In more details.
To investigate this hypothesis in a rigorous, prospective setting, Jari Tiihonen (Karolinska Institute, Stockholm) and colleagues undertook a population-based cohort study of all 21,492 people aged 17–65 years in Sweden diagnosed with schizophrenia before 2006.
In order to avoid survival bias, they also conducted an analysis in a separate cohort of 1230 patients with first-episode schizophrenia during the period 2006–2010. A total of 214,670 age- and gender-matched individuals from the general population served as controls.
The schizophrenia patients were assessed for cumulative exposure to antipsychotic drugs by defined daily dose (DDD) and categorised into four groups: no antipsychotics; small doses or occasional use (0–0.5 DDD/day); moderate doses (0.5–1.5 DDD); and high doses (>1.5 DDD/day).
In total, 1591 (7%) people in the main cohort died during 5 years of follow-up while 45 (4%) of those with first-episode schizophrenia died.
Mortality rates differed among the four antipsychotic exposure groups and displayed a U-shape curve, Tiihonen and co-workers report in Schizophrenia Bulletin. Compared with the general population, hazard ratios (HRs) for overall mortality were 6.3 for no antipsychotic use, 5.7 for high antipsychotic use, and 4.1 and 4.0 for low and moderate antipsychotic use, respectively.