Predictive ability of postnatal growth failure for adverse feeding-related outcomes in preterm infants: an exploratory study comparing Fenton with INTERGROWTH-21st preterm growth charts

Abstract

Background Postnatal growth failure (PGF) can impact the short- and long-term health outcomes in preterm infants. However, PGF rates vary according to the way it is defined and the growth chart used to monitor the postnatal growth. Fenton-2013 growth charts which suggest following intrauterine fetal growth compared to INTERGROWTH-21st, one specifically constructed for monitoring preterm extrauterine growth. Objective Exploratory study to determine the PGF definition at first per oral (PO) that is most predictive of adverse oral feeding-related outcomes in preterm infants. Methods Prospectively collected data of preterm infants 24–32 weeks gestation, who were started on cue-based oral feeds at ≤34 weeks gestation were reviewed. Anthropometric data at first PO (weight, length, and head circumference) were compared according to Fenton and INTERGROWTH-21st growth charts. PGF was defined either as <10th percentile, Z-score change (ZSC) of ≥–1.5 from birth, or ZSC of ≥–2.0. Top-quartile (Q4) of feeding-related outcomes (days from first PO to full PO, post-menstrual age at full PO, days from first PO to discharge, and length of hospital stay) was considered as adverse outcome. Results Of the 125 infants included, the median birth gestation and weight were 29.4 weeks and 1235 g, respectively. Incidence of appropriate, small, and large for gestational age was similar at birth by both growth charts. ZSC −1.5 for weight by Fenton was significantly higher at first PO vs. INTERGROWTH-21st (p=.02), while percentile <10th and ZSC −2.0 rates were similar. The PGF definition based on individual anthropometrics at first PO that has the best area under the curve (AUC) for adverse feeding-related outcomes was used to create a combined PGF definition for each growth chart. The AUC for the combined PGF for the Fenton and INTERGROWTH-21st was similar (p>.05) and both have moderate sensitivity and negative predictive value, but have low specificity, positive predictive value, and positive likelihood ratio for adverse feeding-related outcomes. Conclusions The tested definitions of PGF at first PO have only small to moderate predictive ability for adverse feeding-related outcomes in preterm infants.