This French study from the FACE-BD network investigated associations of smoking status with physical and mental health status of patients with bipolar disorders (BD). The study has several merits, including a large national multicentre sample (N=3860) and systematic structured evaluation of patients with appropriate diagnostic and other measures. The authors found current and former smoking to associate with multiple indicators of high morbidity. The topic of smoking is obviously important for health outcomes and life expectancy of BD patients.

While many aspects of the study are excellent, there are also several limitations and problems with the study and the manuscript:

• The FACE-BD network sample is impressive. However, it would be important to clarify, what is known about epidemiological representativeness of the cohort.
• Most importantly, the study design is cross-sectional. The authors do correctly note that this precludes causal inferences, but at some points this seems to have been forgotten. Overall, although some of the differences between subgroups are impressive, direction of causality remains unclear, and differentiating causes from consequences is not possible. It is important not to make causal claims based on such data. This also limits the range of credible implications.
• In my view, the most clinically important findings involve strong clustering of all kinds of substance use disorders into the current smoking group. As these individuals not only smoke but also are more prone to alcohol and cannabis use disorders as well as misuse of other substances, the specific role of smoking as a part of this clustering is difficult if not impossible to ascertain. It also means difficulty of collecting reliable information on all relevant exposures, given that use of illicit drugs is most likely also more prevalent within this group. On the other hand, I think the authors should highlight concurrent use of multiple substances within the smoker group more – they are not just smokers, and use of multiple substances influences clinical care.
• I would hesitate to make any strong claims based on the subgroup of patients who are former smokers in a cross-sectional study, and do not find such valid. Smoking and quitting can be intermittent and there can be multiple possible pathways as well as confounders. I hope that quitting smoking may mean return to never-smokers levels, but this really needs to be investigated in prospective studies.
• Overall, I can’t help finding the manuscript still somewhat sketchy. Writing could be improved. For example, the methods section could be better organized by using subtitles, and adding missing details. For example, please clarify whether psychiatric comorbidity refers to current comorbidity vs. lifetime comorbidity, and how were individuals smoking 1-4 cigarettes, using pipe or cigars or vaping classified, etc.
• Finally, given that there are several other publications about use of substances from the FACE-BD network, it is important to be clear about precisely what is the novel content and what motivated the current work.

I think the authors have addressed my concerns in an acceptable way.