Abstract
This systematic review explores the role of the endocannabinoid system (ECS) in prodromal psychosis and its potential as a therapeutic target. Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, 22 studies published between 2000 and 2025 were analyzed, comprising preclinical research, genetic studies, neuroimaging investigations, and clinical trials. Converging evidence suggests that ECS alterations precede and potentially contribute to the development of psychotic symptoms, with CB1 receptor modifications and endocannabinoid levels correlating with symptom severity and transition risk to full-blown psychosis. Neuroimaging studies revealed reduced CB1 receptor availability in key brain regions in high-risk subjects, and intervention studies, particularly with cannabidiol—though its therapeutic mechanisms likely extend beyond ECS modulation to include dopaminergic and other neurotransmitter pathways—have shown promising results. Proposed mechanisms of action include stress response attenuation, neuroinflammatory modulation, neurodevelopmental stabilization, and normalization of the dopamine-glutamate interface. Despite limitations of existing studies, primarily small size and short duration, this review provides a solid foundation for developing ECS-targeted interventions as a promising approach to modify disease trajectory during the prodromal phase, potentially offering safer and more effective therapeutic options for individuals at clinical high risk for psychosis.
Keywords
Introduction
Psychotic disorders affect approximately 3% of the general population and contribute substantially to global disease burden (Perälä et al., 2007). The critical importance of early intervention has generated intense interest in identifying and treating individuals during the prodromal period—a stage characterized by attenuated psychotic symptoms (APS), cognitive alterations, and functional decline that precedes full psychosis onset (Fusar-Poli et al., 2013). While early intervention has demonstrated promising results, current pharmacological approaches primarily rely on antipsychotics, which carry significant side effect burdens and limited evidence of disease-modifying effects (Davies et al., 2018).
The endocannabinoid system (ECS) has emerged as a particularly promising therapeutic target due to its critical role in neurodevelopment, synaptic plasticity, and neuromodulatory signaling (Lu and Mackie, 2016). This complex lipid signaling system consists of cannabinoid receptors (primarily CB1 and CB2), endogenous ligands (endocannabinoids) such as anandamide (AEA) and 2-arachidonoylglycerol (2-AG), and enzymes responsible for their synthesis and degradation (Di Marzo et al., 2004). The ECS regulates neurotransmitter release across multiple systems implicated in psychosis, including glutamate, dopamine, and gamma-aminobutyric acid (GABA) circuits, positioning it as a potential “upstream” intervention target (Ferretjans et al., 2012).
Emerging evidence supports ECS dysregulation in psychosis pathophysiology from several lines of research: (1) epidemiological studies consistently link cannabis use, particularly high- Δ-9-tetrahydrocannabinol (THC) varieties, to increased psychosis risk (Di Forti et al., 2019); (2) genetic variations in cannabinoid receptor and enzyme genes modify psychosis risk (Colizzi and Bhattacharyya, 2018); (3) postmortem and neuroimaging studies demonstrate altered CB1 receptor expression in schizophrenia (Ranganathan et al., 2016); and (4) preliminary clinical trials suggest cannabidiol (CBD) may have antipsychotic properties with favorable safety profiles (McGuire et al., 2018).
However, the specific role of the ECS in prodromal psychosis stages and its potential as an intervention target during this critical window remains incompletely synthesized. This systematic review aims to address this gap by: (1) examining evidence for ECS alterations in clinical high-risk (CHR) and prodromal psychosis states; (2) evaluating preclinical and clinical evidence for cannabinoid-based interventions in early psychosis; (3) exploring molecular mechanisms that might inform novel therapeutic approaches targeting the ECS; and (4) identifying key research gaps and future directions for translating this knowledge into clinical applications.
It is important to acknowledge that while CBD has been extensively studied in psychosis research and is often discussed within ECS frameworks, its therapeutic mechanisms likely extend well beyond ECS modulation. Recent evidence suggests that CBD’s antipsychotic effects may be primarily mediated through dopamine D2 receptor interactions rather than CB1/CB2 pathways (Seeman, 2016), along with effects on serotonin 5-HT1A receptors, transient receptor potential cation channel subfamily V member 1 (TRPV1) channels, and additional molecular targets. This mechanistic complexity necessitates careful interpretation when evaluating CBD studies within purely ECS-focused frameworks, while still recognizing the compound’s relevance to understanding broader cannabinoid-based therapeutic approaches in prodromal psychosis.
Methods
Search strategy
This systematic review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Figure 1) with complete adherence documented in Appendix 1. A comprehensive literature search was performed across five electronic databases: PubMed/MEDLINE, Embase, Web of Science, PsycINFO, and the Cochrane Library. The search covered articles published from January 1, 2000, to May 15, 2025. This systematic review is registered with the PROSPERO database under (CRD420251056774).
Flowchart of study search and selection process.
The search strategy combined three concept blocks using appropriate Boolean operators: (1) ECS terms (e.g. “endocannabinoid,” “cannabinoid receptor,” “CB1,” “CB2,” “anandamide,” “2-AG,” “fatty acid amide hydrolase (FAAH),” “Monoacylglycerol lipase (MAGL),” “CBD”); (2) prodromal/high-risk terms (e.g. “prodromal,” “prodrome,” “CHR,” “ultra-high risk (UHR),” “at-risk mental state (ARMS),” “APS,” “first-episode psychosis (FEP)”); and (3) psychosis terms (e.g. “psychosis,” “schizophrenia,” “psychotic disorder”).
Selection criteria
Studies were included if they met the following criteria: (1) focused on the ECS or cannabinoid compounds; (2) included individuals at CHR for psychosis, with prodromal symptoms, FEP patients (particularly studies examining early intervention or biological markers relevant to prodromal states), or used relevant animal models of prodromal psychosis/early intervention; (3) reported original data on biomarkers, neuroimaging findings, genetic associations, or intervention outcomes; and (4) were published in English in peer-reviewed journals. FEP studies were included if they met at least one of the following specific criteria: (a) examined patients within 6 months of psychosis onset to capture neurobiological changes proximal to the prodromal-psychotic transition; (b) investigated biomarkers or interventions specifically designed for early intervention (within the first 2 years of illness onset); (c) compared FEP patients to CHR individuals within the same study; or (d) examined neurobiological mechanisms (e.g. ECS alterations, genetic polymorphisms) that were directly relevant to understanding prodromal pathophysiology based on established theoretical models.
Studies investigating CBD were included given its relevance to cannabinoid-based therapeutics, though CBD’s mechanisms of action extend significantly beyond ECS modulation to include effects on dopaminergic, serotonergic, and other neurotransmitter systems.
Exclusion criteria were: (1) studies exclusively on chronic psychosis without relevance to prodromal states; (2) epidemiological studies of cannabis use without biological or intervention data; (3) case reports or series with fewer than 10 participants; (4) review articles without original data (although these were examined for relevant citations); and (5) studies focusing exclusively on recreational cannabis use without specific assessment of cannabinoid compounds or ECS function.
Study selection and data extraction
All identified records were imported into EndNote X9 for duplicate removal. Two independent reviewers (V.R. and A.S.) systematically screened titles and abstracts using the predefined eligibility criteria. Full texts of potentially eligible articles were then evaluated independently by the same reviewers. All disagreements were resolved through discussion or, when necessary, consultation with a third reviewer (G. Mar), ensuring an unbiased selection process. The inter-rater reliability was high (Cohen’s kappa = 0.87), indicating strong agreement between reviewers.
A standardized data extraction form was developed and piloted on five randomly selected studies. The comprehensive list of variables extracted from each study type is detailed in Appendix 2. Using this form, two reviewers (V.R. and D.D.B.) independently extracted the following data: (1) study characteristics (authors, year, country, design, sample size); (2) participant characteristics (age, gender, high-risk criteria, duration of symptoms); (3) ECS measures or intervention details; (4) outcome measures; (5) main findings; and (6) methodological considerations. For preclinical studies, additional data on animal models, interventions, and translational relevance were extracted. Data extraction discrepancies were identified in 8.7% of cases and were resolved through discussion with a senior researcher (G.M.). This dual-reviewer approach for both study selection and data extraction minimized the risk of selection bias and ensured the accuracy and completeness of the extracted data.
Quality assessment
Study quality was assessed using tools appropriate to each study type: (1) the Newcastle–Ottawa Scale for observational studies; (2) the Cochrane Risk of Bias tool for randomized controlled trials; (3) the NIH Quality Assessment Tool for Before-After Studies for open-label trials; and (4) the SYRCLE risk of bias tool for animal studies. Two reviewers independently conducted quality assessments, with disagreements resolved through discussion.
Data synthesis
Given the heterogeneity in study designs, populations, and outcome measures, a narrative synthesis approach was employed. Studies were categorized into four main groups: (1) preclinical studies (n = 8), (2) genetic and peripheral biomarker studies (n = 7), (3) neuroimaging studies (n = 6), and (4) clinical trials (n = 7). Within each category, evidence was further organized by specific ECS components (e.g. CB1/CB2 receptors, endocannabinoids, metabolic enzymes) or intervention types (e.g. CBD, synthetic cannabinoids, enzyme inhibitors). When sufficient data were available, effect sizes were reported or calculated to facilitate comparisons across studies.
Results
Study selection
The initial database search yielded 1427 records. After removing duplicates, 872 articles were screened based on titles and abstracts, resulting in 112 full-text articles assessed for eligibility. After applying inclusion and exclusion criteria, 22 studies were included in the final analysis (Figure 1). These comprised 8 preclinical studies, 5 genetic/biomarker studies, 5 neuroimaging studies, and 4 clinical trials or intervention studies.
Quality assessment of included studies
The methodological quality of the 26 included studies was systematically evaluated using appropriate tools for each study design. Overall, 11 studies (52%) were rated as high quality with low risk of bias across most domains, while 9 studies (39.1%) were of moderate quality with methodological limitations that warrant caution in interpretation. Two papers described study protocols and were not subject to quality assessment. For preclinical studies, the most common methodological concerns included inadequate reporting of randomization procedures and lack of blinding. Clinical trials generally demonstrated robust methodology, particularly the randomized controlled trials, though open-label studies predictably showed higher risk of performance bias. The strength of evidence was considered in the synthesis and interpretation of findings, with greater weight given to results from methodologically rigorous studies (Table 1).
Quality assessment of included studies.
Selection: randomization, control of confounding factors, sample appropriateness.
Performance: blinding of experimenters and participants, standardized procedures.
Detection: outcome assessment blinding, appropriateness of measures.
Reporting: completeness of results, selective outcome reporting.
Based on the results that emerged from our analysis, we subdivided our findings into several key domains: Preclinical Evidence of ECS Alterations in Models Relevant to Prodromal Psychosis, Genetic and Peripheral Biomarker Evidence in CHR Populations, Neuroimaging Studies of the ECS in CHR States, and Clinical Trials of Cannabinoid-Based Interventions in Prodromal Psychosis. This domain-based organization allowed us to systematically evaluate the evidence across different research methodologies and synthesize a comprehensive understanding of the ECS’s role in prodromal psychosis and its potential as a therapeutic target.
Preclinical evidence of ECS alterations in models relevant to prodromal psychosis
Preclinical studies provide foundational evidence for ECS involvement in psychosis-relevant pathophysiology through three complementary approaches: neurodevelopmental models that simulate early-life risk factors, circuit-specific investigations of neural network dysfunction, and developmental timing studies that identify critical intervention windows (Table 2).
Summary of preclinical studies examining ECS alterations in models relevant to prodromal psychosis.
MIA: Maternal immune activation; THC: Δ-9-tetrahydrocannabinol; 2-AG: 2-arachidonoylglycerol; COX-2: Cyclooxygenase 2; IBA-1: Ionized calcium-binding adapter molecule 1; CBD: Cannabidiol; CB1: Cannabinoid receptor type 1; FAAH: Fatty acid amide hydrolase; MAM: Methylazoxymethanol acetate; PPI: Prepulse inhibition; mRNA: Messenger RNA expression; NMDA: N-methyl-D-aspartate; GFAP: Glial fibrillary acidic protein; MAGL: Monoacylglycerol lipase; VTA: Ventral tegmental area; PFC: Prefrontal cortex.
Neurodevelopmental models
Santoni et al. (2023) demonstrate that Maternal Immune Activation (MIA) in rats affects the ECS during adolescence, particularly in the mesolimbic pathway MIA offspring displayed altered locomotor activity in response to THC and showed a higher bursting activity of Ventral Tegmental Area (VTA) dopamine neurons with a lack of response to cumulative doses of THC. Consistently, MIA adolescent offspring showed enhanced 2-AG-mediated synaptic plasticity and decreased monoacylglycerol lipase activity specifically in mesolimbic structures. Moreover, they displayed higher expression of cyclooxygenase 2 (COX-2) and ionized calcium-binding adapter molecule 1 (IBA-1), markers associated with latent inflammation and persistent microglial activity in the mesolimbic system. These findings reveal specific mechanisms by which prenatal inflammation disrupts endocannabinoid signaling through decreased MAGL activity and enhanced 2-AG-mediated plasticity in the VTA, leading to altered dopamine neuron bursting patterns and abnormal behavioral responses to THC, which may contribute to psychosis vulnerability in adulthood.
N-methyl-D-aspartate (NMDA) receptor hypofunction models, which simulate glutamatergic abnormalities in psychosis, also demonstrated significant neuroinflammatory changes relevant to ECS function. Chronic NMDA antagonist (MK-801) administration in rodents impaired performance in social interaction tests (modeling negative symptoms such as social withdrawal) and novel object recognition tasks (modeling cognitive deficits associated with schizophrenia) (Gomes et al., 2015). This treatment increased Glial fibrillary acidic protein (GFAP)-positive astrocytes in the medial prefrontal cortex and the percentage of Iba-1-positive microglia cells with reactive phenotype in the medial Prefrontal cortex (mPFC) and dorsal hippocampus, indicating enhanced neuroinflammatory responses. Importantly, CBD treatment attenuated both the social interaction deficits and cognitive impairments, as well as the glial activation changes, with effects comparable to the atypical antipsychotic clozapine, suggesting that anti-inflammatory properties may contribute to cannabinoid-mediated therapeutic effects in psychosis models.
Further supporting the therapeutic potential of ECS modulation, Ballmaier et al. (2007) conducted a preclinical study investigating whether cannabinoid CB1 receptor antagonists (rimonabant and AM251) could reverse sensorimotor gating deficits in rodent models of psychosis. Using the prepulse inhibition (PPI) paradigm, researchers demonstrated that both CB1 antagonists significantly counteracted phencyclidine-induced PPI disruption, comparable to the effects of the atypical antipsychotic clozapine. Rimonabant also attenuated PPI disruption caused by dizocilpine and apomorphine. These findings suggest that CB1 receptor antagonism produces an “atypical” antipsychotic profile, potentially by restoring disturbed ECS-glutamate interactions.
Circuit-specific alterations
Hajós et al. (2008) demonstrated that pharmacological activation of CB1 receptors with the agonist CP-55940 (a synthetic cannabinoid agonist) significantly disrupted auditory sensory gating in the hippocampus and entorhinal cortex, and impaired theta and gamma neuronal oscillations in limbic circuitry. These CB1 receptor-mediated disruptions in auditory sensory gating and network oscillations were reversed by the CB1 antagonist AM-251 (a selective CB1 receptor antagonist), and FAAH inhibition during adolescence prevented the emergence of psychosis-like phenotypes. This study demonstrates the direct relationship between ECS overactivation and psychosis-relevant neural circuit dysfunction.
In a translational model targeting the “aberrant salience” hypothesis of psychosis, Szkudlarek et al. (2019) used direct intra-prefrontal cortex infusions to investigate the differential effects of THC and CBD on executive function and affective behaviors in rats. Acute intra-Prefrontal Cortex (PFC) infusions produced anxiogenic effects without impairing executive function, while acute intra-PFC CBD impaired attentional set-shifting and spatial working memory without affecting anxiety or sociability. Importantly, CBD reversed cognitive impairments induced by glutamatergic antagonism and blocked THC’s anxiogenic properties, suggesting CBD’s therapeutic effects may be specific to pathologically aberrant PFC states. The key finding here is that the study demonstrated that THC and CBD effects were mediated through dissociable receptor mechanisms: THC via CB1 receptors and CBD via 5-HT1A receptors.
Developmental timing and interventions
Stark et al. (2020) demonstrated in a methylazoxymethanol acetate (MAM) neurodevelopmental model that peripubertal CBD administration (30 mg/kg from postnatal days 19–39), but not in adulthood, prevented MAM-induced upregulation of dopamine D2 and D3 receptor Messenger RNA expression (mRNA) expression in prefrontal cortex, hippocampus, and nucleus accumbens. This age-dependent effect also normalized regional blood flow changes detected by magnetic resonance imaging. The study suggests that CBD’s therapeutic effects may involve preferential binding to dopamine D3 receptors, where molecular modeling predicted CBD could act as a partial agonist.
Collectively, these preclinical studies demonstrate that ECS disruption occurs early in psychosis-relevant pathological processes, with inflammation, glutamatergic dysfunction, and sensorimotor gating deficits all showing ECS involvement. Importantly, CBD emerges as a consistent protective intervention across multiple models, suggesting broad therapeutic potential that depends critically on developmental timing.
Genetic and peripheral biomarker evidence in CHR populations
Translating preclinical findings to human populations, genetic and biomarker studies provide evidence for ECS alterations in CHR individuals, offering potential pathways for personalized intervention and risk stratification.
Genetic variation in ECS components
Studies examining genetic variations in ECS components have identified several risk variants relevant to prodromal psychosis (Table 3). Suárez-Pinilla et al. (2015) found that polymorphisms in the CNR1 gene (encoding CB1 receptors) were associated with both symptom severity and transition risk in a cohort of 315 CHR individuals followed for 24 months. Specifically, the rs1049353 A allele (characterized as an “at-risk allele”) was associated with higher transition rates (hazard ratio 1.7, 95% confidence interval (CI) 1.2–2.4) and more severe attenuated positive symptoms at baseline, though the specific functional effects of this polymorphism on CB1 receptor activity were not detailed in the study.
Genetic studies of ECS components in CHR populations.
FEP: First-episode psychosis; CHR: Clinical high-risk; HR: Hazard ratio; CI: Confidence interval; OR: Odds ratio; CNR1: Cannabinoid receptor type 1 (gene encoding CB1 receptor); AKT1: AKT serine/threonine kinase 1; EU-GEI: European Network of National Schizophrenia Networks Studying Gene–Environment Interactions; COMT: Catechol-O-methyltransferase; BDNF: Brain-derived neurotrophic factor.
Di Forti et al. (2012) conducted a comprehensive case–control study investigating gene–environment interactions between AKT serine/threonine kinase 1 (AKT1) genetic variation and cannabis use in 489 FEP patients and 278 controls. The study examined the rs2494732 single nucleotide polymorphism and found that while this variant alone was not associated with psychosis risk, it significantly modulated the relationship between cannabis use and psychosis development. Specifically, C/C genotype carriers with a history of cannabis use showed a more than twofold increased likelihood of psychotic disorder (odds ratio (OR) 2.18, 95% CI 1.12–4.31) compared to T/T carriers who used cannabis. Most strikingly, among daily cannabis users, C/C carriers demonstrated a sevenfold increase in psychosis odds compared to T/T carriers (OR 7.23, 95% CI 1.37–38.12), representing a profound gene-environment interaction. Building on this gene-environment framework, Di Forti et al. (2019) conducted a large-scale multicenter case-control study across Europe (European Network of National Schizophrenia Networks Studying Gene–Environment Interactions [EU-GEI]) examining how cannabis use contributes to variation in psychotic disorder incidence. This study provided broader epidemiological context for understanding cannabis-psychosis relationships across different populations and geographic regions, further supporting the importance of genetic factors in modulating cannabis-related psychosis risk identified in their earlier work.
Examining additional genetic pathways Mané et al. (2017) investigated the relationship between Catechol-O-methyltransferase (COMT) Val158Met and Brain-derived neurotrophic factor (BDNF) Val66Met polymorphisms, cannabis use, and age at psychosis onset in 260 Caucasian FEP patients. The study found that early cannabis use and the presence of the met-allele from the BDNF Val66Met polymorphism were significantly associated with earlier age at psychosis onset. Early cannabis use was also significantly associated with male gender, highlighting sex-specific differences in cannabis use patterns and their impact on psychosis emergence.
Peripheral endocannabinoid levels
Complementing genetic findings, peripheral biomarker studies offer insights into dynamic ECS changes during prodromal states, four studies have examined peripheral endocannabinoid levels as potential biomarkers in CHR states (Table 4). Koethe et al. (2009) found significantly elevated cerebrospinal fluid anandamide levels in 27 individuals with initial prodromal states compared to 81 healthy controls. Notably, patients with lower anandamide levels showed a higher risk for earlier transition to psychosis, suggesting a protective role of the ECS in early schizophrenia. This anandamidergic upregulation during the initial prodromal course may represent a compensatory mechanism against psychosis development.
Peripheral endocannabinoid biomarker studies in CHR individuals.
FEP: First-episode psychosis; CHR: Clinical high-risk; UHR: Ultra-high risk; AEA: Anandamide; 2-AG: 2-arachidonoylglycerol; IPS: Inferior parietal sulcus; CSF: Cerebrospinal fluid.
Supporting this protective mechanism hypothesis, Leweke et al. (2007) measured anandamide levels in cerebrospinal fluid and serum from 47 first-episode, antipsychotic-naive schizophrenic patients and 81 healthy volunteers, with subjects further categorized by cannabis use frequency. The study found that schizophrenic low-frequency cannabis users exhibited more than 10-fold higher CSF anandamide levels compared to schizophrenic high-frequency users, healthy low-frequency users, and healthy high-frequency users (all p < 0.01). CSF anandamide levels correlated negatively with disease symptoms in both user groups. These findings suggest that frequent cannabis exposure may down-regulate anandamide signaling in the CNS of schizophrenic patients, potentially representing an important mechanism through h which cannabis impacts mental health in vulnerable individuals.
Extending these findings to trauma-related risk factors Appiah-Kusi et al. (2020b) compared endocannabinoid levels between 33 CHR individuals and 58 healthy controls, examining the impact of childhood trauma. Results revealed significantly elevated N-palmitoylethanolamine and anandamide levels in CHR participants with childhood trauma compared to controls without trauma. Endocannabinoid levels increased linearly with the number of risk factors (CHR status and childhood trauma) and correlated with both CHR symptom severity and childhood maltreatment extent. These findings suggest that ECS alterations may be particularly pronounced in those with multiple risk factors for psychosis, potentially reflecting a neurobiological interaction between developmental trauma and psychosis vulnerability.
However, challenging these consistent findings of elevated endocannabinoids, Joaquim et al. (2022) conducted a comprehensive analysis of plasma endocannabinoid levels in 91 UHR individuals compared to healthy controls, with a 3-year follow-up to assess clinical outcomes. This study found significantly decreased plasma concentrations of both AEA (p = 0.003) and 2-AG (p < 0.001) in UHR subjects compared to controls. Notably, this contrasts with some previous findings of elevated endocannabinoid levels in prodromal states. The study followed participants for 3 years, during which 16 individuals developed psychiatric complaints, though endocannabinoid levels did not significantly differentiate between remitters (those who recovered) and non-remitters (those with persistent symptoms). Weak correlations were observed between clinical parameters and plasma endocannabinoid concentrations in those who developed psychiatric complaints.
These conflicting biomarker findings highlight the complexity of ECS alterations during prodromal states and underscore the need for standardized measurement protocols and longitudinal assessment approaches.
Neuroimaging studies of the ECS in CHR states
Bridging preclinical findings with human clinical applications, neuroimaging studies provide direct evidence of ECS alterations in human high-risk populations, spanning both structural measures of receptor availability and functional responses to cannabinoid challenge.
CB1 receptor availability studies
Positron emission tomography (PET) imaging studies using CB1 receptor-specific radioligands have provided direct evidence of altered cannabinoid receptor availability in CHR populations (Table 5). Borgan et al. (2019) conducted a cross-sectional PET study using two independent cohorts to investigate CB1 receptor availability in FEP without confounds of chronicity or substance use. Study 1 (Finland) included 7 patients taking antipsychotics and 11 controls using [18F]FMPEP-d2, while study 2 (UK) included 20 antipsychotic-naive patients and 20 controls using [11C]MePPEP. Both studies found significantly reduced CB1 receptor availability in patients across multiple brain regions (anterior cingulate cortex, hippocampus, striatum, and thalamus) compared to controls. Lower CB1 availability was associated with greater symptom severity and poorer cognitive functioning, suggesting CB1 receptors may represent a potential therapeutic target for psychotic disorders.
Neuroimaging studies of the ECS in prodromal psychosis.
ECS: Endocannabinoid system; CB1: Cannabinoid receptor type 1; FEP: First-episode psychosis; FAAH: Fatty acid amide hydrolase; CBD: Cannabidiol; fMRI: Functional magnetic resonance imaging; CHR: Clinical high risk; THC: Δ-9-tetrahydrocannabinol; PANSS: Positive and Negative Syndrome Scale; FMPEP: N-(2-fluoromethyl)-N-(1-propylphenyl)-piperidin-3-amine; MePPEP: N-(2-methyl-phenyl)-N-(1-propylphenyl)-piperidin-3-amine; ACC: Anterior cingulate cortex.
Examining ECS-targeted interventions in a different population, D’Souza et al. (2019) conducted a double-blind, placebo-controlled trial testing the FAAH inhibitor PF-04457845 in 70 men with cannabis use disorder. The study found that PF-04457845 significantly reduced cannabis withdrawal symptoms during hospitalization and decreased cannabis use at 4 weeks compared to placebo, suggesting that potentiating endocannabinoid signaling through FAAH inhibition may represent an effective treatment approach for cannabis use disorder.
Transitioning to CBD interventions in CHR populations, a dose-finding study by Davies et al. (2020) compared the effects of a single dose of CBD (600 mg) on neural response during fearful face processing in 33 CHR (CHR for psychosis) patients. Results showed that CHR subjects treated with placebo exhibited greater activation in the parahippocampal gyrus but reduced activation in the striatum compared to healthy controls. In CHR patients who received CBD, activation in these regions was intermediate between the placebo group and control group, suggesting that CBD modulates brain function in regions implicated in psychosis risk and emotion processing, with effects that appear to be task-independent.
Response to cannabinoid challenge
O’Neill et al. (2021) conducted a double-blind, randomized, crossover study investigating CBD’s acute effects on brain function in 15 patients with established psychosis. Participants underwent Functional magnetic resonance imaging (fMRI) during a verbal paired associate learning task after receiving either 600 mg CBD or placebo on separate days. Compared to healthy controls, psychosis patients under placebo showed altered prefrontal activation during encoding and altered mediotemporal/prefrontal activation during recall. CBD partially normalized these dysfunctions, with activation under CBD being intermediate between placebo and healthy controls. CBD also attenuated hippocampal-striatal functional connectivity and produced trend-level symptom reduction.
Bhattacharyya et al. (2010) conducted a randomized, double-blind, within-subject study in 15 healthy volunteers examining the opposite neural effects of THC (10 mg) and CBD (600 mg) using fMRI during multiple cognitive tasks. The study revealed that THC and CBD had opposite effects on brain activation in multiple regions: striatum during verbal memory, amygdala during fear processing, parahippocampal cortex during response inhibition, and temporal/occipital cortices during sensory processing. THC induced psychotic symptoms and anxiety while CBD showed anxiolytic trends. In a separate behavioral experiment with six volunteers, CBD pretreatment significantly blocked THC-induced psychotic symptoms (Positive and Negative Syndrome Scale positive score reduced from 13–9, p < 0.05).
Collectively, these neuroimaging studies demonstrate that ECS dysfunction is detectable in human CHR populations through reduced CB1 receptor availability and altered neural responses. Importantly, CBD consistently shows normalizing effects on brain function across different populations and tasks, supporting its therapeutic potential.
Clinical trials of cannabinoid-based interventions in prodromal psychosis
Cannabidiol trials
Translating the convergent preclinical, genetic, and neuroimaging evidence into clinical applications, three clinical trials investigating cannabinoid-based interventions in CHR populations were identified. It should be noted that while CBD is included in this review due to its cannabinoid origins and relevance to cannabis-psychosis research, its therapeutic effects likely involve multiple mechanisms beyond ECS modulation, including significant interactions with dopaminergic pathways that may be primary to its antipsychotic properties (Table 6). Bhattacharyya et al. (2018) conducted a randomized, double-blind, placebo-controlled trial investigating CBD’s neurocognitive mechanisms in 33 antipsychotic-naive individuals at CHR of psychosis. Participants received either a single oral dose of 600 mg CBD (n = 16) or placebo (n = 17) before undergoing fMRI during a verbal learning task. CHR participants receiving placebo showed reduced brain activation compared to healthy controls in the right caudate during encoding and in the parahippocampal gyrus and midbrain during recall. CBD treatment resulted in activation levels that were intermediate between the placebo and control groups, suggesting that CBD may partially normalize brain dysfunction in regions critical to psychosis pathophysiology.
Clinical trials of cannabinoid-based interventions in prodromal psychosis.
CHR: Clinical high-risk; CBD: Cannabidiol; FAAH: Fatty acid amide hydrolase; RDT: Risky Decision-making Task; AA: Arachidonic acid.
Appiah-Kusi et al. (2020a) conducted a randomized, placebo-controlled study investigating CBD’s effects on stress response in 32 CHR participants and 26 healthy controls. Half of the CHR participants received 600 mg/day CBD for 1 week, while half received placebo, before undergoing the Trier Social Stress Test. Results showed that CHR participants on placebo had blunted cortisol reactivity compared to healthy controls, while those receiving CBD demonstrated intermediate cortisol responses between placebo and control groups. Similarly, CBD-treated CHR participants showed intermediate levels of anxiety and stress responses during public speaking compared to placebo-treated CHR participants and healthy controls.
Other ECS-targeted interventions
Examining genetic factors that may predict ECS-targeted intervention responses, Mayo et al. (2020) conducted a genetic study examining the FAAH 385C->A polymorphism in 423 individuals, focusing on 75 participants (25 per genotype group: two copies of the C allele (homozygous for C) [CC], C allele and one A allele (heterozygous) [AC], two copies of the A allele (homozygous for A) [AA]). The study investigated whether individuals with the FAAH 385A allele, which encodes reduced FAAH enzyme activity, would show elevated anandamide levels and enhanced fear extinction. Results demonstrated that the A allele was dose-dependently associated with elevated basal anandamide levels, facilitated fear extinction, and enhanced extinction recall. AA homozygotes were protected against stress-induced decreases in anandamide and negative emotional consequences of stress. A parallel humanized mouse model confirmed these protective effects against stress-induced anandamide decreases in both peripheral and brain regions (amygdala, prefrontal cortex), suggesting that elevated anandamide signaling may aid treatment for stress-related psychiatric disorders like Post Traumatic Stress Disorder (PTSD).
In summary, the clinical trial evidence, while limited, demonstrates that CBD shows promise for normalizing both brain function and stress reactivity in CHR populations, while genetic studies point toward personalized approaches based on individual ECS-related genetic variations. However, the small sample sizes and methodological heterogeneity limit definitive conclusions.
Discussion
This systematic review reveals converging evidence across preclinical models, biomarker studies, neuroimaging investigations, and clinical trials that the ECS is disrupted in prodromal psychosis and represents a promising therapeutic target. Several key patterns emerge from the data synthesis.
First, ECS alterations appear to precede and potentially contribute to the development of psychotic symptoms and associated neurobiological abnormalities. Preclinical models consistently demonstrate that disruptions in CB1 receptor signaling and endocannabinoid tone occur during critical developmental windows that correspond to prodromal-like states, with timing that often precedes dopaminergic and glutamatergic changes. Human studies support this temporal relationship, with genetic variations in ECS components modifying psychosis risk and peripheral endocannabinoid alterations correlating with symptom severity and transition outcomes.
Second, the evidence suggests a complex, non-linear relationship between endocannabinoid signaling and psychosis risk. Both hyper- and hypo-function of CB1 signaling appear problematic, depending on neural circuit, developmental stage, and environmental context. This complexity may explain the seemingly paradoxical effects of different cannabinoid compounds: THC (a partial CB1 agonist) typically exacerbates psychosis risk, while CBD and FAAH inhibitors (which indirectly enhance endocannabinoid signaling) appear protective. This suggests that the manner in which the ECS is modulated, rather than simple up or down-regulation, is critical for therapeutic effects.
Third, the evidence points to several mechanistic pathways through which ECS-targeted interventions may exert protective effects in prodromal states:
Stress buffer hypothesis: Multiple studies demonstrate that proper endocannabinoid signaling, particularly anandamide tone, buffers against stress-induced neurobiological changes that contribute to psychosis vulnerability. Both CBD and FAAH inhibitors appear to normalize stress reactivity and Hypothalamic–pituitary–adrenal (HPA) axis function in preclinical models and UHR individuals.
Neuroinflammatory modulation: Several studies identified anti-inflammatory effects of CBD and endocannabinoid tone enhancement, which may counteract the neuroinflammatory processes implicated in psychosis progression. Preclinical models show that CBD normalizes microglial activation patterns and inflammatory markers in stress-exposed and genetically vulnerable animals.
Neurodevelopmental stabilization: Evidence from preclinical studies suggests that appropriately timed ECS-targeted interventions may stabilize neurodevelopmental trajectories during vulnerable periods, potentially preventing the structural and functional brain changes that precede psychosis. This may explain the apparent disease-modifying effects observed with CBD treatment.
Dopamine-glutamate interface: The ECS functions at the intersection of dopaminergic and glutamatergic signaling, two systems centrally implicated in psychosis pathophysiology. Several studies demonstrate that normalization of endocannabinoid signaling reduces aberrant dopamine release and improves NMDA receptor function, potentially addressing core pathophysiological processes.
Clinical implications
The findings from this systematic review offer preliminary insights that may inform future clinical approaches for individuals at CHR for psychosis, though current evidence remains insufficient for definitive practice recommendations.
Cannabis use assessment
The evidence strongly supports enhanced assessment of cannabis use patterns in CHR populations. Given the differential effects of THC and CBD observed across studies—with THC consistently associated with increased psychosis risk while CBD shows potential protective effects—detailed characterization of cannabis use (including THC/CBD ratios, potency, quantity per session, and timing of initiation) may significantly inform risk assessment and treatment planning. The contrasting neurobiological effects of THC (through CB1 receptor activation) and CBD (through D2 receptor modulation and other non-CB1 mechanisms) suggest that when abstinence cannot be achieved, harm reduction approaches focused on reducing THC exposure might warrant investigation.
The neuroimaging evidence from O’Neill et al. (2021) and Bhattacharyya et al. (2018) demonstrates that CBD can modulate neural activity in psychosis-relevant brain regions such as the medial temporal lobe and striatum, though translation to clinical practice requires further validation. Additionally, the developmental timing of cannabis exposure appears critical, with several studies indicating that adolescent use during key neurodevelopmental periods confers greater risk than adult-onset use, suggesting the need for developmentally-targeted prevention efforts.
CBD as a potential intervention
The preliminary results from CBD trials suggest it warrants investigation as a potential treatment option for CHR individuals, particularly given its favorable safety profile compared to antipsychotics. The studies by Bhattacharyya et al. (2018) and Appiah-Kusi et al. (2020a) demonstrate CBD’s effects on brain function and stress reactivity with effect sizes comparable to antipsychotics (d = 0.88 in the Bhattacharyya study) but without metabolic, extrapyramidal, or prolactin-related adverse effects. CBD also appears to address anxiety symptoms—frequently co-occurring in prodromal states—and normalize stress reactivity through modulation of the hypothalamic-pituitary-adrenal axis, as demonstrated by Mayo et al. (2020).
However, these findings require replication in larger, longer-term trials before clinical implementation. Preliminary preclinical evidence from Stark et al. (2020) suggests CBD might have disease-modifying potential through stabilizing neurodevelopmental trajectories, though this remains highly speculative and requires confirmation in human studies. Current evidence suggests doses of 600 mg/day show tolerability, but optimal dosing regimens, treatment duration, and long-term safety profiles remain unclear.
Genetic and biomarker considerations
The identification of genetic variations in ECS components represents very early-stage exploratory research that may eventually inform personalized medicine approaches, though clinical utility remains highly speculative. For example, individuals with the rs1049353 A allele in the CNR1 gene, which Suárez-Pinilla et al. (2015) found associated with higher transition rates, might theoretically benefit from targeted interventions, though this requires substantial further research and remains purely investigational. Similarly, the gene-environment interactions demonstrated by Di Forti et al. (2012, 2019), where AKT1 genotype significantly modulated cannabis-related psychosis risk, suggest potential future research directions for genetic screening to identify high-risk individuals, though practical implementation is not currently feasible.
Peripheral endocannabinoid measures show preliminary and exploratory promise as potential biomarkers, with studies by Koethe et al. (2009) and Leweke et al. (2007) demonstrating elevated anandamide levels in prodromal states that inversely correlated with symptom severity. However, conflicting findings—particularly Joaquim et al. (2022) reporting decreased endocannabinoid levels in a larger cohort—highlight the highly preliminary nature of this research and the need for standardized measurement protocols before any clinical utility can be established. It is crucial to emphasize that both genetic and peripheral biomarker approaches remain in the very early stages of research, with significant methodological challenges and conflicting findings that preclude any current clinical application. The contradictory endocannabinoid findings across studies highlight the exploratory nature of this work and the substantial research needed before any clinical utility can be established.
Future research and implementation considerations
The evidence reviewed suggests that comprehensive ECS assessment—through detailed cannabis use characterization and potentially genetic screening—may warrant investigation as future research directions rather than immediate components of CHR evaluation protocols. Treatment algorithms for specialized services may eventually be investigated to include CBD as an early intervention option, particularly for individuals declining antipsychotics or showing poor tolerance, though this requires substantial additional research and remains highly speculative.
Any future implementation of ECS-targeted approaches would require careful ethical consideration and robust informed consent processes, clearly communicating the highly experimental nature of these interventions. The association between cannabinoid compounds and recreational cannabis use necessitates thoughtful patient and family education to address potential misconceptions and ensure appropriate expectations regarding treatment outcomes.
While these findings suggest promising directions for future research and clinical development, current evidence should be considered hypothesis-generating rather than practice-changing. The clinical implications, while encouraging, must be tempered by the limited size, duration, and consistency of current clinical trials. Larger, multicenter trials with longer follow-up periods are essential before ECS-targeted approaches can be recommended for routine clinical use in prodromal psychosis. The summary of key evidence on the role of the ECS in prodromal psychosis and clinical implications derived from this systematic review are summarized in Figure 2.
Summary of evidence for ECS involvement in prodromal psychosis.
Limitations and research gaps
Several important limitations and research gaps should be considered when interpreting these findings. These limitations significantly impact the strength and generalizability of our conclusions and must be carefully considered when evaluating the evidence presented throughout this review.
First, most clinical trials of cannabinoid-based interventions have been relatively small (typically fewer than 35 participants) and of short duration (ranging from single doses to 1–2 weeks), limiting conclusions about long-term efficacy and safety. These small sample sizes reduce statistical power and limit the ability to detect clinically meaningful effects or rare adverse events. The promising results require confirmation in larger, multicenter trials with longer follow-up periods before any clinical recommendations can be considered.
Second, heterogeneity in high-risk criteria, outcome measures, and intervention parameters (doses, formulations, duration) complicates direct comparisons across studies. This methodological heterogeneity means that apparently convergent findings may actually reflect different populations, interventions, or outcome measures, limiting the strength of our synthetic conclusions. Standardization of these elements would strengthen future research.
Third, the mechanistic understanding of how CBD and other ECS-targeted interventions exert their effects remains incomplete. CBD has multiple pharmacological actions beyond the ECS, including effects on serotonin receptors, G protein–coupled receptor 55 (GPR55), and TRPV1 channels, making it difficult to attribute therapeutic effects specifically to ECS modulation. This mechanistic uncertainty limits our ability to predict optimal dosing, identify potential responders, or develop rational combination therapies.
Fourth, the developmental timing of interventions appears critical based on preclinical evidence, but human studies have not adequately addressed age-dependent effects. The lack of age stratification in clinical trials means that potentially critical developmental windows for intervention remain unidentified. Future trials should consider stratification by age and developmental stage to optimize intervention timing.
Fifth, the relationship between exogenous cannabinoid effects (THC, CBD) and endogenous ECS function remains incompletely characterized. More research is needed to understand how these compounds interact with the endogenous system, particularly in the context of genetic variations and environmental stressors. This knowledge gap particularly affects the interpretation of biomarker studies and limits the development of personalized treatment approaches.
Most importantly, the limited clinical trial evidence to date—characterized by small sample sizes, short durations, and variable methodologies—necessitates considerable caution in translating these findings into clinical practice. Current evidence should be considered hypothesis-generating rather than definitive, requiring substantial replication and extension before firm treatment recommendations can be established.
Finally, a significant limitation of this systematic review is the heterogeneity of high-risk populations across studies. We included research examining various high-risk designations—CHR, UHR, ARMS, and individuals with APS—which, while overlapping, represent populations identified through different assessment tools and criteria. These varied designations potentially capture individuals at different stages of prodromal progression and with different risk profiles for transition to psychosis. For instance, the CHR designation typically uses the Structured Interview for Prodromal Syndromes/Scale of Prodromal Symptoms, while the UHR category often employs the Comprehensive Assessment of At-Risk Mental States, with subtle but potentially important differences in inclusion thresholds and symptom categorization. Leweke et al. (2007) employed UHR criteria with emphasis on basic symptoms, while Appiah-Kusi et al. (2020b) focused on attenuated positive symptoms using CHR criteria, potentially capturing different neurobiological stages of prodromal psychosis. Furthermore, baseline transition risk varies considerably across studies (from 15% to 35% at 2 years), reflecting differences in recruitment strategies, exclusion criteria, and demographic factors. This heterogeneity complicates direct comparison of findings, particularly regarding biomarkers and treatment responses. This heterogeneity significantly complicates direct comparison of findings, particularly regarding biomarkers and treatment responses, and may explain some of the contradictory results observed across studies. The fact that different studies report varying levels of endocannabinoid alterations may reflect genuine biological differences between these subpopulations rather than inconsistencies in methodology. Additionally, recent evidence suggests that these high-risk categories may identify individuals on diverse trajectories, including those who will develop non-psychotic disorders or remain persistently at-risk without transition. The inclusion of FEP studies alongside prodromal populations represents a methodological decision that warrants acknowledgment. While FEP occurs after the prodromal period has concluded, we included FEP studies that met specific, pre-defined criteria for relevance to prodromal states. This approach was justified by the temporal proximity of FEP to the prodromal-psychotic transition and the fact that early intervention strategies often target this transition phase. However, this inclusion may have introduced heterogeneity in our sample and should be considered when interpreting findings.
Mechanistic considerations beyond the ECS
This review focuses on ECS-related mechanisms in prodromal psychosis, however it is crucial to acknowledge that CBD—one of the most promising interventions identified—likely exerts its therapeutic effects through mechanisms that extend well beyond ECS modulation. Recent neuroimaging and pharmacological evidence suggests that CBD’s antipsychotic properties may be primarily mediated through dopamine D2 receptor interactions rather than cannabinoid receptor pathways (Seeman, 2016). Specifically, CBD appears to act as a functional antagonist at D2 receptors in regions such as the striatum and prefrontal cortex, areas critically implicated in psychosis pathophysiology.
This D2 receptor activity is particularly relevant to the dopamine-glutamate interface normalization discussed earlier in our review. CBD appears to function as a partial D2 agonist with preferential activity in regions showing hyperactivity, potentially explaining its ability to normalize aberrant dopamine release without causing the motor side effects associated with traditional D2 antagonists. This mechanism may be central to CBD’s capacity to improve NMDA receptor function and restore the balance between dopaminergic and glutamatergic signaling that is disrupted in prodromal psychosis states. Unlike full D2 antagonists used in conventional antipsychotics, CBD’s partial agonism may provide a more nuanced modulation of dopaminergic activity, selectively dampening excessive signaling while preserving normal dopaminergic function.
Additionally, CBD demonstrates significant activity at serotonin 5-HT1A receptors, which may contribute to its anxiolytic effects observed in CHR populations, and interactions with TRPV1 channels and GPR55 receptors that could influence neuroinflammatory processes. This dopaminergic mechanism, combined with its effects on 5-HT1A receptors and potential indirect ECS modulation, positions CBD as a multi-target therapeutic that addresses the complex neurotransmitter imbalances characteristic of prodromal psychosis through complementary pathways.
This mechanistic complexity suggests that CBD belongs in discussions of cannabinoid-based therapeutics due to its plant origins and relevance to cannabis-psychosis research, yet its therapeutic profile may be better understood as a multi-target compound rather than a selective ECS modulator. This mechanistic diversity may actually represent an advantage in treating prodromal psychosis, where multiple neurotransmitter systems are likely dysregulated. However, it also complicates the interpretation of CBD studies within purely ECS-focused frameworks and suggests that the promising results observed with CBD may not necessarily validate ECS modulation as the primary therapeutic mechanism in prodromal psychosis intervention.
Contrasting findings and mechanistic uncertainties
Our review identified several important discrepancies in the literature that warrant further discussion. The most notable inconsistency concerns the precise mechanism by which CBD exerts its apparent antipsychotic effects. Within our included studies, Bhattacharyya et al. (2018) demonstrated that CBD partially normalized alterations in parahippocampal, striatal, and midbrain function in CHR individuals, though the study focused on functional outcomes rather than proposing specific receptor mechanisms. This highlights a fundamental mechanistic uncertainty: while our review focuses on ECS alterations in prodromal psychosis, CBD’s beneficial effects appear to operate largely independent of direct ECS modulation. The question remains whether CBD’s therapeutic profile results from compensating for ECS dysfunction through alternative pathways, or whether its effects are entirely unrelated to the ECS alterations observed in prodromal states. This complexity is exemplified by the fact that while THC (a CB1 agonist) typically increases psychosis risk, CBD appears to exert beneficial effects primarily through non-ECS mechanisms such as D2 receptor modulation, despite both compounds being derived from the same plant.
Adding to this mechanistic complexity, studies examining endocannabinoid biomarkers show contradictory patterns. Leweke et al. (2007) found elevated anandamide levels in CSF of prodromal patients, which inversely correlated with symptom severity, suggesting a potential compensatory mechanism. Conversely, Borgan et al. (2019) demonstrated reduced CB1 receptor availability in CHR individuals, suggesting downregulation or altered receptor function. These seemingly contradictory findings suggest a more nuanced relationship between endocannabinoid signaling and psychosis than simple hyper- or hypo-function models.
The mechanistic complexity of CBD extends beyond simple receptor interactions to include potential downstream effects on multiple signaling cascades. As noted by O’Neill et al. (2021), CBD’s effects extend beyond the ECS to include actions on multiple neurotransmitter systems and signaling pathways. For instance, CBD’s effects on D2 receptors may indirectly influence dopamine-mediated modulation of endocannabinoid release, creating complex feedback loops that are difficult to disentangle. Similarly, its 5-HT1A receptor activity may affect stress-related endocannabinoid signaling through HPA axis modulation. This multi-target profile is further illustrated by Mayo et al. (2020), who directly compared a FAAH inhibitor with CBD in UHR individuals and found different efficacy profiles, with FAAH inhibition primarily reducing anxiety while CBD showed broader effects on APS. These interconnected pathways make it challenging to determine which mechanisms are primary versus secondary to CBD’s therapeutic effects, and whether the compound’s multi-target nature is essential for its efficacy or represents pharmacological redundancy around a core mechanism of action.
These mechanistic uncertainties are compounded by inconsistent clinical findings. While CBD showed significant effects on positive symptoms, its impact on negative symptoms and cognitive deficits—often more persistent and disabling features of psychosis—was less consistent and sometimes non-significant. O’Neill et al. (2021) reported variable individual responses to CBD, with some participants showing minimal improvement despite group-level positive effects. Additionally, optimal dosing remains unclear, with studies using a relatively narrow range (typically 600 mg/day) without systematic dose-finding investigations. Some studies also reported mild adverse effects including sedation, gastrointestinal discomfort, and medication interactions that, while not severe, deserve more prominent discussion. The lack of clear dose-response relationships and variable individual responses suggests that CBD’s effects may depend on complex interactions between multiple neurotransmitter systems that vary across individuals based on genetic background, disease stage, or environmental factors. Without understanding which of CBD’s many molecular targets are responsible for specific therapeutic effects, it becomes difficult to optimize dosing, predict treatment response, or develop rational combination therapies.
A particularly striking example of these inconsistencies is evident in peripheral endocannabinoid studies. While Koethe et al. (2009) and Leweke et al. (2007) found elevated anandamide and 2-AG levels in prodromal populations, Joaquim et al. (2022) reported significantly decreased plasma concentrations of both endocannabinoids in a larger UHR cohort (n = 91) with longer follow-up (3 years). This direct contradiction highlights fundamental challenges in endocannabinoid biomarker research, potentially reflecting differences in analytical methodologies, sample timing relative to symptom onset, medication exposure, or the heterogeneous nature of high-risk populations captured by different assessment tools.
Collectively, these mechanistic uncertainties highlight a fundamental challenge in cannabinoid research: the complexity of determining causal relationships between specific molecular targets and clinical outcomes when dealing with compounds that act on multiple receptor systems. The divergent effects of THC and CBD, despite both being cannabis-derived compounds, underscore the importance of mechanistic specificity rather than simply targeting “the ECS” as a therapeutic approach. This mechanistic ambiguity complicates the interpretation of CBD studies within ECS-focused frameworks and raises questions about whether ECS dysfunction is a primary therapeutic target or simply one component of a broader neurobiological disruption.
Future directions
Several promising research directions emerge from this systematic review:
Long-term safety and effectiveness studies: A critical research priority is to conduct longitudinal studies examining the long-term safety and tolerability of CBD and other ECS-targeted interventions. Current evidence on CBD safety is largely derived from short-term trials (typically 7–12 weeks), leaving significant gaps in our understanding of extended treatment effects. Future research should assess hepatic function, drug-drug interactions (particularly with commonly prescribed psychiatric medications), hormonal effects, and potential impacts on brain development in young adults. Additionally, given the potential for CBD to modulate multiple neurotransmitter systems, monitoring for delayed-onset adverse effects or tolerance development is essential. Studies should also investigate optimal dosing regimens, therapeutic windows, and whether intermittent dosing strategies might mitigate potential long-term risks while maintaining efficacy.
Longitudinal biomarker studies: Prospective studies combining peripheral endocannabinoid measurements, neuroimaging, and clinical assessments could better characterize the dynamic changes in ECS function throughout the prodromal period and identify optimal intervention windows. These studies should also systematically monitor safety parameters and potential biomarkers of adverse effects alongside efficacy measures to develop a comprehensive risk-benefit profile specific to different CHR subpopulations.
Novel ECS-targeted compounds: Beyond CBD, more selective compounds targeting specific components of the ECS (e.g. selective FAAH or MAGL inhibitors, allosteric modulators of CB1/CB2) warrant investigation as potential interventions with improved specificity. Development of these compounds should incorporate rigorous safety monitoring, including assessment of psychiatric adverse effects—particularly important given the history of serious psychiatric side effects with some cannabinoid receptor antagonists like rimonabant. Comparative safety studies between different ECS modulators could help identify compounds with optimal therapeutic indices.
Combination approaches: Studies examining combinations of CBD with other interventions (e.g. cognitive-behavioral therapy, neuroprotective agents, anti-inflammatory compounds) could identify synergistic effects that maximize benefits while minimizing risks. These combinations might allow dose reduction of CBD, potentially mitigating any dose-dependent adverse effects while maintaining or enhancing therapeutic efficacy. Safety assessments should be a primary outcome in these combination trials, not merely a secondary consideration.
Personalized medicine approaches: Research integrating genetic, biomarker, and clinical data to predict which individuals would most benefit from ECS-targeted interventions could improve treatment selection and outcomes. Crucially, these approaches should also identify individuals at heightened risk for adverse effects based on factors such as genetic polymorphisms in cannabinoid metabolism pathways, pre-existing medical conditions, or concomitant medications. Development of safety algorithms alongside efficacy algorithms is essential for truly personalized medicine. -Withdrawal and discontinuation effects: Studies specifically designed to assess potential withdrawal or discontinuation effects following extended CBD or ECS-targeted treatment are needed. While current evidence suggests minimal dependence potential for CBD, the ECS’s role in homeostasis and adaptation warrants careful investigation of potential rebound phenomena or neuroadaptations following prolonged modulation.
These future research directions acknowledge both the therapeutic promise of ECS-targeted interventions and the imperative to establish their comprehensive safety profiles, particularly for long-term use in young, vulnerable populations at critical neurodevelopmental stages. A balanced approach that simultaneously advances efficacy and safety science will be essential for responsible clinical implementation.
Conclusion
This systematic review provides substantive evidence that the ECS represents a mechanistically relevant and clinically promising therapeutic target in prodromal psychosis. Alterations in ECS function appear to precede and potentially contribute to psychosis development, while targeted interventions, particularly CBD, show favorable efficacy and safety profiles in preliminary clinical trials. The evidence supports a model in which appropriate modulation of endocannabinoid signaling may buffer against stress-induced neurobiological changes, stabilize neurodevelopmental trajectories, and address core pathophysiological processes involved in psychosis progression.
While significant research gaps remain, the convergence of findings across preclinical models, biomarker studies, neuroimaging investigations, and clinical trials provides a strong foundation for further development of ECS-targeted interventions. These approaches have the potential to address the critical need for safer, more effective treatments for individuals at clinical high risk for psychosis, potentially altering disease trajectory during a critical window of opportunity. Future research should focus on confirming preliminary findings in larger trials, refining the understanding of mechanisms and optimal treatment parameters, and developing personalized approaches that maximize benefits for this vulnerable population.
Footnotes
Appendix 1
PRISMA checklist for systematic review on the ECS as a therapeutic target in prodromal psychosis.
| Section and topic | Item number | Checklist item | Location where item is reported |
|---|---|---|---|
| TITLE | |||
| Title | 1 | Identify the report as a systematic review. | Page 1 |
| ABSTRACT | |||
| Abstract | 2 | See the PRISMA 2020 for Abstracts checklist. | Page 1 |
| INTRODUCTION | |||
| Rationale | 3 | Describe the rationale for the review in the context of existing knowledge. | Pages 3–4 |
| Objectives | 4 | Provide an explicit statement of the objective(s) or question(s) the review addresses. | Page 4 |
| METHODS | |||
| Eligibility criteria | 5 | Specify the inclusion and exclusion criteria for the review and how studies were grouped for the syntheses. | Pages 4–5 |
| Information sources | 6 | Specify all databases, registers, websites, organizations, reference lists, and other sources searched or consulted to identify studies. Specify the date when each source was last searched or consulted. | Page 5 |
| Search strategy | 7 | Present the full search strategies for all databases, registers, and websites, including any filters and limits used. | Page 5, Supplementary Table 1 |
| Selection process | 8 | Specify the methods used to decide whether a study met the inclusion criteria of the review, including how many reviewers screened each record and each report retrieved, whether they worked independently, and if applicable, details of automation tools used in the process. | Page 5 |
| Data collection process | 9 | Specify the methods used to collect data from reports, including how many reviewers collected data from each report, whether they worked independently, any processes for obtaining or confirming data from study investigators, and if applicable, details of automation tools used in the process. | Pages 5–6 |
| Data items | 10a | List and define all outcomes for which data were sought. Specify whether all results that were compatible with each outcome domain in each study were sought. | Page 6 |
| 10b | List and define all other variables for which data were sought. Describe any assumptions made about any missing or unclear information. | Page 6 | |
| Study risk of bias assessment | 11 | Specify the methods used to assess risk of bias in the included studies, including details of the tool(s) used, how many reviewers assessed each study and whether they worked independently. | Page 6 |
| Effect measures | 12 | Specify for each outcome the effect measure(s) used in the synthesis or presentation of results. | Page 6 |
| Synthesis methods | 13a | Describe the processes used to decide which studies were eligible for each synthesis. | Page 6 |
| 13b | Describe any methods required to prepare the data for presentation or synthesis, such as handling of missing summary statistics or data conversions. | Page 6 | |
| 13c | Describe any methods used to tabulate or visually display results of individual studies and syntheses. | Page 6 | |
| 13d | Describe any methods used to synthesize results and provide a rationale for the choice(s). If meta-analysis was performed, describe the model(s), method(s) to identify the presence and extent of statistical heterogeneity, and software package(s) used. | Page 6 | |
| 13e | Describe any methods used to explore possible causes of heterogeneity among study results. | Not applicable | |
| 13f | Describe any sensitivity analyses conducted to assess robustness of the synthesized results. | Not applicable | |
| Reporting bias assessment | 14 | Describe any methods used to assess risk of bias due to missing results in a synthesis. | Page 6 |
| Certainty assessment | 15 | Describe any methods used to assess certainty in the body of evidence for an outcome. | Page 6 |
| RESULTS | |||
| Study selection | 16a | Describe the results of the search and selection process, from the number of records identified in the search to the number of studies included in the review, ideally using a flow diagram. | Page 7, Figure 1 |
| 16b | Cite studies that might appear to meet the inclusion criteria, but which were excluded, and explain why they were excluded. | Page 7 | |
| Study characteristics | 17 | Cite each included study and present its characteristics. | Pages 7–9, Tables 1–6 |
| Risk of bias in studies | 18 | Present assessments of risk of bias for each included study. | Pages 7, Table 1 |
| Results of individual studies | 19 | For all outcomes, present for each study: (a) summary statistics for each group and (b) an effect estimate and its precision, ideally using structured tables or plots. | Pages 8–16, Tables 2–6 |
| Results of syntheses | 20a | For each synthesis, briefly summarize the characteristics and risk of bias among contributing studies. | Pages 7–16 |
| 20b | Present results of all statistical syntheses conducted. If meta-analysis was done, present for each the summary estimate and its precision and measures of statistical heterogeneity. | Not applicable | |
| 20c | Present results of all investigations of possible causes of heterogeneity among study results. | Not applicable | |
| 20d | Present results of all sensitivity analyses conducted to assess the robustness of the synthesized results. | Not applicable | |
| Reporting biases | 21 | Present assessments of risk of bias due to missing results for each synthesis assessed. | Page 7 |
| Certainty of evidence | 22 | Present assessments of certainty in the body of evidence for each outcome assessed. | Pages 16–18 |
| DISCUSSION | |||
| Discussion | 23a | Provide a general interpretation of the results in the context of other evidence. | Pages 18–20 |
| 23b | Discuss any limitations of the evidence included in the review. | Pages 21–22 | |
| 23c | Discuss any limitations of the review processes used. | Page 22 | |
| 23d | Discuss implications of the results for practice, policy, and future research. | Pages 22–24 | |
| OTHER INFORMATION | |||
| Registration and protocol | 24a | Provide registration information for the review, including register name and registration number, or state that the review was not registered. | Not applicable |
| 24b | Indicate where the review protocol can be accessed, or state that a protocol was not prepared. | Not applicable | |
| 24c | Describe and explain any amendments to information provided at registration or in the protocol. | Not applicable | |
| Support | 25 | Describe sources of financial or non-financial support for the review, and the role of the funders or sponsors in the review. | Page 24 |
| Competing interests | 26 | Declare any competing interests of review authors. | Page 24 |
| Availability of data, code, and other materials | 27 | Report which of the following are publicly available and where they can be found: template data collection forms; data extracted from included studies; data used for all analyses; analytic code; any other materials used in the review. | Not applicable |
Appendix 2
Variables extracted from studies for systematic review on the ECS as a therapeutic target in prodromal psychosis.
| All studies | Preclinical studies | Clinical and neuroimaging studies | Genetic/biomarker studies |
|---|---|---|---|
| Author, year | Study design | Study design | Study design |
| Country/region | Species/animal model | Sample size and characteristics | Sample size and characteristics |
| Study design | Age/developmental stage | Inclusion/exclusion criteria | Inclusion/exclusion criteria |
| Sample size | Model type (e.g. neurodevelopmental, NMDA hypofunction) | Control group (if applicable) | Control group (if applicable) |
| Participant characteristics | Intervention/manipulation | Assessment methods | Assessment methods |
| Prodromal psychosis definition/criteria | ECS component examined | Neuroimaging techniques | Genetic variants/polymorphisms studied |
| ECS measurement methods | Outcome measures | Outcome measures | Endocannabinoid levels measured |
| Main outcomes measured | Behavioral assessments | Intervention type (if applicable) | Measurement techniques |
| Key findings | Biochemical/molecular assessments | Duration of intervention (if applicable) | Statistical analyses |
| Effect sizes (when reported) | Histology/imaging techniques | Outcome measurement tools | Effect estimates and precision |
| Quality assessment | Statistical analyses | Statistical analyses | Correlation with symptoms |
| Methodological considerations | Timing of effects | Confounding variables controlled | Risk factors identified |
| Limitations | Developmental considerations | Attrition rates (if applicable) | Genetic–environmental interactions |
| Proposed mechanisms | Translational relevance | Effect estimates and precision | Predictive value for transition |
| Therapeutic implications | Sex differences | Safety/tolerability data | Biomarker validity measures |
| Funding source | Strain differences | CB1 receptor binding potential | Peripheral vs. central measures |
| Dose-response effects | Brain regions examined | Sensitivity/specificity (when reported) | |
| Pharmacological properties | Functional connectivity findings | Baseline vs. follow-up measures | |
| Methodological quality | Symptom correlations | Time course of biomarker changes | |
| Control conditions | Side effects reported | Association with cannabis use | |
| Technical considerations | Dropouts and reasons | Measurement reliability | |
| Route of Administration | Cognitive/functional outcomes | Methodological quality | |
| Funding source | Funding source | Funding source |
PRISMA: Preferred reporting items for systematic reviews and meta-analyses; ECS: Endocannabinoid system; CB1 receptor: Cannabinoid receptor type 1; NMDA: N-methyl-D-aspartate.
Author contributions
VR carried out the comprehensive literature search and systematic review process. VR and FC wrote the manuscript, with AS providing substantial contribution to the data extraction and analysis. GM and GMA critically reviewed the manuscript and provided expert input on the neuropharmacological and clinical aspects, respectively. All authors approved the final version of the manuscript.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was conducted as part of the academic responsibilities of the authors without specific grant funding.
Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: None of the authors has received funding or have financial relationships with pharmaceutical companies producing or marketing cannabinoid-based products or endocannabinoid system modulators. None of the authors has been involved in developing drugs related to the endocannabinoid system or has patents in this area.