**********
N.B. These are the comments made by the Academic Editor when reviewing an earlier version of this paper. Prior to publication, the manuscript has been revised in light of these comments and to address other editorial requirements.
*********

Review of the original submission:
Summary

Harms and colleagues tackle in this important study on a relevant question; should patients with acute ischemic stroke receive prophylactic antibiotics to reduce the short term incidence of infections?. The investigators opt to use the stronger method available, a randomized controlled trial (RCT).

On intention to treat analysis (ITT) the incidence of infection was not significantly different between the groups but a per protocol (PP) efficacy analysis showed a significantly lower incidence of infection in patients allocated the antibiotic moxifloxacin 400 mgs daily or placebo for 5 days (41.9% vs17.1%, p=0.03). Based on these results the Authors interpret that preventive administration of moxifloxacin is superior to placebo in preventing infections after stroke.

The authors also address some secondary (but important) endpoints that allowed them to conclude that the neurological outcome, survival rate, and antibiotic resistance were not altered in patients receiving prophylactic moxifloxacin.

At last, the authors identify that the expression of MHC class II was decreased in monocytes of patients who developed infections.

General Comment

This is an important and timely trial with relevant information. The authors have contributed in the past to major steps in the understanding of the complex interaction between the CNS and the immune system. This study follows these well paved tracks but deserves a number of clarifications before this reviewer totally agree with their conclusions.

Specific questions

It is a little worrisome the number of patients excluded for the ITT analysis (17.5% of patients) for varied reasons. Reassuringly, patients were declared "invalid" in a blinded way but some criteria are arguable. Thus, deaths were excluded in the ITT analysis but Authors must acknowledge the theoretical plausibility that they could be due to the investigational treatment (although were more frequent in the placebo group). The authors should recognize these limitations and accord their main conclusion (in the text and abstract) to the ITT analysis.

The Authors should better emphasize that these results only apply to patients with moderate to severe ischemic stroke and cortical involvement.

The study criteria of 9 to 36 hours for stroke symptoms onset and the election of 11 days for the cut off of infection should be explained, as otherwise seem arbitrary.

When referring to the day that infections occurred it is not clear whether the clock of the study was initiated at clinical stroke onset or at treatment onset, which could occur 36 hours later.

Presentation in table form of the infections, both in ITT and PP analyses, is recommended.

Analysis of antibiotic resistance is very important although less obvious whether this information should accompany the primary results of this trial. I recommend deletion of this part which could constitute an independent secondary paper (or maybe a complementary file). It would also lighten a somehow lengthy article.

The Barthel Index is an accepted measure of functional activities but I recommend adding data using the modified Rankin scale. This would facilitate a more homogeneous comparison with other studies and the possibility of meta-analysis.

Review of the first revised manuscript:
The authors have appropriately addressed my main points.