Psychosis and Substance Use Comorbidity: Cannabis, Stimulants, Diagnostic Challenges, and Integrated Treatment Approaches

The co-occurrence of psychotic disorders and substance use disorders (SUDs) represents one of the most clinically challenging intersections in psychiatry. These conditions share overlapping neurobiological substrates, confound each other's diagnosis, worsen each other's prognosis, and demand integrated treatment models that most healthcare systems are poorly equipped to deliver. The relationship between substances and psychosis is not unidirectional — substance use can precipitate psychosis, psychotic disorders can drive substance use, and shared genetic and neurodevelopmental vulnerabilities can give rise to both independently.

Epidemiological data consistently demonstrates that rates of SUDs among individuals with schizophrenia-spectrum disorders are dramatically elevated compared to the general population. The Epidemiologic Catchment Area (ECA) study found that approximately 47% of individuals with schizophrenia met criteria for a comorbid SUD at some point in their lifetime, compared to roughly 16% in the general population — an odds ratio of approximately 4.6. Cannabis and stimulants (amphetamines, methamphetamine, cocaine) are among the substances most consistently implicated in psychosis onset, exacerbation, and diagnostic confusion.

More recent data from NIMH-funded studies and international registries have refined these estimates. Among first-episode psychosis (FEP) cohorts, cannabis use disorder is present in 25–45% of cases depending on the population studied, while stimulant use disorders are found in 10–25%. These figures are not merely epidemiological curiosities — they carry direct implications for treatment response, relapse risk, functional outcome, and mortality. This article examines the neurobiology, diagnostic complexities, and evidence-based treatment approaches for psychosis–substance use comorbidity with a focus on cannabis and stimulants.

The relationship between cannabis and psychosis is mediated primarily through the endocannabinoid system (ECS) and its interactions with mesolimbic and mesocortical dopamine pathways. The principal psychoactive constituent of cannabis, Δ9-tetrahydrocannabinol (THC), acts as a partial agonist at cannabinoid type 1 (CB1) receptors, which are densely expressed in the prefrontal cortex, hippocampus, amygdala, basal ganglia, and cerebellum. CB1 receptors are located presynaptically on GABAergic and glutamatergic terminals, meaning THC modulates the balance of inhibition and excitation across critical cortical and subcortical circuits.

The psychotogenic mechanism of THC involves several converging pathways:

• Dopaminergic dysregulation: THC administration increases dopamine release in the nucleus accumbens and striatum, as demonstrated by PET imaging studies using [¹¹C]raclopride. This mirrors the dopamine hypothesis of schizophrenia, which posits that mesolimbic hyperdopaminergia underlies positive symptoms. However, THC also reduces dopamine synthesis capacity in the striatum with chronic use, potentially contributing to negative symptoms and amotivation.
• Glutamatergic disruption: CB1 activation on cortical glutamatergic neurons reduces glutamate release in the prefrontal cortex, contributing to NMDA receptor hypofunction — a mechanism increasingly recognized as central to schizophrenia pathophysiology. This dovetails with the glutamate hypothesis and may explain why cannabis-associated psychosis shares phenomenological features with both dopaminergic and glutamatergic models.
• GABAergic disinhibition: THC reduces GABA release from interneurons in the hippocampus and cortex, disrupting gamma oscillatory activity and impairing the precise timing of neural circuits essential for working memory, sensory gating, and reality monitoring.
• Neuroinflammation: Chronic cannabis use activates microglial cells and increases pro-inflammatory cytokines, a process that intersects with neuroinflammatory models of schizophrenia.

The role of cannabidiol (CBD) adds complexity. Unlike THC, CBD does not directly activate CB1 receptors and may have antipsychotic properties, potentially through modulation of anandamide metabolism (via FAAH inhibition), 5-HT1A partial agonism, and PPARγ activation. The landmark trial by McGuire et al. (2018) demonstrated that CBD (1000 mg/day) as an adjunct to antipsychotics was associated with significant reductions in positive symptoms compared to placebo in patients with schizophrenia. The progressive shift in the cannabis market toward high-THC, low-CBD strains (with THC concentrations increasing from ~4% in the 1990s to 15–25% or higher in modern products) is therefore of significant clinical concern.

Genetic moderators are critical. The most replicated finding involves the catechol-O-methyltransferase (COMT) gene. The Val158Met polymorphism affects dopamine catabolism in the prefrontal cortex: individuals homozygous for the Val allele (Val/Val), who have higher COMT activity and lower prefrontal dopamine, show increased vulnerability to THC-induced psychotic symptoms. The seminal study by Caspi et al. (2005) demonstrated a gene–environment interaction whereby adolescent cannabis use significantly increased the risk of developing schizophreniform disorder in Val/Val carriers but not in Met/Met carriers. Subsequent studies have yielded mixed replication, but the broader principle — that genetic background moderates cannabis-psychosis risk — is well-supported by polygenic risk score analyses showing that individuals with high schizophrenia polygenic risk scores are disproportionately affected by cannabis exposure.

Stimulant-induced psychosis (SIP) provides perhaps the most direct pharmacological evidence supporting the dopamine hypothesis of psychosis. Amphetamines, methamphetamine, and cocaine all increase synaptic dopamine availability, though through distinct mechanisms: amphetamines reverse the dopamine transporter (DAT) and promote vesicular release, while cocaine blocks DAT reuptake. The psychotogenic potential of these substances correlates with the magnitude and speed of dopamine elevation in the mesolimbic pathway.

Key neurobiological features of stimulant-induced psychosis include:

• Striatal dopamine flooding: PET studies show that amphetamine-induced dopamine release in the striatum is significantly greater in individuals who develop psychotic symptoms compared to those who do not. The magnitude of [¹¹C]raclopride displacement directly correlates with positive symptom severity.
• Dopamine sensitization: Repeated stimulant exposure produces progressive sensitization of the mesolimbic dopamine system, meaning that equivalent or smaller doses produce increasingly robust dopamine release over time. This behavioral sensitization model (Robinson & Berridge, 1993) has been proposed as a mechanism for the transition from substance-induced psychosis to persistent psychotic disorder.
• Neurotoxicity: Methamphetamine is particularly neurotoxic, producing oxidative stress, mitochondrial dysfunction, and structural damage to dopaminergic and serotonergic terminals in the striatum and prefrontal cortex. Neuroimaging studies show reduced DAT density and gray matter volume in chronic methamphetamine users, changes that overlap substantially with findings in schizophrenia.
• Glutamate and GABA disruption: Beyond dopamine, stimulants alter glutamate signaling in the prefrontal cortex–striatal–thalamic circuit, and chronic use is associated with reduced prefrontal GABA concentrations measured by magnetic resonance spectroscopy (MRS).

The prevalence of stimulant-induced psychosis is substantial. Epidemiological reviews estimate that 10–60% of methamphetamine users will experience at least transient psychotic symptoms during active use, with the wide range reflecting variation in dose, duration, route of administration, and population characteristics. Among individuals presenting with methamphetamine-associated psychosis, approximately 25–30% will develop a persistent psychotic disorder that continues beyond the period of intoxication and withdrawal — a conversion rate that has significant public health implications, particularly in regions with high methamphetamine prevalence such as parts of Southeast Asia, Australasia, and the Western United States.

Cocaine-induced psychosis follows similar dopaminergic principles but tends to be more transient, likely reflecting cocaine's shorter half-life and the absence of the same degree of neurotoxicity. However, kindling phenomena have been described with cocaine, wherein repeated episodes of cocaine-induced psychosis become progressively easier to trigger and longer in duration.

One of the most clinically important questions in this field is: How often does substance-induced psychosis convert to a primary psychotic disorder? The answer carries enormous implications for monitoring, early intervention, and patient counseling.

The most comprehensive data on this question comes from the landmark Danish register-based cohort study by Starzer et al. (2018), published in the American Journal of Psychiatry, which followed over 6,788 individuals diagnosed with substance-induced psychosis over periods of up to 20 years. Key findings included:

• Cannabis-induced psychosis: 47.4% converted to a schizophrenia-spectrum disorder (primarily schizophrenia or schizoaffective disorder) or bipolar disorder over the follow-up period. This was the highest conversion rate among all substances.
• Amphetamine-induced psychosis: 30.3% converted to a persistent psychotic or bipolar disorder.
• Overall conversion rate across substances: 32.2% received a diagnosis of schizophrenia-spectrum or bipolar disorder.
• Median time to conversion: 3.1 years for cannabis-induced psychosis, 4.4 years for amphetamine-induced psychosis.

A systematic review and meta-analysis by Murrie et al. (2020) corroborated these findings across 35 studies, reporting pooled transition rates of 34% from cannabis-induced psychosis to schizophrenia and 22% from amphetamine-induced psychosis to schizophrenia. The overall pooled transition rate from any substance-induced psychosis to schizophrenia was 25% (95% CI: 18–34%).

These conversion rates are strikingly high and argue strongly against dismissing substance-induced psychosis as a benign, self-limited condition. They also raise the question of whether many cases of "substance-induced" psychosis are in fact prodromal presentations of primary psychotic disorders that are unmasked or accelerated by substance use, rather than de novo pharmacological events.

Risk factors for conversion from substance-induced psychosis to schizophrenia-spectrum disorder include:

• Younger age at first psychotic episode
• Family history of psychotic disorders
• Longer duration of initial psychotic episode
• Prominent negative symptoms during the substance-induced episode
• Continued substance use after the initial episode
• Male sex (modestly increased risk)
• Cannabis as the primary substance (vs. other substances)

Distinguishing substance-induced psychotic disorder (SIPD) from a primary psychotic disorder with comorbid substance use is one of the most difficult differential diagnostic tasks in clinical psychiatry. The distinction matters enormously for prognosis, treatment planning, duration of antipsychotic treatment, and patient education — yet the clinical tools available to make this distinction are imprecise.

The DSM-5-TR defines substance/medication-induced psychotic disorder (F1x.x59) as psychotic symptoms that develop during or soon after substance intoxication or withdrawal, are not better explained by an independent psychotic disorder, and do not occur exclusively during delirium. The critical qualifier is that the clinician must determine whether the symptoms are in excess of what would be expected from intoxication alone and whether they represent an independent psychotic process. The ICD-11 parallels this in its category of substance-induced psychotic disorders (6C4x.5) but offers similarly limited operational guidance for distinguishing primary from induced psychosis.

Practical diagnostic pitfalls include:

• Temporal overlap: Many patients with emerging schizophrenia begin using substances during the prodromal phase, making it impossible to establish clear temporal precedence. Cannabis use often precedes the first psychotic break by months to years, creating the appearance of causation when the relationship may be bidirectional or confounded.
• Persistent symptoms during abstinence: The canonical diagnostic marker — that substance-induced psychosis resolves with sustained abstinence — requires a period of confirmed sobriety that is often difficult to achieve or verify clinically. DSM-5-TR suggests that symptoms persisting beyond approximately one month after cessation of substance use should raise suspicion for a primary psychotic disorder, but this timeline is an approximation. Methamphetamine-associated psychosis can persist for months after last use.
• Phenomenological similarities: The positive symptoms of substance-induced psychosis (paranoid delusions, auditory hallucinations, referential thinking) overlap substantially with those of schizophrenia. Some distinguishing features have been proposed — for instance, visual hallucinations and tactile hallucinations are more common in stimulant-induced psychosis, while third-person auditory hallucinations and formal thought disorder are more characteristic of schizophrenia — but these are statistical tendencies, not diagnostic criteria.
• Negative and cognitive symptoms: The presence of prominent negative symptoms (avolition, flat affect, alogia) and cognitive deficits (working memory, executive function impairment) tends to favor a primary psychotic disorder, as substance-induced psychosis typically features predominantly positive symptoms. However, chronic stimulant use can produce pseudo-negative symptoms related to dopamine depletion.
• Polysubstance use: Many patients use multiple substances simultaneously, making it impossible to attribute psychotic symptoms to a single agent. This is the rule rather than the exception in clinical practice.

The diagnostic uncertainty inherent in this distinction means that provisional diagnoses are common, and longitudinal reassessment is essential. Many clinicians adopt a pragmatic approach: treating the psychosis aggressively regardless of presumed etiology while simultaneously addressing substance use, then reassessing the diagnostic formulation after a sustained period of abstinence.

Biomarker research has yet to yield clinically useful tools for this distinction. Some studies have examined inflammatory markers (CRP, IL-6), HPA axis function (cortisol), and neuroimaging signatures (prefrontal gray matter volume, striatal DAT availability), but none have achieved sufficient sensitivity or specificity for individual-level diagnostic application.

The pharmacological treatment of psychosis in the context of active or recent substance use requires careful consideration of both efficacy and tolerability, as patients with comorbid SUDs often respond differently to antipsychotic medications and face a distinct side-effect profile.

Antipsychotic Selection

The evidence base for antipsychotic choice in dual-diagnosis patients is considerably weaker than for schizophrenia alone. The CATIE trial (Clinical Antipsychotic Trials of Intervention Effectiveness), while not specifically designed for dual-diagnosis populations, included a significant proportion of participants with comorbid SUDs (~37%), and post hoc analyses revealed important patterns. Olanzapine showed the longest time to discontinuation overall but was associated with significant metabolic side effects, while perphenazine was roughly as effective as the second-generation agents for many outcomes.

Several principles guide antipsychotic selection in this population:

• Clozapine has the strongest evidence for reducing substance use in patients with schizophrenia and comorbid SUD. Multiple studies, including a randomized trial by Brunette et al. (2006), have shown that clozapine is associated with reduced cannabis, alcohol, and cocaine use compared to other antipsychotics. The proposed mechanism involves clozapine's unique pharmacological profile — including relatively low D2 occupancy, potent 5-HT2A/2C antagonism, and effects on the noradrenergic system — which may reduce the rewarding effects of substances and address the anhedonia that drives self-medication. However, clozapine is reserved for treatment-resistant cases due to the risk of agranulocytosis (cumulative incidence ~0.8%), metabolic syndrome, and the need for regular blood monitoring, which is particularly challenging in this population.
• Long-acting injectable (LAI) antipsychotics are increasingly favored for dual-diagnosis patients due to improved adherence. Paliperidone palmitate and aripiprazole lauroxil have pharmacokinetic profiles that maintain therapeutic drug levels even with inconsistent follow-up. The PROACTIVE trial and similar studies have demonstrated that LAI formulations reduce relapse rates by approximately 30–40% compared to oral equivalents in nonadherent populations.
• Aripiprazole, a D2 partial agonist, has theoretical advantages in substance-comorbid psychosis because it stabilizes dopamine transmission without producing the complete D2 blockade that can worsen anhedonia and drive compensatory substance use. Some studies suggest it may reduce craving, though the evidence remains preliminary.
• Avoid highly sedating or metabolically burdensome agents when possible, as weight gain, metabolic syndrome, and sedation disproportionately affect this population's already compromised physical health and engagement in treatment.

Treatment Outcomes

Response rates to antipsychotic treatment in dual-diagnosis patients are generally lower than in schizophrenia without SUD. While approximately 60–70% of patients with first-episode schizophrenia respond to initial antipsychotic treatment (typically defined as ≥50% reduction in positive symptom scales), response rates in dual-diagnosis populations are estimated at 40–55%, with continued substance use being the primary driver of reduced treatment response and elevated relapse risk.

The number needed to treat (NNT) for preventing relapse with antipsychotic maintenance therapy in schizophrenia is approximately 3–5 (Leucht et al., 2012 meta-analysis). In dual-diagnosis populations, the NNT is likely higher (estimated 5–8), reflecting the added variance introduced by ongoing substance use, medication nonadherence, and treatment disengagement.

Relapse rates at 12 months in dual-diagnosis patients are approximately 50–70%, compared to 25–40% in schizophrenia without SUD, even with adequate pharmacotherapy. This underscores the necessity of psychosocial interventions that target substance use directly.

Unlike the well-established pharmacotherapies for alcohol and opioid use disorders, there are no FDA-approved medications specifically for cannabis or stimulant use disorders, which represents a significant gap in the treatment of dual-diagnosis patients. Nonetheless, several pharmacological strategies are under investigation and some are used off-label in clinical practice.

Cannabis Use Disorder

• N-acetylcysteine (NAC): A modulator of glutamate homeostasis, NAC has shown mixed results. A randomized controlled trial by Gray et al. (2012) in cannabis-dependent adolescents found that NAC (1200 mg twice daily) doubled the odds of negative urine cannabinoid tests compared to placebo (OR = 2.4). However, a subsequent larger trial in adults (Gray et al., 2017) failed to replicate this effect. NAC remains of interest due to its favorable safety profile and potential neuroprotective effects.
• Gabapentin: Some evidence suggests gabapentin may reduce cannabis withdrawal symptoms and use, potentially through GABAergic and calcium channel modulation.
• CBD: As mentioned, CBD has both potential antipsychotic and anti-craving properties, making it a theoretically appealing intervention for dual-diagnosis patients, though rigorous efficacy trials for cannabis use disorder in psychotic populations are lacking.

Stimulant Use Disorder

• No medication has demonstrated consistent, replicated efficacy for methamphetamine or cocaine use disorder in the general population or in dual-diagnosis cohorts. This remains one of the most significant unmet pharmacological needs in addiction medicine.
• Mirtazapine: Has shown modest promise in some trials for methamphetamine use reduction, potentially through 5-HT2/5-HT3 antagonism and noradrenergic effects.
• Combination pharmacotherapy (injectable naltrexone + extended-release oral bupropion): The ADAPT-2 trial (Trivedi et al., 2021) demonstrated a significant reduction in methamphetamine use in the general population (OR for negative urine samples = 1.43, p = 0.003). Applicability to psychotic populations is unknown and requires study, and bupropion carries a theoretical seizure risk that requires caution.
• Contingency management (CM): While not pharmacological, CM has the strongest evidence base for stimulant use disorders (discussed further in the psychosocial section). The NIDA-funded meta-analyses consistently show that CM produces the largest effect sizes of any behavioral treatment for stimulant use (Cohen's d ≈ 0.46–0.62).

The treatment of co-occurring psychosis and substance use disorders requires integrated dual-diagnosis treatment (IDDT) — the simultaneous, coordinated treatment of both disorders within a single treatment team or program, rather than sequential or parallel treatment in separate systems. The evidence for integrated models has accumulated since the 1990s, driven largely by SAMHSA-funded demonstration projects and the landmark work of Drake, Mueser, and colleagues.

Key Components of Integrated Treatment

• Motivational interviewing (MI): Addresses the ambivalence about substance use change that is pervasive in this population. MI is typically Stage 1 in integrated treatment and has moderate effect sizes for treatment engagement (Cohen's d ≈ 0.28–0.35 across addiction populations).
• Cognitive-behavioral therapy for psychosis and substance use (CBTp-SU): Adapted CBT protocols address both psychotic symptoms (reality testing, coping with hallucinations) and substance use (functional analysis, relapse prevention). The MIDAS trial (Barrowclough et al., 2010) randomized 327 patients with schizophrenia-spectrum disorders and substance misuse to MI + CBT vs. treatment as usual. The intervention group showed modest but significant improvements in general functioning and days of substance use at 12 and 24 months, though effects on psychotic symptoms were not significant.
• Contingency management: Provides tangible reinforcements (typically monetary vouchers or prizes) for verified abstinence or treatment attendance. Contingency management has the largest and most consistent evidence base for reducing stimulant use (NNT ≈ 4–7 for achieving sustained abstinence periods). Its application in psychotic populations has been studied less extensively, but available evidence suggests comparable efficacy. Barriers to implementation include funding for incentives and philosophical resistance within treatment systems.
• Family psychoeducation: Critical for reducing expressed emotion, improving medication adherence, and supporting both symptom management and substance abstinence. Meta-analyses of family interventions in schizophrenia show relapse reduction with NNT ≈ 7.
• Assertive community treatment (ACT): Intensive, mobile, multidisciplinary team-based care that is particularly suited to dual-diagnosis patients who are poorly served by traditional outpatient models. Dual-diagnosis ACT teams integrate addiction counseling directly into the team. Evidence shows ACT reduces hospitalization days and improves housing stability, with more equivocal effects on substance use per se.

Outcomes and Limitations

Despite the face validity and broad endorsement of integrated treatment, the outcome evidence is modest in magnitude. A Cochrane review by Hunt et al. (2019) examined psychosocial interventions for dual-diagnosis patients and concluded that while integrated approaches are associated with improved treatment retention compared to standard care, the evidence for superiority on substance use outcomes or psychotic symptom reduction is inconsistent. The review noted significant heterogeneity across studies, small sample sizes, and variable definitions of "integrated treatment." This does not mean integrated treatment is ineffective — rather, the evidence base is insufficiently powered and standardized to detect what may be clinically meaningful but statistically modest effects.

The prognosis of psychosis with comorbid substance use depends on a complex interplay of clinical, demographic, and treatment-related factors. Identifying these factors is essential for risk stratification and treatment intensity allocation.

Factors Associated with Poorer Prognosis

• Continued substance use: By far the strongest and most consistent predictor of poor outcomes. Continued cannabis use after a first psychotic episode approximately doubles the risk of relapse within 12 months. The study by Schoeler et al. (2016), a meta-analysis of 16,565 patients, found that continued cannabis use was associated with significantly increased relapse risk (OR = 1.75), more frequent hospitalizations, greater positive symptom severity, and poorer overall functioning.
• Earlier age of onset of substance use: Adolescent-onset cannabis use (before age 15) confers the highest risk for both conversion to persistent psychosis and poorer long-term outcomes, reflecting the vulnerability of the developing brain during critical neurodevelopmental windows.
• Methamphetamine as the primary substance: Compared to cannabis and other substances, methamphetamine-associated psychosis is associated with more severe neurocognitive deficits (particularly executive function and processing speed), greater structural brain changes, and a more chronic, treatment-refractory course in a subset of patients.
• Longer duration of untreated psychosis (DUP): DUP is a robust predictor of outcome across all psychotic disorders, and it tends to be longer in dual-diagnosis patients due to delays in treatment seeking and diagnostic confusion. Each month of untreated psychosis is associated with incrementally worse symptomatic and functional outcomes.
• Poor medication adherence: Non-adherence rates in dual-diagnosis schizophrenia are estimated at 50–70%, compared to 40–50% in schizophrenia without SUD.
• Homelessness and social isolation: Amplify both substance use and psychotic symptom severity in a reinforcing cycle.

Factors Associated with Better Prognosis

• Substance cessation or significant reduction: Achievable in a meaningful proportion of patients, particularly with integrated treatment, and is consistently associated with improved symptom trajectories.
• Good premorbid functioning: Higher premorbid IQ, better social functioning, and educational attainment prior to illness onset predict better outcomes regardless of substance use status.
• Female sex: Associated with later onset, better treatment response, and better overall functioning in schizophrenia-spectrum disorders, though sex differences may be attenuated in substance-comorbid populations.
• Early engagement in integrated treatment: Enrollment in early intervention programs (e.g., coordinated specialty care) within the first two years of psychosis onset is associated with markedly better outcomes. The RAISE-ETP (Recovery After an Initial Schizophrenia Episode — Early Treatment Program) trial (Kane et al., 2016) demonstrated that coordinated specialty care, which included substance use interventions, significantly improved quality of life, symptoms, and participation in work/school compared to usual community care.
• Absence of significant negative symptoms: Negative symptoms are the strongest predictors of long-term functional disability in schizophrenia and are less responsive to both pharmacological and psychosocial treatment.

The emergence of synthetic cannabinoids (SCs) — sold under names like K2, Spice, and various proprietary labels — has introduced a new dimension to the psychosis–substance use interface. Unlike THC, which is a partial CB1 agonist, most synthetic cannabinoids are full CB1 agonists with binding affinities 10–800 times greater than THC. This pharmacological distinction translates into significantly more severe psychotogenic effects.

Clinical reports and case series consistently describe synthetic cannabinoid-associated psychosis as more severe, more acute in onset, and more likely to involve agitation, disorganization, and bizarre behavior compared to natural cannabis-associated psychosis. Emergency department presentations for SC-related psychosis have increased dramatically since the mid-2010s, particularly among incarcerated populations, homeless individuals, and adolescents.

The diagnostic challenge with SCs is compounded by the fact that standard urine drug screens do not detect most synthetic cannabinoids. Specialized assays are available but are not routinely used in most clinical settings. This means that SC-related psychosis may be misdiagnosed as a primary psychotic disorder or as psychosis related to other substances.

The long-term conversion rates from SC-induced psychosis to persistent psychotic disorders are not yet well-established due to the relative novelty of these substances and the difficulty of tracking users longitudinally, but preliminary data suggest rates that may equal or exceed those associated with natural cannabis.

Similarly, novel stimulants such as synthetic cathinones ("bath salts") — including mephedrone, MDPV, and alpha-PVP — produce psychosis through potent dopamine transporter inhibition and/or reversal. Alpha-PVP, in particular, has been associated with severe, prolonged psychosis with extreme agitation, often described clinically as "excited delirium." These substances present additional diagnostic challenges because they may not be detected on standard toxicology panels.

Despite significant progress, the field of psychosis–substance use comorbidity is marked by substantial evidence gaps and active research frontiers:

• Biomarkers for diagnostic distinction: No clinically validated biomarker currently exists to distinguish substance-induced psychosis from primary psychotic disorders. Research into inflammatory markers (e.g., IL-6, TNF-α), neurocognitive profiles, and neuroimaging signatures (e.g., striatal dopamine synthesis capacity measured via [¹⁸F]DOPA PET) is ongoing. Preliminary data suggest that individuals whose psychosis converts to schizophrenia may have lower baseline prefrontal gray matter volume and higher polygenic risk scores for schizophrenia, but these tools are not ready for clinical deployment.
• Pharmacogenomics: The moderating role of COMT, AKT1, and other genetic variants on cannabis-psychosis risk raises the possibility of individualized risk assessment. The AKT1 rs2494732 polymorphism has been associated with an approximately twofold increased risk of psychosis with daily cannabis use. Integrating polygenic risk with substance exposure history could eventually enable precision prevention strategies.
• CBD as adjunctive treatment: Following the McGuire et al. (2018) trial, several larger trials are underway to assess CBD's efficacy and optimal dosing in both psychotic symptoms and cannabis use reduction. Results from these studies are expected to clarify whether CBD has a clinical role in dual-diagnosis treatment.
• Glutamate-targeted therapies: Given the convergence of cannabis- and stimulant-related psychosis mechanisms on glutamatergic systems, agents targeting mGlu2/3 receptors, glycine transporters (e.g., sarcosine), and D-serine pathways are under investigation for psychosis generally and may have particular relevance for substance-comorbid populations.
• Digital phenotyping and ecological momentary assessment (EMA): Smartphone-based monitoring of real-time substance use, mood, and psychotic experiences is an emerging methodology that could improve both research accuracy and clinical care by capturing the temporal dynamics of substance use and symptom fluctuation in naturalistic settings.
• Implementation science: Perhaps the most urgent frontier. Despite decades of evidence supporting integrated treatment, the majority of mental health systems globally continue to operate in silos, with separate funding streams, training requirements, and clinical cultures for addiction and psychosis services. Translating the integrated treatment evidence into routine clinical practice remains a formidable implementation challenge.

The co-occurrence of psychosis and substance use disorders — particularly cannabis and stimulant use — is common, neurobiologically grounded, diagnostically challenging, and prognostically significant. Clinicians encountering this comorbidity should be guided by the following principles:

• Do not dismiss substance-induced psychosis as benign. Conversion rates to persistent psychotic disorders are 25–47% depending on the substance, with cannabis carrying the highest risk. Longitudinal monitoring is essential.
• Maintain diagnostic humility. The distinction between substance-induced and primary psychotic disorders is often provisional and should be revisited after sustained abstinence. Prominent negative symptoms, cognitive deficits, family history of psychosis, and persistent symptoms beyond one month of abstinence favor a primary psychotic disorder.
• Treat both disorders simultaneously. Integrated dual-diagnosis treatment is the standard of care, though the evidence for its superiority over standard treatment remains modest in magnitude. Sequential treatment ("get clean first, then we'll treat your psychosis" or vice versa) is associated with worse outcomes.
• Consider clozapine earlier in treatment-resistant dual-diagnosis cases. It has the strongest evidence for reducing both psychotic symptoms and substance use.
• Long-acting injectable antipsychotics should be considered early for patients with adherence difficulties, which are the norm rather than the exception in this population.
• Monitor for metabolic complications aggressively. Dual-diagnosis patients face elevated cardiovascular risk from both antipsychotic medications and stimulant use.
• Advocate for system-level change. The persistence of siloed treatment systems for psychosis and addiction is a structural barrier to evidence-based care.