Compare body fat percent (BF%) and neonatal outcomes of preterm small for gestational age (SGA) infants based on customized (SGAcust) versus population (SGApop) growth curves.
This was an observational study of 204 preterm low birthweight (<2500 g) infants. Using population-based Fenton growth curves and customized Centile Calculator-Gestation Related Optimal Weight software by Gardosi, infants were classified as SGA (weight< 10th percentile) or appropriate for gestational age (AGA). Air-displacement plethysmography (PEA POD, Life Measurement, Inc, Concord, CA) was performed at a mean adjusted gestational age of 43.0 ± 10.4 weeks to determine BF%. Neonatal outcomes, hospital length of stay (LOS) and BF% were compared using analysis of variance, Student's t-test, Chi-square and logistic regression between SGA and AGA defined by the different growth curves.
Population-based curves identified 47 SGA infants with mean BF% of 14.6 ± 6.7% compared to 15.2 ± 6.2% for the 157 AGA infants (p=0.53). Customized curves identified 109 SGA infants with a mean BF% of 14.1 ± 6.3% compared to 16.2 ± 6% for the 95 AGA infants (p=0.02). When both SGAcust and SGApop were included together in a multivariable model predicting BF%, SGAcust remained a significant predictor (p=0.02) whereas SGApop was not shown to be a significant independent predictor (p=0.50). RDS was significantly less in both SGA groups compared to AGA groups (SGApop OR = 0.43, CI [0.2-0.9]; SGAcust OR = 0.29, CI [0.1-0.6]) but was no longer significant after controlling for gestational age. LOS was not significantly different between SGApop and AGApop (p=0.78), but LOS was significantly different between SGAcust (40.9 ± 26.3 days) and AGAcust (48.9 ± 27.5 days; p=0.03).
Customized growth curves reclassified 30% of the preterm, low birthweight infants from AGA to SGA compared to a commonly used population-based growth curve developed for preterm infants. Customized growth curves better identified low infant BF%, neonatal morbidity risk, and LOS.
© 2011 Mosby, Inc. Published by Elsevier Inc. All rights reserved.