Background Official suicide statistics for England are based on deaths given suicide verdicts and most cases given an open verdict following a coroner's inquest. Previous research indicates that some deaths given accidental verdicts are considered to be suicides by clinicians. Changes in coroners' use of different verdicts may bias suicide trend estimates. We investigated whether suicide trends may be over-or underestimated when they are based on deaths given suicide and open verdicts. Method Possible suicides assessed by 12 English coroners in 1990/91, 1998 and 2005 and assigned open, accident/misadventure or narrative verdicts were rated by three experienced suicide researchers according to the likelihood that they were suicides. Details of all suicide verdicts given by these coroners were also recorded. Results In 1990/91, 72.0% of researcher-defined suicides received a suicide verdict from the coroner, this decreased to 65.4% in 2005 (p trend <0.01); equivalent figures for combined suicide and open verdicts were 95.4% (1990/91) and 86.7% (2005). Researcher-defined suicides with a verdict of accident/misadventure doubled over that period, from 4.6% to 9.1% (p <0.01). Narrative verdict cases rose from zero in 1990/91 to 25 in 2005 (4.2% of researcher-defined suicides that year). In 1998 and 2005, 50.0% of the medicine poisoning deaths given accidental/misadventure verdicts were rated as suicide by the researchers. Conclusions Between 1990/91 and 2005, the proportion of researcher-defined suicides given a suicide verdict by coroners decreased, largely due to an increased use of accident/misadventure verdicts, particularly for deaths involving poisoning. Consideration should be given to the inclusion of 'accidental' deaths by poisoning with medicines in the statistics available for monitoring suicides rates. © 2012 Cambridge University Press.