Methods
Study Population
The study population (n=3548) consisted of infants enrolled the Illinois Special Supplemental Nutrition Program for Women, Infants and Children (WIC) program who were born between Jan 1, 2010 – Dec 31, 2010 with low birth weight (1500-2500g). Infants with less than 2 follow up visits and any weights outside normal parameters (+5 or -6 Z-score on the WHO 2006 growth standards) (n=418) were excluded from the study leaving a total of 3130 infants included in the analysis. Data were collected by the Illinois Department of Human Services' Cornerstone data management systems. The study was approved by the Institutional Review Board of Loyola University Chicago Health Sciences Campus.
Infant Formula
Formula use at baseline was determined by the first food package assigned to the infant upon WIC enrollment, which generally occurs within 2 weeks after hospital discharge. Formula assignments were not randomized but determined by infant’s pediatrician. Parents provided a medical prescription to the WIC program to receive nutrient enriched infant formulas for their infants. Nutrient enriched infant formulas included in this study were SimilacNeosure by Abbott Nutrition and Enfamil Enfacare by Mead Johnson. Standard term infant formulas included in this study were Enfamil Infant, Enfamil Prosobee, and Enfamil Gentlease by Mead Johnson. Table 1 provides detailed nutritional information for these formulas.
Infant Weight
The baseline infant weights were generally measured within 2 weeks after birth and at every subsequent follow-up WIC visit (every 6 months) with an average of 2.5 visits per infant within the first 13 months of life. Infant weight was measured using WIC anthropometric guidelines at WIC offices or at a pediatrician’s office.
Covariates
Demographic information was collected by WIC personnel trained in motivational interviewing to ensure the accuracy of the information collection. Information collected during initial WIC interview included gender and race/ethnicity, which were self-reported by the infant’s parent or legal guardian at the infant’s initial WIC visit. Infants’ due dates were verified with mother’s obstetrician as part of her proof of pregnancy eligibility with the WIC program. Mother’s maternal age was calculated based on the mother’s date of birth, which is verified with her identification card, passport, Illinois Medicaid card or acceptable documentation per WIC guidelines. Mother’s pre-pregnancy BMI and mother’s education level were collected during the mother’s initial prenatal WIC interview, if the mother was enrolled in the WIC program during pregnancy. If the mother was not enrolled in WIC before the infant was born, this information was collected during the infant’s initial WIC visit using the same methods. Breastfeeding status and household smoking status were collected by WIC personnel by interviewing the infants’ mothers or legal guardian during every WIC follow up visit. For this study, breastfeeding was coded as a categorical variable (yes/no) with breastfeeding considered as “yes” if the infants were ever breastfed. Infants with records of any household smoking at any WIC visit were coded as being exposed to a household smoker.
Statistical Analysis
The statistical analysis was performed using STATA software (StataCorp. 2011. Stata Statistical Software: Release 12. College Station, TX: StataCorp LP). Characteristics of infants by use of nutrient enriched formulas were compared. Categorical variables were compared using a Chi-square test and continuous variables were compared using an unpaired t-test. A multilevel mixed-effects linear regression model with random intercepts (baseline weight) was used to determine the association between infant formulas (nutrient enriched vs. standard term) used at baseline and weight gain. To account for differences in weight gain by birthweight we stratified analyses by baseline birth weight, (1500-1999g and 2000-2500g) and adjusted for birth weight in the regression model. Given the known differences in infant growth during the first 6 months vs. 6-12 months, the analyses was also stratified by age (0-6 months and 6-12 months) to satisfy the assumption of a linear association between time and weight gain.
We examined three different models in order to examine the potential impact of covariates on the association between nutrient enriched formula and infant weight gain. Model 1 examined the unadjusted association between nutrient enriched formula and infant weight gain. Model 2 added birth weight, gender and gestational age to Model 1. Model 3 added history of breastfeeding, exposure to household smoking, mother’s age, mother’s pre-pregnancy BMI, mother’s education level, and infant’s race and ethnicity to Model 2. In all models, infant’s weight, age in months, gestational age, mother’s age, mother’s pre-pregnancy BMI, and mother’s education were fitted as continuous variables, whereas formula use, gender, history of breastfeeding, exposure to household smoking, and race/ethnicity were fitted as categorical variables.