Pyoderma gangrenosum (PG) is an important disease with significant complications. The objectives of this study were to determine incidence and mortality of PG and strength of reported associations. A retrospective cohort study was completed using computerized medical records from the General Practice Research Database, a large representative UK database. Patients with PG and three groups of age-, sex-, and practice-matched controls--general population, rheumatoid arthritis (RA), and inflammatory bowel disease (IBD) controls--were included in the study. Incidence and mortality were determined and validation undertaken to inform diagnostic accuracy. In all there were 313 people with the median age of 59 (interquartile range 41-72) years, and of them 185 (59%) were female. The adjusted incidence rate standardized to European standard population was 0.63 (95% confidence interval (CI) 0.57-0.71) per 100,000 person-years. The risk of death was three times higher than that for general controls (adjusted hazard ratio=3.03, 95% CI 1.84-4.73, P<0.001), 72% higher than that for IBD controls (adjusted hazard ratio=1.72, 95% CI 1.17-2.59, P=0.013), with a borderline increase compared with RA controls (adjusted hazard ratio=1.55, 95% CI 1.01-2.37, P=0.045). Disease associations were present in 110 (33%) participants: IBD, n=67 (20.2%); RA, n=39 (11.8%); and hematological disorders, n=13 (3.9%). To our knowledge, there are no previous population-based studies of the epidemiology of PG, an important disease with significantly increased mortality.