Risk Factors Influencing Neurodevelopmental Outcome and Survival in Preterm Infants: A Prospective Cohort Study.

BACKGROUND/

OBJECTIVES

Survival of infants born very preterm or with very low birth weight has improved without a comparable reduction in neurodevelopmental morbidity, and the exposures acting during neonatal intensive care are usually studied in isolation. We aimed to identify which prenatal, perinatal and postnatal factors are independently associated, domain by domain, with neurodevelopment at 24 months' corrected age (CA) and with the composite outcome of neurodevelopment and survival.

METHODS

Single-centre prospective cohort of 146 infants (135 assessed at 24 months' CA and 11 who died between day 8 of life and the assessment) with gestational age ≤ 32 weeks and/or birth weight ≤ 1500 g admitted to a level III unit (2018-2022). Neurodevelopment was assessed with the Bayley-III; a score < 85, rather than the conventional <70 being taken as pathological, was used, with the third edition yielding systematically higher scores. Eighty-one candidate exposures were screened; for each domain a primary composite outcome (score < 85 or death) and a secondary survivors-only outcome were analysed by exploratory multivariable logistic regression.

RESULTS

Twenty-nine of 164 eligible survivors (17.68%) were lost to follow-up and did not differ appreciably from the analysed cohort (all standardised differences ≤ 0.10). Severe intraventricular haemorrhage was associated with language, sepsis and anaemia requiring transfusion with motor, and male sex and sepsis with socio-emotional development, whereas no covariate was associated with cognition. No nutritional, metabolic or postnatal-growth variable was independently associated with any domain, and no association survived Benjamini-Hochberg control of the false discovery rate (optimism-corrected c-statistics 0.598-0.739).

CONCLUSIONS

These findings are associations rather than evidence of causality; they are hypothesis-generating and warrant multicentre external validation and longer follow-up.