The role of E in affecting the course of severe ROP has remained controversial. Most E trials were done prior to the International Classification of ROP (ICROP), and in previous meta-analyses the E effect on stage 3+ ROP (as per ICROP) was not routinely assessed. Since stage 3+ ROP predicts progression to visual loss, and since the cohort at risk for ROP has increased, we performed an analysis to ask: By combining previous randomized clinical trial (RCT) data, can we demonstrate a reduction in risk for stage 3+ ROP in <1500 gm infants with E? Results: Of the 62 papers on E(1961-1994), six RCTs met our criteria for analysis. Authors of 5/6 re-confirmed the accuracy of their data being used by us. 714 VLBW infants were randomized into the control and 704 into the E groups. Outcomes of 331 were not reported, mostly due to death prior to eye examination. Pooled“exact odds ratios” (ORexact) were computed for estimating the relative risk for any ROP and stage 3+ ROP for all infants enrolled (“intention-to-treat” analysis). The pooled incidence rates for any stage of ROP were similar-43.5% in control vs 39.8% in E group- and the ORexact was non-significant. The pooled incidence rates for stage 3+ ROP was 5.2% in the control and 2.43% in the E group. ORexact was 0.44 (95% CI, 0.21, 0.87).Conclusions: A 56% reduction in the risk for 3+ ROP was seen with E; a reciprocal of this rate reduction showed that 35 infants (95% CI, 20, 180) would be given E prophylaxis to prevent one case of 3+ ROP. The cost, risks, and outcomes of such E therapy should be assessed against the backdrop of the costs, risks, and outcomes of cryo/laser surgery for stage 3+ ROP. The role of E in reducing stage 3+ ROP, especially in < 1.0 Kg birth weight infants, needs to be assessed using larger RCTs.