Optimal birth weight percentile cut-offs in defining small- or large-for-gestational-age

Aims: It remains questionable what birth weight for gestational age percentile cut-offs should be used in defining clinically important poor or excessive foetal growth. We aimed to evaluate the optimal birth weight percentile cut-offs for defining small- or large-for-gestational-age (SGA or LGA). Methods: In a birth cohort-based analysis of 17 979 120 non-malformation singleton live births, U.S. 1995-2001, we assessed the optimal birth weight percentile cut-offs for defining SGA and LGA. The 25th-75th percentile group served as the reference. Primary outcomes are the risk ratios (RR) of neonatal death and low 5-min Apgar score (< 4) comparing SGA or LGA versus the reference group. More than 2-fold risk elevations were considered clinically significant. Results: The 15th birth weight cut-off already identified SGA infants at more than 2-fold risk of neonatal death at pre-term, term or post-term, except for extremely pre-term births < 28 weeks (continuous risk reductions over increasing birth weight percentiles). LGA was associated with a reduced risk of low 5-min Apgar score at pre-term, but an elevated risk at term and post-term. The 97th cut-off identified LGA infants at 2-fold risk of low 5-min Apgar at term. Conclusion: The commonly used 10th and 90th birth weight percentile cut-offs for defining SGA and LGA respectively seem largely arbitrary. The 15th and 97th percentiles may be the optimal cut-offs to define SGA and LGA respectively.