Risk Factors for Severe Retinopathy of Prematurity Among Very Preterm Infants: A Unit-Based or Population-Based Approach?

To the Editor .— In the interesting article by Darlow et al,1 the authors correctly remark that, although many factors (including prenatal factors) have been suggested to have some causal association with retinopathy of prematurity (ROP), most studies have been unit based or from hospital groupings, rather than being population based, and may have introduced bias into the results. These limits in the available studies may explain the fact that the same prenatal variables are reported as associated to an increased risk of ROP by some single-center, cohort studies and as not associated or even protective in others (eg, the use of antenatal steroids [no association according to refs 2–4; an association with a protective effect based on refs 5 and 6] or the maternal history of preeclampsia [no association according to ref 7; an …