BackgroundThe last comprehensive international systematic review of the incidence of psychotic disorders was published in 2004. New epidemiological data, from different settings, now permit a broader understanding of global variation. We synthesized this literature to examine variation by demographic characteristics and study methodology.MethodsWe followed PRISMA guidelines and registered our systematic review on PROSPERO (ID: CRD42018086800). We searched PubMed, Embase, Web of Science, PsychInfo and bibliographies, and directly contacted authors, to obtain citations of original research published between 2002 and 2017 on incidence of non-organic adult-onset psychotic disorder. Quality of yield was assessed. Data were evaluated using univariable random-effects meta-analysis, and meta-regression. ResultsFrom 56,721 records, 177 met inclusion criteria. The pooled incidence of all psychotic disorders was 26.6 (95% confidence interval [CI]:22.0-31.7) per 100,000 person-years. Pooled incidence of non-affective disorders was 18.7 (95%CI:14.8-23.7) per 100,000 person-years, and 4.6 (95%CI:3.1-6.8) for affective psychotic disorders. Heterogeneity was high (I2≥99.0%). Men were at higher risk of all psychotic disorders (incidence rate ratio [IRR]: 1.44, 95%CI:1.27-1.62) and non-affective disorders (IRR:1.60, 95%CI:1.44-1.87), but not of affective psychotic disorders (IRR:0.87, 95%CI:0.75-1.00). Ethnic minorities were also at excess risk of all psychotic disorders (IRR:1.75, 95%CI:1.53-2.00), including non-affective disorders (IRR:1.71, 95%CI:1.40-2.09). Meta-regression revealed that population registers reported higher rates of non-affective disorders (IRR:9.64, 95%CI:2.72-31.82), schizophrenia (IRR:2.51, 95%CI:1.24-5.21) and bipolar disorder (IRR:4.53, 95%CI:2.41-8.51) than first contact study designs. DiscussionThere is marked variation in incidence of psychotic disorders by personal characteristics and place. Some geographical variation could be partially explained by differences in case ascertainment methods.