Hypnotics, sleep, and mortality in elderly people.

OBJECTIVE

To re-assess relationships between mortality, hypnotic use, subjective insomnia, and sleep duration.

DESIGN

A prospective study examining 5-year mortality among hypnotic drug users and respondents with subjective insomnia identified in a longitudinal study of health, activity, and lifestyle (Nottingham Longitudinal Study of Activity and Ageing).

SETTING

General community.

PARTICIPANTS

1042 survey respondents, aged over 65 years, randomly selected from the community and stratified at age 75 years.

MAIN OUTCOME MEASURE

Recorded mortality.

RESULTS

During the 5-year period, 352 respondents died. The mortality rate was significantly greater among those taking some form of medication for sleep (n = 208) than for those not taking sleep medication (n = 812; chi-square = 4.91, df = 1, P = 0.027). When sleep medication users were categorized as either "hypnotic users" (ie, users of medication with recognized hypnotic or sedative actions) or "other users" (including analgesics and other over-the-counter medicines), only "other users" showed significant excess mortality (chi-square = 7.27, df = 1, P = 0.007). Logistic regression showed that "other users" were 2.5 times more likely to die than "non-users" even when gender, health risk, and usual sleep duration were controlled. There were no significant relationships between mortality and subjective insomnia or reported duration of sleep.

CONCLUSIONS

Earlier reported relationships between excess mortality and use of medication for sleep are replicated in this study. Among elderly people, however, this relationship does not derive from the pharmacological characteristics of prescription hypnotics. Rather, it appears that reported self-medication to promote sleep, using a variety of non-sedative products, provides an epidemiological "marker" for a group within which levels of morbidity and mortality are particularly high. Excess mortality associated with very short or long sleep duration was not replicated in this study. Overall, these findings provide little epidemiological support for a wide-spread interaction between benzodiazepine hypnotic use and sleep disordered breathing in old age.