This is a well-conducted and well-written study aimed at identifying subgroups of OCD patients based on their comorbidity profile, using latent class analysis. The authors evaluated 419 adult patients with OCD and 15 comorbid disorders, according to the SCID-I and three proxy diagnoses. Two major groups were identified, 311 patients with low amount of comorbidity (147 ‘simplex’ and 186 with major depression) and 108 patients with high amount of comorbidity (49 anxiety-related, 27 autism/social phobia-related and 10 psychosis/bipolar-related subgroups). Some clinical correlates were also explored and, compared to the former group, the latter (high-comorbid) presented higher childhood trauma and aggression/checking dimension scores, higher illness severity, and lower scores on some personality characteristics. Chronic course was also more frequent in the high-comorbid group.

Considering that OCD is a very heterogeneous disorder and comorbidity is the rule among OCD patients, using a new analytical approach (i.e. latent class analysis) for identifying and characterizing more homogeneous subgroups based on their comorbidity profile is valuable. The study focuses on OCD external boundaries, and has potential theoretical (e.g. etiological) and practical implications (e.g. impact on help-seeking and treatment response).

I just have minor suggestions or comments, listed below:

Abstract: Make clear that the participants were all adults and describe at least the main correlates of the high-comorbid group. In my view, the significant outcomes are too generic and should be more informative.
Limitations: Besides BDD and grooming disorders, hypochondriasis and some impulse-control disorders (e.g. pathological gambling, intermittent explosive disorder, compulsive buying, internet addiction) were not assessed. Some non-significant results (e.g. remission rate and course of illness after two years) can be due to type II error, considering the small number of patients in some subgroups.

Results: In the first paragraph, please describe the percentages of the most prevalent comorbidities (> 20%): MDD, tic disorders, social phobia, panic disorder and/or agoraphobia, and ADHD.
Table 2: I suggest describing the significant results in bold type and changing the abbreviation of social phobia to SP. Panic disorder/agoraphobia is also included in the schizophrenia/psychoses subgroup, right? It is important to describe this, otherwise it may be considered part of the anxiety-related subgroup.
Discussion: The possible overlap between autism and social phobia clinical characteristics should be acknowledged and/or briefly discussed since all subjects in class 4 of 5 (autism) also presented social phobia. Likewise, the occurrence of panic disorder/agoraphobia in class 5 of 5 (psychosis/bipolar) should be discussed. For example, in a recent study by Domingues-Castro et al. (Journal of Affective Disorder 256: 324-30, 2019), panic disorder with agoraphobia was an independent predictor of bipolar disorder comorbidity in a large sample of adult patients with OCD.