PurposeThe aim of the study was to evaluate the effect of the consumption of anxiolytic or hypnotic drugs on total mortality in a general population.MethodsWe followed a cohort of 7225 men and 7726 women aged 40–42 years who underwent health surveys in 1985–1989 in two Norwegian counties, with respect to deaths. Mean follow‐up period was 18 years. The subjects were categorised according to frequency of anxiolytic or hypnotic drug use during the last month: daily, every week, less than every week and not used during the last month.ResultsThe proportion of anxiolytic or hypnotic drug users was 6.6% among men and 16.2% among women. Altogether 402 men and 290 women died. There was an increase in risk of death with an increase in frequency of use. Crude hazard ratios for men and women daily using anxiolytics or hypnotics were 3.1 (95%CI 2.0, 4.8) and 2.7 (1.9, 4.0), respectively, as compared with non‐users last month. After adjusting painkiller use and smoking the hazard ratios were lowered, being 2.4 (1.5–4.0) (men) and 2.1 (1.4–3.2) (women). After additional adjustments for other possible confounders the hazard ratios were further attenuated to 1.5 (0.9–2.7) for men and 1.7 (1.1–2.6) for women.ConclusionsDaily users of anxiolytic or hypnotic drugs in our study showed higher crude mortality than non‐users. However, after adjusting lifestyle and socio‐economic variables the difference was markedly reduced suggesting that the remaining excess mortality is due to residual confounding. Copyright © 2007 John Wiley & Sons, Ltd.