Admission Physiology Appeared More Informative Than Small-for-Gestational-Age Status for Predicting Severe In-Hospital Disposition in Preterm Infants Born at 28-36 Weeks: A Retrospective Single-Center NICU Cohort Study.

backgroundWhether small for gestational age adds prognostic information after NICU admission physiology is known remains uncertain in preterm infants. This distinction is clinically important because SGA is an anthropometric birth-size classification rather than, by itself, a diagnosis of fetal growth restriction.

methodsWe conducted a retrospective cohort study of 270 singleton infants born at 28 to 36 weeks’ gestation and admitted to a tertiary NICU in Mashhad, Iran, during 2022 to 2023. SGA was defined using INTERGROWTH-21st standards. The primary endpoint was severe unfavorable disposition, defined as transfer to another facility or in-hospital death versus healthy discharge, with discharge against medical advice excluded by endpoint definition. We compared prespecified binary logistic models containing gestational age and SGA alone or combined with admission physiologic variables, and internally validated the binary models using 1000 bootstrap resamples. A secondary logistic model described admission features associated with SGA status.

resultsSeventy-eight infants (28.9%) were classified as SGA. Compared with AGA infants, SGA infants had lower platelet counts, lower red blood cell counts, higher venous PO2 values, and longer hospital stay. In the secondary SGA model, lower platelets and lower red blood cell counts were independently associated with SGA status. For the primary endpoint, 208 infants were eligible, including 31 transfer or death events. A model containing gestational age, log-transformed C-reactive protein, and venous pH showed better fit and discrimination than the gestational age plus SGA model, with optimism-corrected area under the curve 0.78, Brier score 0.11, and calibration slope 0.92. Adding SGA did not improve fit or discrimination, with likelihood-ratio p = 0.841 and adjusted OR 1.10, 95% CI 0.43 to 2.80. Replacing binary SGA with birth-weight z score also did not improve the model. When discharge against medical advice was included in a broader nonhealthy composite, discrimination fell substantially.

conclusionIn this single-center cohort of infants born at 28–36 weeks and admitted to a tertiary NICU in Mashhad, SGA as an anthropometric birth-size classification was associated with a distinct admission hematologic profile and longer hospital stay, but it did not improve prediction of severe hospital disposition after gestational age, admission venous pH, and CRP were considered. For this short-term disposition endpoint in this setting, admission physiology appeared more informative than birth-size category. These findings are hypothesis-generating, context-specific, and should not be extrapolated to longer-term outcomes or other NICU settings without external validation.