The Road From Early Warning Signs to Full Psychosis—and Where Cannabis Fits | Research

The takeaway

A review of 60 studies mapped how psychosis develops through prodromal stages, finding cannabis confers a 4-fold risk increase (10-fold with genetic vulnerability) among multiple environmental factors.

Psychiatrists specializing in early psychosis, families worried about psychosis risk, and policymakers weighing cannabis and mental health evidence.

What the researchers found

This systematic review synthesized 60 studies spanning 25 years (2000–2025) to map the trajectory from early warning signs to full psychotic episodes. The prodromal period—before diagnosable psychosis—turns out to be more complex than previously understood.

Early symptoms are mostly non-specific: depression (52%), worry (41%), and anxiety (38%). These look nothing like psychosis. Attenuated psychotic symptoms emerge later. Negative symptoms (social withdrawal, flat affect, reduced motivation) typically precede positive symptoms (hallucinations, delusions) by about 12 months.

Three distinct trajectories were identified: persistent (35%), fluctuating (42%), and improving (23%). The fact that nearly a quarter of people at ultra-high risk actually improve underscores that prodromal symptoms don't inevitably lead to psychosis.

Among environmental risk factors, cannabis stood out with specific numbers: a 4-fold increased transition risk, escalating to 10-fold in individuals with genetic vulnerability. Childhood trauma affected 61% of ultra-high-risk individuals. These environmental factors may be particularly important because the review notes that premorbid deterioration may reflect environmental factors more than genetic ones.

Neurobiological prediction is advancing: machine learning using eye-tracking achieved 77.6% accuracy, and multimodal assessment reached AUC = 0.84 for predicting psychosis transition.

Why it matters

If we can identify who's on the path to psychosis early enough, we might be able to intervene. The 4-fold cannabis risk (10-fold with genetic vulnerability) provides specific, quantifiable risk information that's more actionable than vague warnings. And the finding that 23% of at-risk individuals improve naturally helps avoid over-pathologizing early symptoms.

The numbers in context

60 studies reviewed. Early prodromal: depression 52%, worry 41%, anxiety 38%. Negative symptoms precede positive by ~12 months. Trajectories: persistent 35%, fluctuating 42%, improving 23%. Cannabis: 4-fold risk (10-fold with genetic vulnerability). Childhood trauma: 61% of UHR. Machine learning prediction: 77.6% (eye-tracking), AUC 0.84 (multimodal).

How the study worked

Systematic review of 60 studies (2000–2025) following PRISMA guidelines. Included 35 prospective cohorts (58.3%), 9 cross-sectional studies (15%), 9 RCTs (15%), 5 case-control studies (8.3%), and 2 qualitative studies (3.3%).

Who was studied

60 studies (2000-2025) examining prodromal symptoms and psychotic onsets.

What this study cannot tell us

Systematic reviews aggregate existing evidence with all its limitations. The 4-fold cannabis risk figure comes from observational studies that can't prove causation. Prodromal assessment tools may over-identify people who would never develop psychosis. The improving trajectory (23%) complicates intervention decisions. Machine learning prediction tools aren't yet validated for clinical use.

How to read the evidence

Systematic review of 60 studies across multiple designs—rated strong for comprehensiveness, though individual included studies vary in quality.

When this study was published

Published in 2025, synthesizing 25 years of prodromal psychosis research (2000–2025).

The bigger picture

The 4-fold/10-fold cannabis risk quantifies what RTHC-00193 described qualitatively (THC linked to psychosis risk). RTHC-00163's finding that the cannabis-psychosis link disappeared after adjusting for other substances adds an important caveat—the 4-fold figure may be partially confounded. RTHC-00202's genetic study of the CNR1 rs1049353 polymorphism explores one of the genetic vulnerabilities that could turn a 4-fold risk into a 10-fold risk. This review places cannabis within the broader context of multiple interacting risk factors, not as a single cause.

Replication

Not stated in abstract.

Funding

Not reported in abstract.

Conflicts of interest

Not reported in abstract.

Questions still open

Could early cannabis cessation reduce the 4-fold transition risk in prodromal individuals? Should cannabis screening be integrated into early psychosis prevention programs? Can the genetic vulnerability markers that turn 4-fold into 10-fold risk be identified with routine testing?

Read the original research

The psychosis continuum: Systematic review on prodromal markers, symptom progression, and early intervention strategies.

Asian journal of psychiatry, 113, 104725

Asian Journal of Psychiatry is a reputable journal focusing on psychiatric research and clinical practice.

View on PubMed ↗Publisher’s paper ↗

Citation

Ricci, Valerio; Sarni, Alessandro; Martinotti, Giovanni; Maina, Giuseppe. (2025). The psychosis continuum: Systematic review on prodromal markers, symptom progression, and early intervention strategies.. Asian journal of psychiatry, 113, 104725. https://doi.org/10.1016/j.ajp.2025.104725