Background: Mental disorders may represent a risk factor for severe COVID-19. This study evaluated the risk of COVID-19 mortality, hospitalization, and intensive care unit (ICU) admission in patients with pre-existing mental disorders or exposure to psychopharmacological treatments. Methods: In this PRISMA-compliant systematic review and meta-analysis (PROSPERO-CRD42021233984), we searched Web of Science, PsycINFO, Cochrane, PubMed, and Ovid/PsycINFO databases through March 5, 2021 for studies reporting data on COVID-19 outcomes in psychiatric patients compared to controls. Risk of COVID-19 mortality was defined as the primary outcome, with hospitalization and ICU admission as secondary outcomes. We modeled random-effects meta-analyses to estimate crude (OR) and adjusted (aOR) odds ratios for these three outcomes. Heterogeneity was assessed using the I2 statistic, publication bias was tested with Egger regression and visual inspection of funnel plots. Newcastle Ottawa Scale was used to explore study quality. We estimated effect sizes associated with any pre-existing mental disorder as well as specific categories of mental disorders (psychotic disorders, mood disorders, substance use disorders, anxiety disorders, intellectual disabilities and developmental disorders), and exposure to psychopharmacological drug classes (antidepressants, antipsychotics, and anxiolytics). Severe mental illness was compared with other mental disorders. We also conducted subgroup analyses and meta-regressions to assess the effects of baseline COVID-19 treatment setting, patient age, country, pandemic phase, quality assessment score and adjustment for confounders. Findings: A total of 32 studies were included in the systematic review and 23 in the meta-analysis, representing 658,702 patients with COVID-19 (41,944 with mental disorders). The presence of any mental disorder was associated with increased risk of COVID-19 mortality (Crude OR 1·99, 95% CI 1·56-2·54; aOR 1·34, 95% CI 1·15-1·56). This association was confirmed in psychotic disorders (Crude OR 2·11, 95% CI 1·47-3·03; aOR 1·77, 95% CI 1·34-2·33), and mood disorders (Crude OR 1·98, 95% CI 1·48-2·64; aOR 1·44, 95% CI 1·14-1·82). An increased mortality risk with substance use disorders (Crude OR 1·76, 95% CI 1·30- 2·38; aOR 1·12, 95% CI 0·89-1·39) and intellectual disabilities and developmental disorders (Crude OR 1·47, 95% CI 1·08-2·01) was only established in unadjusted estimates. We detected no effect for anxiety disorders (Crude OR 1·06, 95% CI 0·75-1·51; aOR 1·01, 95% CI 0·76-1·34). COVID-19 mortality was also associated with exposure to antipsychotics (Crude OR 3·71, 95% CI 1·74-7·91; aOR 2·43, 95% CI 1·81-3·25) and anxiolytics (Crude OR 2·58, 95% CI 1·22-5·44; aOR 1·47, 95% CI 1·15-1·88), but with antidepressant exposure only in unadjusted models (Crude OR 2·23, 95% CI 1·06-4·71; aOR 1·18, 95% CI 0·93-1·50). Overall, mental disorders were consistently associated with increased risk of hospitalization (Crude OR: 2·24, 95% CI 1·70-2·94; aOR: 1·77, 95% CI 1·29-2·42), but not ICU admission (Crude OR 1·71, 95% CI 1-2·93; aOR 1·26, 95% CI 0·78-2·05). Increased rates of hospital and ICU admission were found in substance use disorders (Hospitalization - Crude OR: 2·66, 95% CI 1·79-3·95; aOR: 1·87, 95% CI 1·16-3·03; ICU - Crude OR 2·59, 95% CI 1·54-4·38) and for unadjusted estimates in mood disorders (Hospitalization Crude OR 2·26, 95% CI 1·33-3·86; aOR 1·26, 95% CI 0·64-2·50; ICU - Crude OR 2·39, 95% CI 1·15-5·00; aOR 1·13, 95% CI 0·70-1·84). Psychotic disorders were not associated with an increased risk of hospitalization (Crude OR 1·68, 95% CI 0·86-3·29; aOR: 1·34, 95% CI 0·61-2·94), and even showed a decreased risk of ICU admission in adjusted estimates (Crude OR 0·77, 95% CI 0·42- 1·42; aOR 0·75, 95% CI 0·62-0·91). Adjusted effect sizes were significantly higher for mortality (p=0·005) and lower for ICU admission (pInterpretation: Pre-existing mental disorders, in particular psychotic and mood disorders, as well as exposure to antipsychotics and anxiolytics, were associated with COVID-19 mortality. Substance use disorders were associated with COVID-19 hospitalization. These findings support the need for targeted approaches for the management and prevention of COVID-19 in patients with mental disorders. Funding: NoneDeclaration of Interest: LDP reports grants from Boehringer-Ingelheim and Janssen R&D, outside the submitted work. PFP reports grants from Lundbeck, personal fees from Angelini, personal fees from Menarini, non-financial support from Boehringer Ingelheim, outside the submitted work. BV, FB, AB, RT, ML, MEB, IB, PFP and LDP are members of the ECNP Immunopsychiatry Thematic Working Group. All other authors have nothing to disclose.