The study reviewed electronic medical records for 10,531 patients who were prescribed hypnotics and 23,674 matched controls in an American healthcare database system between 2002 and 2007. Data were available for each individual for an average of 2.5 years.
After correction for several variables including age, gender, smoking, body mass index, ethnicity, marital status, alcohol use and prior cancer it was found that there was an increased risk of death among users of hypnotics compared to non-users, even at low levels of use. Among those prescribed hypnotics there were 638 deaths (6.06%) whereas among non-users there were 295 deaths (1.25%). For individuals prescribed up to 18 doses per year the hazard ratio was 3.60 (95% CI 2.92 to 4.44), for 18–132 doses per year this increased to 4.43 (95% CI 3.67 to 5.36) and for more than 132 doses per year it was 5.32 (95% CI 4.50 to 6.30).
The authors note that there are limitations to this study. The data relating to hypnotics covers the issuing of prescriptions. It is not known for sure that these prescriptions were dispensed and subsequently taken. Additionally, the data did not contain and could not be corrected for social and psychological problems as these are protected by privacy laws. Finally, despite the corrections made there may be residual confounding in the data.
The authors conclude that, "receiving hypnotic prescriptions was associated with greater than threefold increased hazards of death" and that "poor health did not explain the observed excess mortality".
National advice recommends only using hypnotics to treat insomnia when it is severe, disabling or causing the patient extreme distress.
Action: Clinicians should be aware of this study and continue to implement national recommendations. The absolute risk of death remains small and the association cannot be described as causal from this study.