Cannabis consumption is a modifiable risk factor associated with psychosis, but not all cannabis users develop psychosis. Animal studies suggest that an antecedent active immune system interacts with subsequent cannabis exposure and moderates the cannabis–psychosis association, supporting the two-hit hypothesis. The clinical investigations are few, and it is unclear if the immune system is a biological candidate moderating the cannabis–psychosis association or whether cannabis increases inflammation, which in turn, augments psychosis likelihood.

We explored the mediating and moderating role of blood inflammation using PROCESS macro. We used data from a cross-sectional study, including 153 first-episode psychosis patients and 256 community-based controls. Participants answered the Cannabis Experience Questionnaire (cannabis frequency, age of onset, and duration), and plasma cytokines were measured [interleukin (IL)-1β, IL-6, IL-4, IL-10, tumour necrosis factor-α (TNF-α), interferon-γ (IFN-γ), transforming growth factor-β (TGF-β); multiplex]. We computed an inflammatory composite score (ICS) to represent the systemic inflammatory state. Confounders included sex, age, ethnicity, educational level, body mass index, tobacco smoking, lifetime use of other drugs, and antipsychotic treatment.

Mediation: Cannabis consumption was not associated with increased inflammation, thus not supporting a mediating effect of inflammation. Moderation: Daily use and age of onset <17 interacted significantly with the ICS to increase the odds of psychosis beyond their individual effects and were only associated with psychosis among those scoring medium–high in the ICS.

Immune dysregulation might be part of the pathophysiology of psychosis, not explained by cannabis use or other confounders. We provide the first and initial evidence that immune dysregulation modifies the cannabis–psychosis association, in line with a two-hit hypothesis.

Inflammation is a possible biological mechanism underlying the association between childhood maltreatment and psychosis. Previous investigations on this regard were mainly conducted on chronic schizophrenia and lacked control for confounders. We aim to investigate the role of familial liability, childhood maltreatment and recent stress in determining cytokine abnormalities at the onset of psychosis.

We recruited 114 first-episode psychosis (FEP) patients, 57 unaffected biological siblings of FEP patients, and 251 community-based controls. Plasma cytokines (IL-1β, IL-6, TNF-α, IFN-γ, IL-4, IL-10 and TGF-β) were measured and differences across the groups analysed after adjusting for potential confounders.

FEP had a higher pro- and anti-inflammatory cytokine profile (IL-1β, IL-6, TNF-α, IL-10 and TGF-β), which was not observed in unaffected siblings. Siblings presented decreased IL-1β when compared with patients and controls. Childhood maltreatment was associated with higher levels of TGF-β in both patients and siblings when compared with controls. Physical childhood abuse was associated with increased levels of TGF-β in FEP patients but with decreased levels in controls. Other childhood maltreatment subtypes or recent stressors did not affect cytokine levels in any of the groups.

Normal or reduced cytokines in siblings represent possibly a protective factor and suggest that the identified inflammatory profile in FEP can be a real pathophysiological component of psychosis. Experience of childhood maltreatment may contribute as long-term immune priming for the TGF-β pathway, and increased levels of this cytokine in both patients and siblings exposed to childhood maltreatment point to a possible biological candidate of familial risk for psychosis.