This is a very interesting article about the relationship between smoking behavior and the impact in clinical and prognosis variables of bipolar disorder. The sample is huge and the methodology is appropriate. The results are not new but can be useful to improve the knowledge about the impact of smoking in the patients. The analysis about how the relationship between smoking and the presence of metabolic syndrome is mediated by the use of antipsychotic is new and clinically relevant.

I have some comments for the authors:
1.- The introduction section is updated and complete. Please introduce more information about how smoking is a factor of bad prognosis in BD (quality of life, relapses, severity of the symptoms…). What is principal reason because the general practitioner drifts the patient to this centers?
2.- Why are the patients recruited in the expert less severe than usual patients? Please give us more information about how this type of centers works.
3.- The methods section is very complete. Was evaluated the number of hospitalizations?
4.- It is surprising the higher proportion of patients (42%) that never smoke. As authors explain in the introduction section approximately 70% of the patients are smokers. How it can be explained?
5.- Figure 1: please include if these differences are statistically different.
6.- It would be good to include more data about the reasons to quit smoking in these population and the difficulties to obtained cessation. There are several articles in the last two years about this issue.

I have not new comments for the authors