Table 3.
Characteristics of included studies in patients/risk of psychosis comparing cannabis users (P-C) with non-users (P-NC): Neurochemical differences.
| Study | Method | Measure | Population | Cannabis use definition |
N
|
Findings | Limitations | Confounders considered: |
||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| P-C | P-NC | HC-C | HC-NC | ETOH | Other drug | AP | Tob | |||||||
| (A) Endocannabinoid system | ||||||||||||||
| Dean et al. (2001) | Post-mortem autoradiography | [3H]CP-55940 to index CB1 receptor density at dorsolateral prefrontal cortex, caudate-putamen and temporal lobe regions | Schizophrenia assessed after psychologist and psychiatrist case notes review using diagnostic instrument for brain studies (DIBS) | THC in blood at time of death | 5 | 9 | 4 | 9 | Increased DLPFC binding in patients vs. controls. Increased binding in caudate-putamen in cannabis groups independent of whether patients or controls | Small numbers. Did not adjust for key confounders (see right) | N | N | N | N |
| Zavitsanu et al. (2004) | Post-mortem autoradiography | [3H]SR141716A to index CB1 receptor density binding at anterior cingulate cortex | DSM III/DSM IV schizophrenia using case notes review using SCAN and DIBS | Lifetime ever use | 5 | 5 | 9 | No difference P-C vs. P-NC. Patients increased CB1 receptor binding compared with controls | Small group size | N | N | Y | N | |
| Monterrubio et al. (2006) | Peripheral blood (cross-sectional) | Fatty acid levels in blood | Schizophrenia treated with clozapine | Ever used. Last use ≥6 months ago | 6 | 6 | In cannabis users only: arachadonic acid correlated with total fatty acid. Linoliec acid correlated with stress | Small size; Discontinued users | N | Y | Y | N | ||
| Leweke et al. (2007) | Cerebrospinal fluid (CSF) (cross-sectional) | Anandamide levels | Schizophrenia | High frequency in cannabis group: ≥20 times per life. Low frequency in non-cannabis group: ≤5 times per life | 22 | 25 | 26 | 55 | P-NC markedly higher anandamide in CSF than P-C. No difference between HC and HC-NC | Needs replication | N | N | Y | N |
| Deng et al. (2007) | Post-mortem autoradiography | [3H]SR141716A and [3H]CP-55940 both to index CB1 receptor density at superior temporal gyrus | DSM IV schizophrenia using case notes review using SCAN and DIBS | Not clear | 4 | 4 | 8 non-psychiatric controls – unclear if any cannabis use history | P-C vs. P-NC no difference between groups. No difference between patients vs. controls | Small numbers. Insufficiently clear about cannabis use history | N | N | Y | N | |
| Eggan et al. (2008) | Post-mortem immunocytochemistry | CB1 receptor mRNA in dorsolateral prefrontal cortex (Brodmann area 9) | Schizophrenia/ schizoaffective | Not clear | 7 | 16 | 23 | No difference between P-C and P-NC. Reduction of around 10–15% in CB1 receptor transcript expression in patients vs. controls | Study not designed to determine difference of P-C vs. P-NC. Tested as a possible confounding variable | N | Y | Y | N | |
| Eggan et al. (2010) Includes 14 patients and 14 HCs from Eggan et al. (2008) | Post-mortem immunocytochemistry | CB1 receptor protein expression in dorsolateral prefrontal cortex (Brodmann area 46) | DSM IV Schizophrenia/schizoaffective using case notes review and structured interview with relative. Controls: Controls with no psychiatric history (NP) and depressed (DEP) | Lifetime history of cannabis use | 6 | 15 | 0 NP, 3 DEP | 26 NP, 7 DEP | P-C vs. P-NC no significant difference between groups. Reduction of around 19–23% in CB1 receptor density in patients with psychosis vs. controls | Study not designed to determine difference of P-C vs. P-NC. Tested as a possible confounding variable | N | Y | Y | N |
| Ho et al. (2011) | Structural MRI, cognitive assessment (WAIS subscales) by CB1 receptor genotype and cannabis interaction (cross-sectional) | White matter & WAIS subscales | Schizophrenia patients, P-C arranged by CB1 receptor genotype (12 tagged SNPs) | Cannabis abuse or dependence | 52 | 183 | Three CB1 receptor polymorphisms associated with decreased WM volume. One also associated with decreased processing speed and attention in P-C | Needs replication in larger cohort. Possible confounding from other substance use | Y | Y | Y | N | ||
| Onwuameze et al. (2013) (from same sample as Ho et al. (2011)) | Structural MRI by MAPK14 genotype and CB1 receptor and cannabis interaction (cross-sectional) | White matter | Schizophrenia patients, P-C arranged by MAPK14 Receptor genotype (nine tagged SNPs) | Cannabis abuse or dependence | 52 | 183 | MAPK14 and CB1 receptor specific alleles associated with small white matter brain volume in heavy cannabis use. Independent and additive effect | As Ho et al. (2011). Also no functional correlate of WM loss demonstrated | Y | Y | Y | N | ||
| Ceccarini et al. (2013) | PET (cross-sectional) | [18F]MK-9470 mSUV (indexes CB1 receptor) | Schizophrenia | Ever use. Last use for all participants ≥6 months ago | 35 | 32 | 12 | Patients with a history of heavy cannabis use no significant difference in binding vs. medium, low or never use | Not designed to test P-C vs P-NC | N | N | N | N | |
| Volk et al. (2014) (same participant group as Eggan et al. (2008) | Post-mortem autoradiography | [11C]OMAR binding (indexes CB1 receptor) | Schizophrenia/ schizoaffective | Not clear | 7 | 14 | 21 | [11C]OMAR binding did not differ between P-C vs. P-NC | Study not designed to determine difference of P-C vs. P-NC. Tested as a possible confounding variable | N | N | N | N | |
| Ranganathan et al. (2016) | PET (cross-sectional) | [11C]OMAR VT (indexes CB1 receptor) | Male Schizophrenia | Ever use. Lifetime cannabis use disorder excluded. 1/25 patients with recent use | 16 | 7 | Total 18 HC. Lifetime use unclear | No significant correlations between cannabis use and VT | Not designed to test P-C vs. P-NC | N | N | N | N | |
| (b) Dopamine system | ||||||||||||||
| Dean et al. (2003) | Post-mortem autoradiography | [3H]Mazidol for DAT; Tyrosine Hydroxylase | Schizophrenia | Blood test +ve | 5 | 9 | 4 | 10 | No significant difference between CBS users and non-users | Small sample. Limited cannabis information | N | N | Y | N |
| Bowers and Kantrowitz (2007) | Peripheral blood (cross-sectional) | Plasma homovanillic acid | Inpatient FEP vs. inpatient non-psychosis | Urine test +ve | 5 | 15 | 18 | P-C group elevated HVA levels vs. P-NC and others (p=0.001) | Small sample. Limited cannabis information | N | N | N | N | |
| Safont et al. (2011) | SPECT (cross-sectional) | [123I]IBZM striatal/frontal ratio to index D2/D3 receptor availability | Untreated FEP patients | Use 3 units/day for last 3 months (n=14) | 14 | 23 | 18 | No significant difference between P-C and P-NC | Used S/F ratio but frontal binding may be altered in cannabis use | N | Y | Y | N | |
| Kuepper et al. (2013) | PET: 8 mg inhaled THC vs. placebo Acute challenge on dopamine function (interventional) | [18F] Fallypride displaced | Psychosis patients, first-degree relatives of patients, healthy controls | Self-report ever use | 8 pts 7 rel’s | 9 | THC induced significant striatal displacement of fallypride in patients and relatives but not controls | No comparison between P-C and P-NC | Y | Y | Y | Y | ||
| Mizrahi et al. (2014) | PET: stress task to induce dopamine release (cross-sectional) | [11C]PHNO displaced | Clinical high risk | Use at least 3 times/week | 12 | 12 | Decreased displacement in P-C group and increased in P-NC group in striatum | Unable to determine whether blunted dopamine release is marker of cannabis use or addiction | N | Y | Y | Y | ||
| (c) Glutamate system | ||||||||||||||
| Rentzsch et al. (2011) | See Table 2 (e) (Mismatch Negativity believed to index glutamatergic function via NMDA receptor) | |||||||||||||
| Pesa et al. (2012) | See Table 2 (e) (Mismatch Negativity believed to index glutamatergic function via NMDA receptor) | |||||||||||||
| Rigucci et al. (2017) | Magnetic resonance spectroscopy (cross-sectional) | Glutamate medial prefrontal cortex & neurocognitive assessment (MATRICS battery) | Early psychosis and healthy controls | Current cannabis use | 18 | 17 | 33 | Decreased glutamate in P-C vs. P-NC. Impaired working memory P-C vs. P-NC | Neuro-cognitive impairment in P-C rather than sparing. ?atypical sample. Requires further replication | Y | N | Y | Y | |
| (d) GABAergic system (indexed through cortical inhibition) | ||||||||||||||
| Wobrock et al. (2010) | Transcranial Magnetic Stimulation (cross-sectional) | Short-interval cortical inhibition (SICI), intracortical facilitation (ICF) | First-episode schizophrenia | P-C: lifetime use of ≥20 times per lifetime; P-NC Lifetime use of ≤5 times | 12 | 17 | Reduced SICI and enhanced ICF for P-C vs. P-NC indicating GABAergic deficit and intracortical disconnectivity | SICI and ICF not direct measures of GABA-A. No comparison with HC groups | N | Y | Y | N | ||
| Goodman et al. (2017) | Transcranial magnetic stimulation (cross-sectional) | SICI/ICF | Schizophrenia/schizoaffective disorder | DSM IV cannabis dependence | 12 | 11 | 10 | 13 | Increased SICI for PC vs. P-NC indicating increased GABA-A mediated inhibition, reduced SICI HC-C vs. HC-NC. No significant difference for ICF | SICI and ICF not direct measures of GABA-A | Y | Y | Y | N |
(a) Endocannabinoid System.
(b) Dopamine System.
(c) Glutamate System.
(d) GABAergic System.
P-C: Psychosis/at-risk patients with cannabis use; P-NC: Psychosis/at-risk patients without cannabis use; HC-C: Non-psychosis controls with cannabis use; HC-NC: Non-psychosis controls without cannabis use; ETOH: Alcohol; AP: medication for psychosis; Tob: Tobacco.