BackgroundThe excess mortality following first-contact psychosis is well recognized. However, the causes of death in a complete incidence cohort and mortality patterns over time compared with the general population are unknown.MethodAll 2723 patients who presented for the first time with psychosis in three defined catchment areas of the UK in London (1965–2004, n=2056), Nottingham (1997–1999, n=203) and Dumfries and Galloway (1979–1998, n=464) were traced after a mean of 11.5 years follow-up and death certificates were obtained. Data analysis was by indirect standardization.ResultsThe overall standardized mortality ratio (SMR) for first-contact psychosis was 184 [95% confidence interval (CI) 167–202]. Most deaths (84.2%, 374/444) were from natural causes, although suicide had the highest SMR (1165, 95% CI 873–1524). Diseases of the respiratory system and infectious diseases had the highest SMR of the natural causes of death (232, 95% CI 183–291). The risk of death from diseases of the circulatory system was also elevated compared with the general population (SMR 139, 95% CI 117–164) whereas there was no such difference for neoplasms (SMR 111, 95% CI 86–141). There was strong evidence that the mortality gap compared with the general population for all causes of death (p<0.001) and all natural causes (p=0.01) increased over the four decades of the study. There was weak evidence that cardiovascular deaths may be increasing relative to the general population (p=0.07).ConclusionsPeople with first-contact psychosis have an overall mortality risk that is nearly double that of the general population. Most excess deaths are from natural causes. The widening of the mortality gap over the last four decades should be of concern to all clinicians involved in delivering healthcare.