Projecting Infant Weight Gain per Day to 112 Days from First 35 Days of Life

The American Academy of Pediatrics Task Force on Clinical Testing of Infant Formulas [1] stated, “Determination of rate of gain in weight is the single most valuable component of the clinical evaluation of an infant formula” and “recommends, that weight gain be determined over an interval of 3 to 4 months.” A gastrointestinal tolerance study (denoted ST) measured infants’ weight gain per day from 14 to 35 days of age (CW2). The objectives of this project were 1) to estimate weight gain per day from 14 to 112 days of age (CW5) using the CW2 in ST and historical data from seven 4-months growth and tolerance studies by simulation and bootstrap, and 2) to determine if ST would have supported normal physical growth in infants had the study been continued to 112 days of age by showing that the experimental formula is non-inferior to the control formula using weight gain accretion of 3 g/day as the non-inferiority margin. 1. American Academy of Pediatrics, Committee on Nutrition, (CON-AAP). “Clinical Testing of Infant Formulas With Respect to Nutritional Suitability for Term Infants.” June 1988 (Report prepared under FDA contract 223-86-2117) (available at https://wayback.archive-it.org/7993/20170722090324/https://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryInformation/InfantFormula/ucm170649.htm) (accessed May 23, 2019). The simulation made use of the observed linear relationship between CW2 and CW5 from growth data in the historical studies, which included 16 formula groups. The simulated data for CW2 and CW5 were generated using multivariate normal distribution with mean vector μ and covariance matrix ∑, estimated from ST and the historical studies. In addition, the bootstrapped distribution of the difference in mean CW5 between the experimental and control groups in ST was derived using the historical studies. Software SAS® 9.4 and SAS® Enterprise Guide 7.1 (SAS Institute Inc, Cary, NC) were used for this report. Based on the analyses, we conclude that growth observed from 14 to 35 days of age in ST when extended to 112 days of age will demonstrate normal physical growth with high probability (> 98%). Using such a predictive model may be a complementary solution for bridging growth and tolerance studies of shorter duration to 4 months for minor changes to infant formulas. Abbott Laboratories.