• Alina Rodriguez, Yingbo Wang, Anohki Ali Khan, Rufus Cartwright, Mika Gissler, and Marjo-Riitta Järvelin.
• Department of Epidemiology and Biostatistics, Imperial College London, London, United Kingdom.
• PLoS Med. 2019 Feb 1; 16 (2): e1002746e1002746.

BackgroundAntenatal corticosteroid therapy (ACT) is used clinically to prepare the fetal lung for impending preterm birth, but animal and human studies link corticosteroids to smaller birth size. Whether ACT is associated with birth size is debated; therefore, we assessed differences in birth size in treated versus untreated pregnancies.Methods And FindingsThis observational register-based study used data from the Finnish Medical Birth Register (FMBR) covering all births in Finland (January 1, 2006-December 31, 2010). We used unadjusted and adjusted regression analyses as well as propensity score matching (PSM) to analyze whether birth size differed by ACT exposure. PSM provides a stringent comparison, as subsamples were created matched on baseline and medical characteristics between treated and untreated women. All analyses were stratified by timing of birth. The primary study outcome was birth size: birth weight (BWT), birth length (BL), ponderal index (PI), and head circumference (HC) measured immediately after birth and recorded in the FMBR. Additional analyses explored indicators of neonatal health in relation to ACT exposure and birth size. A total of 278,508 live-born singleton births with ≥24 gestational completed weeks were registered in the FMBR during the 5-year study period. Over 4% of infants were born preterm, and 4,887 women were treated with ACT (1.75%). More than 44% of the exposed infants (n = 2,173) were born at term. First, results of unadjusted regression analyses using the entire sample showed the greatest reductions in BWT as compared to the other analytic methods: very preterm -61.26 g (±SE 24.12, P < 0.01), preterm -232.90 g (±SE 17.24, P < .001), near term -171.50 g (±SE 17.52, P < .001), and at term -101.95 g (±SE 10.89, P < .001). Second, using the entire sample, regression analyses adjusted for baseline and medical conditions showed significant differences in BWT between exposed and unexposed infants: very preterm -61.54 g (±SE 28.62, P < .03), preterm -222.78 g (±SE 19.64, P < .001), near term -159.25 g (±SE 19.14, P < .001), and at term -91.62 g (±SE 11.86, P < .03). Third, using the stringent PSM analyses based on matched subsamples, infants exposed to ACT weighed less at birth: -220.18 g (±SE 21.43, P < .001), -140.68 g (±SE 23.09, P < .001), and -89.38 g (±SE 14.16, P < .001), born preterm, near term, and at term, respectively. Similarly, significant reductions in BL and HC were also observed using the three analytic methods. There were no differences among postterm infants regardless of analytic method. Likewise, we observed no differences with respect to PI. Additional analyses showed that exposed and unexposed infants had generally similar Apgar scores at birth, yet the ACT-treated infants received greater medical care during the first 7 days of life and beyond. Our study is mainly limited by lack of data in FMBR specifying the interval between treatment and birth as well as other potential confounders that could not be tested.ConclusionsIn this study, ACT was consistently associated with reduction in birth size for infants born preterm, near term, or at term. Further investigation is warranted alongside reevaluation of guidelines. Efforts need to be made to correctly identify and target patients who will deliver preterm. Reduced growth should be considered when deliberating early care decisions.

23 keywords...

hide…