In this systematic review it was investigated whether cannabis use has an effect on brain morphology in psychosis patients and in subjects at clinical or genetic risk for psychosis. The present review focused on structural MRI and additionally reviewed post mortem studies examining brain structure based on cannabinoid receptor density. Our systematic search strategy and literature review uncovered consistent brain structural abnormalities in CB1 receptor-enhanced brain areas such as the cingulate, the prefrontal cortex and the cerebellum.
Few structural neuroimaging studies have investigated cannabis use in non-psychiatric populations yet and results were inconsistent. Limited evidence of major effects of cannabis on brain structure has so far been reported [8
] with the strongest effects being found in medial temporal regions [10
]. We focused here on psychiatric patients affected with early or chronic phases of psychosis. Some methodological limitations must soon be acknowledged. For example, comparing results between studies presented in this systematic review is hindered by differences in subject selection and design of the studies. Not all studies used DSM-IV criteria for cannabis dependence or abuse and studies varied in how they set criteria to define their cannabis using (C+) and non-using (C-) group. A solution to overcome could have been to include cannabis as a continuous variable rather than creating dichotomous groups as it was done by a few studies in this review [40
]. However, the problem with most studies is also that cannabis intake is poorly measured across studies yielding to high between-samples heterogeneity. This point is very important because smoking styles and quantities vary largely and it has also been reported, that the THC content of smoked cannabis has increased over the past 20 years [61
]. Another problem for quantitatively comparing the different included studies is the variety regarding the inclusion of a control group: some studies included a non-psychiatric sample and others did not. Within those studies that included a non-psychiatric sample, only few also analysed cannabis using healthy controls.
Despite these factors that make it difficult to compare the included studies at meta-analytical level, a few conclusions can be drawn. Within the 15 in vivo
structural MRI studies included in this review, 11 found that cannabis use (as individually defined in each study) was associated with a decrease in global or specific brain structures in psychosis patients [40
] or subjects at clinical/genetic risk for psychosis [39
]. These effects seemed to be particularly strong in brain regions rich on CB1 receptors, such as the cingulum [54
], the dorsolateral prefrontal cortex [44
] and the cerebellum [52
]. Two studies reported no difference between users and non users [45
] in psychosis and two studies reported that early onset cannabis users showed more white matter than cannabis naïve patients and controls [46
]. Conversely, none of the post mortem studies found an effect of cannabis use on cannabinoid receptor density except for Dean et al.
] who reported an increase of CB1 receptors in the caudate-putamen from subjects who had recently ingested cannabis.
The effects of cannabis on the patient and control groups were differentially reported: while one study found that the effects of cannabis on brain structure are equally both in ARMS and controls [39
], other studies showed that the negative effects of cannabis use on brain structure were significantly more pronounced in psychosis patients and genetic high risk subjects than in healthy controls [40
]. In general, this review shows that effects of cannabis on brain in psychosis subjects seem to be more distinct than in studies investigating non-psychiatric samples [8
]. This could be an indication that people with schizophrenia or at genetic high risk for this disease may have a particular sensitivity to brain tissue loss on exposure to cannabis, which is also in line with previous research reporting that the use of alcohol and amphetamines is associated with greater structural brain abnormalities than this would be expected in healthy individuals with a comparable level of exposure [64
]. However, the question remains whether brain abnormalities are only an expected consequence of substance misuse or whether they also predispose it: cortical and hippocampal dysfunctions in schizophrenia could also be responsible for the greater reinforcement of drugs leading to more drug problems, which underlines the latter assumption [66
4.1. Potential Mechanisms Underlying Cannabis Action on Brain Structure in Psychosis
There could be two ways in which cannabis affects brain structure in psychosis: cannabis could either directly affect brain morphology or the volumetric changes might be indirectly caused through the psychotic symptoms which are associated with cannabis use [67
]. The direct mechanism could be explained as follows: two post mortem
] showed significant differences in CB1 receptor binding between schizophrenia patients and healthy controls. This suggests that changes in the endogenous cannabinoid system may be involved in the pathophysiology of schizophrenia. The endogenous cannabinoid system is fully reviewed in a separate paper published in the present issue. Dean et al
] additionally showed that acute cannabis use was associated with change in density of CB1 in tissue. Therefore, a plausible consequence of chronic cannabis use would be that these structures change in volume, which may also happen in other regions rich on CB1 receptors. However, change in receptor density due to cannabis was only shown in one post mortem
study out of four [59
It was postulated that the interaction of endogenous cannabinoids with CB1 receptors is critically involved in brain development during the sensitive period of adolescence through its regulating role in the release of glutamate. THC disturbs this normal physiological process through its action on CB1 receptors. Consequently, glutamate release is hindered which leads to neurotoxic effects and cortical structural abnormalities [13
]. In contrast, two studies showed that cannabis use during adolescence was associated with more white matter compared to cannabis-naïve patients [46
] suggesting that cannabis-naïve individuals who develop schizophrenia might have a more vulnerable brain structure compared with that of cannabis users who develop the disease. A direct neurotoxic effect of cannabis on brain was shown by Jockers-Scherubl et al
] who reported that schizophrenia patients with regular cannabis use had significantly raised nerve growth factor serum levels compared to controls and schizophrenic patients not consuming cannabis. Additional evidence directly implicating abnormal glutamate levels in the early phases of psychosis is available from recent molecular imaging studies [69
Next to direct mechanisms, there might also be even more complex interactions between cannabis use and genetic factors that lead to brain morphologic changes, known to be involved in schizophrenia. Ho et al
. (2011) found evidence for gene environment interactions, showing that rs12720071 genotype significantly interacts with marijuana use on white matter in schizophrenic patients. More indirectly, cannabis use was shown to be associated with poorer clinical outcome [67
], which in turn could also lead to a higher “toxic” effect of the psychotic state on the brain [72
Most studies included in this review had small sample sizes, leading to limited statistical power. Also, large differences in secondary variables across studies (i.e. gender, medication status, etc.) and the high overlap between cannabis and other illicit drug use may have played a confounding role. In many studies, it could therefore not explicitly be excluded that the observed effects were secondary to cannabis in contrast to other drug use. A further caveat is that there may be differences between oral ingestion and smoking cannabis; however studies did not explicitly present the methods of cannabis intake. Similar, not all studies presented mean dosage of cannabis intake prevailing any analyses on dose-response effects. For future studies we suggest including control group of cannabis-using subjects. Most studies so far have only compared cannabis using and non-using patients with cannabis naïve controls. However, this approach does not enable conclusions regarding to whether brain structural differences are caused through cannabis use or the disease of schizophrenia.
This review suggests that cannabis use in psychosis is associated with volume loss of global and specific brain structures, whereby the effects seem to be particularly strong in CB1 rich brain regions such as the cingulum, the dorsolateral prefrontal cortex and the cerebellum. As the current literature did not uncover strong similar effects in healthy samples yet, psychosis patients and subjects at risk for psychosis might be particularly vulnerable to brain volume loss due to cannabis exposure.