Psoriasis has been associated with lymphohematopoietic and solid cancers; however, reports have been inconsistent. Cancer incidence was compared between psoriasis and psoriasis-free patients, and the roles of psoriasis duration and treatment were explored in this observational study using the UK General Practice Research Database. Among 67,761 patients, 1,703 patients had incident cancer; of whom 54% had a history of psoriasis. Incidence rate ratios for lymphohematopoietic and pancreatic cancers were 1.81 (95% confidence interval (CI) 1.35-2.42) and 2.20 (95% CI 1.18-4.09), respectively. In a nested case-control analysis, adjusted odds ratios (ORs) for cancer overall were 1.50 (95% CI 1.30-1.74) for psoriasis of >or=4 years duration and 1.53 (95% CI 0.97-2.43) for patients receiving systemic treatment (marker of disease severity). Lymphohematopoietic malignancy risk was highest in patients with systemic treatment. The OR for patients without systemic treatment was 1.59 (95% CI 1.01-2.50) for psoriasis of <2 years and 2.12 (95% CI 1.45-3.10) for that of >or=2 years duration. Risks of bladder/kidney and colorectal cancers were increased with longer-duration psoriasis. Psoriasis patients may have an increased overall risk of incident cancer (mainly lymphohematopoietic and pancreatic). Longer-term psoriasis and more severe disease may increase the risk of some cancers. These observations need further confirmation, particularly because of the potential of findings by chance in observational studies with subgroup analyses.