Longitudinal modelling of growth in neonates exposed to antenatal steroids to quantify associations with final height: a cohort study.

OBJECTIVE: To assess the associations of antenatal steroids with child growth. DESIGN: Longitudinal observational cohort study started in 1994. SETTING: A single tertiary neonatal centre in Sheffield, UK. PARTICIPANTS: Of 254 individuals recruited, two were excluded, 48 born at term; 202 (57% boys, 87% white ethnicity) modelled had a median of 19 height measurements each (Q1:12 to Q3:21) up to median age 15.8 years (Q1:9.9 to Q3:16.9). INTERVENTIONS: Data on administration of antenatal steroids were collected alongside gestational age and parental height. MAIN OUTCOME MEASURES: Height was modelled with SuperImposition by Translation and Rotation (SITAR) to extract each person's peak velocity and age at peak velocity via the SITAR random effects of 'size', 'timing' and 'intensity' and to predict height at 18 years. The association of each random effect and final height with exposure to antenatal steroids was assessed by multiple regression to adjust for covariates. RESULTS: In girls with covariates available (n=59/87), exposure to antenatal steroids was positively associated with SITAR 'size' and 'intensity' of growth when adjusted for gestational age, maternal and paternal height, equating to a final height 2.8 cm (95% CI 0.3 to 5.3 cm) greater than for those not exposed to antenatal steroids. In boys (n=66/115), exposure to antenatal steroids had no association with final height. CONCLUSIONS: This observational cohort study showed greater height of girls exposed to antenatal steroids not seen in boys. Analysis of existing long-term follow-up data from neonates is indicated to increase understanding of the associations of neonatal interventions on growth.