variable and were excluded from the analysis.
• Age: Grades Three through Five
• Ethnicity: About 94% Caucasian.

Physical Characteristics

• Normal weight: 60% (N=99) 
• Overweight: 29.3% (N=48)
• Gained weight: 6.7% (N=11)
• Lost weight: 3.7% (N=6).

Paired T-test to determine differences between baseline and Year Two beverage consumption in all subjects, categorized by BMI Z-score, showed a significant decrease in milk (ounces and percentage) and total caloric intake and a significant increase in diet soda (ounces and percentage) in all subjects and normal-weight subjects.

Beverage Consumption (daily ounces and percentage) and Total Caloric Intake (daily kcal) of All Normal Weight, Overweight, Gained Weight and Lost Weight Subjects at a Baseline, Year Two and Change from Baseline to Year Two

• Subjects who were overweight significantly increased diet soda consumption (ounces). Subjects who gained weight did not significantly change beverage consumption. However, diet soda consumption (percentage) approached significance (P=0.058), subjects who lost weight (N=6) did not significantly change beverage consumption or total caloric intake.
• Repeated measures ANOVA to determine diffferences in beverage consumption from baseline to Year Two among subjects categorized by BMI Z-score group revealed no significant differences for milk, 100% juice, sugar-sweetened beverage consumption or total caloric intake.
• Change in milk consumption was inversely correlated with sugar-sweetened beverage consumption.
• Significant increase from baseline to Year Two in diet soda consumption was found for all subjects and subjects who were overweight and subjects who gained weight had significantly higher soda consumption than normal-weight subjects.
• Regresssion analyses indicated that baseline BMI Z-score and Year Two diet soda consumption accounted for 83.1% of the variance in Year Two BMI Z-scores (model, R²=0.83; P<0.001)
• Higher baseline BMI Z-scores and greater consumption of diet soda at Year Two were associated with a higher Year Two BMI Z-score.

• This study is a longitudinal design and little data exists reporting associations between beverage consumption and BMI in school-age girls and boys.
• Shifts in beverage consumption were found in this convenience sample across two years. Diet soda consumption was the only type of beverage associated with the two-year BMI Z-score and consumption was greater in subjects who were overweight and subjects who gained weight, as compared to normal-weight subjects at two years.
• Additional longitudinal data examining associations between beverage consumption and BMI is needed in children and adolescents, as consumption of regular soda and diet soda has become more of a social norm.

Limitations

• Observational from previously collected data in a small convenience sample of school-age children.
• Sample was mainly Caucausian and from one rural state, thus generalizability of the study is limited.
• Data collected for only one school day and did not include weekend days
• Data collected from fall 1992 through spring 1996 and thus prevents comparisons to beverage consumption intake of children since 1996.