To the best of our knowledge, the current study examined the largest sample of patients with cannabis‐related psychiatric disorders in Japan. Moreover, to date, this is the only study in Japan to examine risk factors for the onset of dependence and chronic psychosis due to cannabis use by considering of the impact of clinical genetic factors, preceding psychiatric disorders, and the mode of cannabis use.

As factors associated with the onset of dependence due to cannabis use, the univariate analysis identified the number of years of cannabis use, the use of cannabis other than dried cannabis, and usage on ≥4 days per week, while the multivariate analysis identified the number of years of cannabis use and the use of cannabis products other than dried cannabis. These findings are consistent with those of previous studies by Coffey et al, ⁶ Grant and Pickering, ⁷ and Noack et al ⁸ Moreover, the current results suggest that cannabis use for a longer period and at a higher frequency, as well as intake of products with a high THC content, may have a larger impact on the onset of dependence than clinical genetic factors, the combined use of other psychoactive substances, or the presence of psychological distress before starting to use cannabis. In other words, as with alcohol, ⁴² a higher frequency and longer period of THC intake, as well as a higher concentration of THC intake (ie, the total intake of THC), are associated with a higher risk of developing dependence.

We obtained unexpected results regarding the onset of residual and late‐onset psychotic disorders. In this study, as factors associated with the onset of these disorders, the univariate analysis identified the use of cannabis other than dried cannabis, while the multivariate analysis identified the presence of a comorbid psychiatric disorder before starting to use cannabis and the use of cannabis other than dried cannabis, with ORs below the decimal point (=negative association). This finding suggests the possibility that the duration and frequency of cannabis use, early start of cannabis use, and clinical genetic factors may not be associated with the onset of chronic psychosis due to cannabis use. Rather, the associated factors were (a) no prior history of use of cannabis products with high THC levels and (b) no history of psychiatric disorders prior to cannabis use.

This result is not consistent with previous studies that identified risk factors for the onset of chronic psychosis, including genetic factors, ²⁸ , ²⁹ cannabis use early in life, ¹³ the use of cannabis products with a high THC content, ³⁰ and cannabis use to alleviate symptoms caused by psychiatric disorders. ³¹ Further, the current results contradict the findings of several previous studies. Particularly, our finding of a negative association between chronic psychosis due to cannabis and the presence of a comorbid psychiatric disorder prior to cannabis use is contrary to the reported association between chronic psychosis due to cannabis and the use of cannabis for alleviating pre‐existing psychiatric disorders. ³¹ Further, our finding of a negative association between chronic psychosis due to cannabis and the use of cannabis products other than dried cannabis (resins and liquids with a high THC content) is contrary to the reported association between chronic psychosis due to cannabis and the use of cannabis products with a high THC content. ³⁰

The present data reveal an association between the absence of comorbid psychiatric disorders prior to cannabis use and chronic psychosis due to cannabis use. This finding could have resulted from several factors. In the case of patients who had some type of psychiatric disorder before starting to use cannabis, chronic psychosis was not considered to be after effect of cannabis use (residual and late‐onset psychotic disorders), but instead, as a disease state in which latent schizophrenia was induced by cannabis use, and thus was counted as a comorbid psychiatric disorder. In that sense, the finding that only a small number of patients who were diagnosed with residual and late‐onset psychotic disorders had comorbid psychiatric disorders before starting to use cannabis was an expected result given the definition of the diagnostic criteria and disease state.

In contrast, it is difficult to explain the observed association between chronic psychosis due to cannabis use and no history of using cannabis products with a high THC content. A mentioned above, the current findings indicate that the frequency and duration of exposure to THC—the constituent responsible for the psychotropic activity of cannabis—were not associated with the onset of chronic psychosis. In that context, it is unsurprising that the use of cannabis products with a high THC content was not related to the onset of chronic psychosis. These suggest that the onset of cannabis‐related chronic psychosis may be more influenced by individual/constitutional factors, including clinical genetic factors and those that precede the onset of comorbid psychiatric disorders, than the total amount of THC exposure, although we found no associations between individual/constitutional factors and chronic psychosis in this study.

The reason why a significant negative association was observed between the use of high THC‐containing products and chronic psychosis is not clear. Two possibilities may be involved in the above result. One possibility is the heterogeneity of cannabis‐related chronic psychosis in which not only pure psychosis but also other disorders such as organic personality disorder, organic mood disorder, and dementia are included. The other possibility is the small sample size of subjects which may have influenced the statistical output. Therefore, further investigation should be required in this respect.

The current study has several limitations that should be considered. First, the “number of years of cannabis use” was automatically calculated from the difference between the “age at the time of final cannabis use” and the “age at the time of first cannabis use.” Periods of voluntary drug abstinence and periods of forced drug abstinence while serving a prison term were not considered. Therefore, the number of years of cannabis use and the total amount of cannabis used may not be positively correlated. Second, the ICD‐10 “Residual disorders and late‐onset psychotic disorder” is a heterogenic group including rare psychiatric cannabis‐related conditions in clinical practice, such as flashbacks, organic personality disorder, organic mood disorder, and dementia, in addition to chronic psychosis. Therefore, the risk factors identified in this study may be inconsistent with those of purely chronic psychosis. Third, some cases of chronic psychosis due to cannabis use may not have been listed as “residual and chronic late‐onset psychotic disorder,” but could have been listed as a comorbid psychiatric disorder, namely, cannabis‐induced schizophrenia. Forth, because this study was conducted in patients who were undergoing treatment at a psychiatric hospital, the sample may not have been representative of cannabis users in the general population. In our sample, 43.7% of all subjects had some sort of comorbid psychiatric disorder at the time when the survey was conducted. Of this subset, approximately half (22.5% of all subjects [16 of 71 subjects]) had been suffering from some type of psychiatric disorder before they started to use cannabis. Thus, because the subjects in the current study had various psychiatric problems before they started to use cannabis, they were likely to have different characteristics compared with cannabis users in the general population. Fifth, the sample size was too small for in‐depth statistical examination, and the possibility of type II errors cannot be excluded. Sixth, because all attending psychiatrists used an interview‐based method in which they asked patients questions about the mode of cannabis use in the past, it was difficult to avoid recall bias. Finally, in this study, we collected information from several attending psychiatrists at nine mental health hospitals, and the evaluation criteria might have varied between institutions. It should be emphasized, however, that each of these nine mental health hospitals treat many patients with drug dependence daily, and all attending psychiatrists had a high level of expertise for diagnosing and treating drug dependence.

Despite the limitations described above, the current study represents an important contribution as the only study of its kind in Japan. We examined the largest sample of patients to date with cannabis‐related psychiatric disorders in Japan to investigate the onset of dependence and chronic psychosis due to cannabis use while considering clinical genetic factors, preceding psychiatric issues, and the mode of cannabis use.

The current study examined risk factors for the onset of dependence and chronic psychosis due to cannabis use in 71 patients with cannabis‐related psychiatric disorders who underwent treatment at nine mental health hospitals in Japan. The results suggested that the onset of dependence may be related to long‐term cannabis use and the use of cannabis products with a high THC content. However, regarding the onset of chronic psychosis, no association was observed for total intake of THC (including duration and frequency of use), or for psychiatric vulnerability (including clinical genetic factors and the presence of preceding psychiatric disorders). These findings indicate that the onset of chronic psychosis is related to unknown factors.