We thank Costa et al. 1 for their comments. They question how we explain that lutein supplementation affects retinal development and the severity of retinopathy of prematurity (ROP). The first issue raised is whether the lutein supplementation was sufficient to produce effects. Certainly, an 'optimal' amount of dietary lutein for the preterm infant remains unknown. Nevertheless, lutein supplementation in our clinical trial 2 establishes three things.First, our study shows supplementation raises plasma lutein to levels approximating those observed in healthy human milk-fed term infants, that is, dietary lutein was absorbed and made available to tissues. Of the four randomized clinical trials in which lutein and zeaxanthin were given to preterm infants, 2-4 we provided carotenoids as a part of the ingredient matrix of infant formula. The daily dietary lutein was derived from the lutein levels found in human milk and colostrum. The three Italian trials 3-5 provided a single daily dose of 0.5 ml of lutein drops (containing 0.14 mg of lutein and 0.006 mg of zeaxanthin). Nutritional studies indicate that lutein, like other carotenoids, is better absorbed from a lipid-containing diet, for example, milk; delivery can be less certain when oral or orogastric tube drops are provided. In any case, our results show earlier and higher levels than were obtained in Romagnoli et al.'s 3 trial, the other study where blood levels were measured. As expected, that trial found no differences between groups on plasma cholesterol or triglyceride levels. 3 It is highly unlikely that there would be an unexplained systematic difference in plasma lipids in our larger study populations.Second, Romagnoli et al. 3 previously proposed further investigations would be needed to evaluate whether lutein supplementation can influence functional rather than structural vision outcomes in preterm infants. 3 Our trial demonstrated that dietary lutein improved rod photoreceptor function in these infants.Finally, no trial to date has been adequately designed specifically to test the hypothesis that lutein affects ROP outcomes. Among the four trials, only Dani et al. 4 performed a power analysis to test a lutein effect on ROP, but the study ended early; it also assumed an effect might be maximal with less severe ROP. But it is intriguing that Rubin et al. 2 (seven sites), Dani et al. 4 (five sites) and Manzoni et al. 5 (three sites) each showed a trend of decreased ROP severity (stages three to five or Plus disease) among the luteinsupplemented groups. Romagnoli et al. 3 (one site) had few subjects. Consequently, it remains undetermined whether lutein supplementation can decrease the incidence or severity of ROP in at-risk preterm infants.We hypothesize that lutein is more likely to have an effect on severe retinopathy. The pathogenesis of ROP is hypoxic ischemic injury to the immature retina leading to a proliferative vascular response. Secondary (postnatal) preventions for ROP aim to suppress the retinal oxidant stress and inflammatory responses. The f...