We read, with great interest, the article by Antti Mustonen et al entitled ‘Adolescent cannabis use, baseline prodromal symptoms and the risk of psychosis’, it is a tremendous study and is well conceptualised.^{Reference Mustonen, Niemelä, Nordström, Murray, Mäki and Jääskeläinen1} We would like to make certain comments.

In the Method section, under the subheading of Psychosis diagnoses, ICD diagnoses have been mentioned as (F22-F29, F302, F312…) whereas the fourth character in ICD-10 codes is always a ‘dot’.² Second, the codes F302 and F312 are also mentioned under the category of psychosis. If these codes refer to F30.2 and F31.2, respectively then these fall in the category of affective disorder without psychosis. These errors in reporting make the article quite difficult to follow.

There is also a mismatch between the figures and the text in the article. In the Results section, Figure 2, explaining the cumulative incidences of psychosis in four groups with and without cannabis use and prodromal symptoms in the Northern Finland Cohort 1986 is not self-explanatory as the figures (n = 13/134, 5/134…) has not been explained in the text and I found it difficult to understand the origin and meaning of these figures.

Similarly, while mentioning the association between adolescent cannabis use and subsequent psychosis (on page 230 in the first paragraph of the Associations between adolescent cannabis use and subsequent psychosis subsection) the authors state that 18 out of 375 (4.8%) cannabis users received a diagnosis of psychosis during the 15-year follow-up (4 narrow-defined schizophrenia, 4 schizophrenia spectrum disorder, 0 bipolar disorder with psychotic features, 7 major depression with psychotic features, 3 other psychosis). As 7 out of 18 participants had major depression with psychotic features, which is a mood disorder, inclusion of it in the criteria would lead to inaccurate results. If these 7 participants are excluded, there would be 11 participants who had psychosis. There should have been more participants for a better exploration of the hypothesis formulated for the study. It would be of great help if the authors could clarify these points. Thank you.