Reviewer comment DOM 22-0597-OP.

This paper describes a systematic review and meta-analysis of RCTs of at least six months duration of very low carbohydrate (<50g/day) ketogenic diets to reduce HbA1c in adults with impaired glucose metabolism (“pre-diabetes”) or type 2 diabetes.

The paper is very well written, well presented, methodologically rigorous and takes a balanced, nuanced approach to the topic.

It is interesting in that it highlights the very substantial paucity of evidence in this area, with a tiny number of trials and participants. I would have the following minor comments.

I’m afraid to say that I cannot get my head around how the authors have had to treat HbA1c outcomes differently from different studies. Surely they all had baseline and follow-up measures? The description of this in the abstract is a bit convoluted but it doesn’t get any easier to assimilate in the main text and I think this is an issue that needs to be “ironed out” better by the authors, especially as there are so few trials available and HbA1c is the primary outcome.

Also the reduction in A1c of “0.65%” is somewhat confusing – is that equivalent to 6.5 mmol/mol, or are the authors describing a tiny relative change in A1c? It’s very hard to tell.

I could not find the study on PROSPERO.

It’s been 15 months since the deadline for inclusion of potentially eligible trials – it would be worth ensuring there are no new ones.

I was concerned reading the paper that the authors could have included studies with less stringent carbohydrate restriction or a duration of follow-up less than six months, but in fairness the authors have comprehensively addressed that limitation in the discussion, which is excellent.

Thanks for the clarifications and amendments