Background Retinopathy of prematurity (ROP) is one of the leading causes of childhood blindness. Use of antenatal steroid can reduce neonatal morbidity and mortality in preterm births, but its effect on ROP remained controversial. We aim to determine the association between antenatal steroid and risk of ROP by a systematic review and meta-analysis.

Methods Reported studies on the association between antenatal steroid and risk of ROP or severe ROP were identified from MEDLINE and Embase databases from their inception to November 2016. Outcome measures were ORs with 95% CIs. Extracted data were pooled using a random-effect model or fixed-effect model where appropriate. Heterogeneity was assessed, and sensitivity analysis was performed.

Results A total of 434 relevant studies were identified, and 28 studies were eligible for the meta-analysis, involving 20 731 neonates with 4202 cases of ROP. Among the 28 studies included, 13 studies provided data evaluating the association between antenatal steroid use and severe ROP, involving 4999 neonates with 792 cases of severe ROP. Antenatal steroid administration was associated with a reduced risk of ROP development (OR_unadjusted=0.82, 95% CI 0.68 to 0.98; OR_adjusted=0.67, 95% CI 0.47 to 0.94) and progression to severe ROP (OR_unadjusted=0.58, 95% CI 0.40 to 0.86).

Conclusion Antenatal steroid administration is associated with a reduced risk of ROP development and progression to severe ROP. Our results strengthened the indications of antenatal steroid therapy to high-risk mothers giving preterm births, especially in low-income and middle-income countries where antenatal steroid are not yet widely used.