1. Use of date-based method of randomization is not a strict randomization procedure. This issue should not be brushed aside by authors as ‘lacunae for purist’. If a proper randomized control trial would have been done, the results could have been different. Similarity of results of this study to previous studies is not the test of robustness of this study.

2. Use of flow rate of 4 L/min was not justified, when standard practice guidelines for neonatal resuscitation recommends at least 5 L/min of flow rate(2).

3. Use of Apgar score at 5 min is probably not be the best primary outcome variable. Grimace and tone are normally not tested at 1 min and 5 min if baby requires active resuscitation at these time points. It is mentioned, of the two persons present at each delivery, one person only monitored the time and outcome measured. However, if both the residents get involved in resuscitation in a difficult case, or if there is only one resident at time of deliver (not uncommon), the apgar recording will be on subsequent recall. Then one is not sure of the time as well as the actual Apgar score. Also, the 5 minute Apgar could be affected by 100% oxygen, which was delivered to all the babies who ‘failed treatment’ in room air resuscitation (RAR) group. This cannot be accounted for by intention to treat analysis. So, these babies should have been excluded at the time of analysis. Moreover, there can be significant inter rater variability among different residents, in recording Apgar scores within and between different centers.

4. Sample size calculation: This seems to be done as for superiority trial. However, the hypothesis and conclusion in abstract stated is that of equivalence trial, which would have required much larger sample size.

5. Statistical analysis: Cluster adjustment for each center should have been done, as the data is not one random sample, but collated data from four centers.

1. If 1 min Apgar score in RAR group was significantly higher, it could mean this group had babies with less severe asphyxia. In absence of data on cord pH, it is difficult to say that babies in 2 groups had suffered comparable asphyxia to begin with. Comparable Apgar score at 5 min could mean that use of RAR caused some harm to the babies and their Apgar score became comparable at 5 min. It is possible that the residents were so unconvinced of RAR that they used 100% oxygen when baby was severely asphyxiated or was born to some parents known to the residents/doctors in that hospital. This could account for lower Apgar score in 100% oxygen group.

2. Results presented in Table 2 are not clear. Example, though authors mention there is significant difference in 1 minute Apgar Scores in the two groups, the median and range mentioned in the table are identical. Similarly, though the median and 5-95 centile values of ‘time to first cry’ and’ duration of resuscitation’ are almost similar, the p value is highly significant.

3. Total number of live births and number of babies who required intubation and chest compressions in the two groups was not mentioned. Since the study was completed six years ago, data on follow up of the babies in the two groups should also have been mentioned.