Comments on abstract, title, references

The abstract is well written.The aim to determine incidence, onset and early course of ROP in a large cohort of prematurely born babies is clear. The study is a multi central retrospective data review, findings are well presented and methodology is clear. The title is informative and relevant. Only 15 references are cited. They are however relevant, but not the most recent ones, cited correctly. The appropriate basic key studies are included in the reference list, however not the most recent studies are cited.

Comments on introduction/background

The context for the present study is well illustrated. Adequate and appropriate literature review is given to get an idea of what is already known about the topic. The research question is outlined and is based on the G-ROP group previous research. The potential importance of the present research is also well documented. The research goal of using previously collected G-ROP data in order to demonstrate incidence, timing of onset and early course of ROP is clearly outlined.

This is a secondary analysis of a previous G-ROP study data. G-ROP study was conducted in 29 US and Canada hospitals. Prematurely born children from January 2006 -end of December 2011 were included. Inclusion criteria were clearly stated. The variables were clearly defined and measured appropriately. Validity and reliability of the study are high. The study can be easily replicated.

Comments on data and results

Data are presented with 3 tables and with 4 figures. The first table presents demographic data of almost 7500 infants included in the study, with absolute numbers and percentage. The table is relevant and clearly presented. The next 2 tables are about ROP rates at different gestational ages and birth weights. The columns are divided into smaller fragments with a total number and percentage as the last column. No real statistics is utilized here, data are clearly presented. The 4 figures used in this paper are spot diagrams with percentage of all eyes with certain type of ROP at certain gestational age. No real statistics is used, but data are the results are well presented and the peak incidence of ROP can be seen immediately. The text in the results gives additional information to the figures and is not repetitive. Practically meaningful results are clearly highlighted.

Comments on discussion and conclusions

The first paragraph of the discussion states findings of the paper relatively extensively. The results address the aim. The results of this study have filled the gap of the knowledge. Previous large studies from the field are cited and results are compared. There is a separate paragraph on strengths and limitations. The major limitation was to use retrospective data, but the authors provide information why this should not significantly alter the results. The take-home massage is clearly stated in a separate paragraph of conclusions and is the need to reevaluation the ROP screening criteria in North America.

Overall the article brings important information about ROP incidence among prematurely born children of different GA and BW in high income countries. At the end it raises the need for perhaps changing the screening criteria in order to minimize the burden of excessive screening on preterm babies and in the same time not to miss out any serious ROP potentially needing treatment. The impact of this paper is very high, especially because a large cohort of preterm babies have been included in the study. There are however some weaknesses such as retrospective data review. A prospective, multicentric study, perhaps including other hospitals (in Europe and Asia) would give us more accurate insight into this topic.