These analyses are based on previously collected data. The FCS was a longitudinal study that enrolled 106 children from two-parent families with a 3–6 year-old child in 1987. The families were third and fourth generation descendants of subjects in the original Framingham Heart Study [17]. Of the original 106 families, 100 provided dietary data for the children at baseline (preschool) and throughout follow-up (adolescence). Diet, physical activity and other lifestyle factors were evaluated annually by means of interviews, questionnaires, and clinical measurements over a period of 10 years [18,19,20].
Dietary data
Dietary data were collected annually using multiple sets of 3-day diet records. During early years of the study (prior to age 10), parents completed the diaries for the children, with input from other caregivers inside and outside of the family. A study nutritionist instructed each family in the completion of the diet record including accurate estimation of portion size. Nearly 90% of subjects completed diet records for eight or more of the 11 years in the study. Dietary data were analyzed for nutrient intake using the Nutrition Data System for Research (NDSR) of the University of Minnesota [21]. Mean servings per day at each age in the 30 USDA food groups were calculated by linking NDSR food codes with Pyramid Serving data files of the Continuing Survey of Food Intake by Individuals [22]. For our calculation of fruit juice, only 100% fruit juices and 100% juice blends such as 100% cranberry juice blends (i.e., blended with other 100% juices) were included. Part-juice beverages and tomato juice were excluded from fruit juices. Intake of whole fruit (including cut fruit) and juices are expressed as USDA-defined cup-equivalents per day. The most common types of juice beverages in the preschool years were apple and orange juices.
Outcome variables
Each child’s intake of whole and total fruit throughout childhood was examined to determine whether the child met the recommendations for fruit intake at each age. Based on Dietary Guidelines for Americans (DGA) recommendations, the following levels were considered to meet guidelines for total fruit intake: 1 cup for 2–3 year-olds, 1–1½ cups for 4–8 years, 1½ cups for 9–13 years old, and 1½ cups for 14–18 year-old girls and 2 cups for 14–18 year-old boys [3]. Diet quality was based on the 2015 Healthy Eating Index (HEI-2015) total score which is designed to measure conformance with the 2015 USDA DGAs. HEI-2015 is comprised of 13 component scores with a maximum total score of 100. Two fruit outcomes are included: whole fruit and total fruit intake. As an overall measure of diet quality, the HEI-2015 has been shown to be both reliable and valid [23].
Each child’s height and weight were recorded at each annual clinic exam. Weight (to the nearest 1/4 pound) was measured using a standard counterbalance scale, and height was measured (to the nearest 1/4 in) using a measuring bar on the same scale. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared.
Statistical analysis
Children were categorized into four age groups: preschool (3–6 years old) and three follow-up ages (7–9, 10–13, and 14–17 years old). These age groups were chosen to reflect the child’s growing level of independence with respect to food and beverage choices as well as emerging peer influence on those choices. The youngest age group includes children in preschool/kindergarten when parents exert the greatest control over food choices. The second and third age groups include early elementary school and middle school ages, respectively, while the oldest children were those in their high school years who have the greatest level of independence in food choices. For analyses related to the association between 100% fruit juice consumption and subsequent total and whole fruit intakes, preschool fruit juice intake was categorized as < 0.5 cups, 0.5- < 1.0 cups, and ≥ 1.0 cups. To increase power for some analyses, categories of juice intake were collapsed to include < 0.75 cups vs. ≥0.75 cups. Mixed linear regression models for repeated measures data was used to examine the association between juice consumption at 3–6 years of age and total and whole fruit intake as well as HEI scores throughout childhood. Logistic regression modeling was used to estimate the likelihood of meeting dietary guidelines throughout childhood and adolescence. Potential confounding by age, sex, parental education, mother’s BMI, energy intake, physical activity, and television and video viewing time was explored. Only sex was found to confound the results for dietary outcomes and thus was retained in these final models. For the BMI analysis, the final model included age, sex, maternal education, maternal BMI, physical activity, and TV and video viewing time.
