- Email: [email protected]
Is usage of hypnotics associated with mortality?
Sleep Medicine 10 (2009) 279–286 www.elsevier.com/locate/sleep
Original Article
Is usage of hypnotics associated with mortality? Lena Mallon a,*, Jan-Erik Broman a, Jerker Hetta b b
a Department of Neuroscience, Psychiatry, Uppsala University, University Hospital, SE-75185 Uppsala, Sweden Department of Clinical Neuroscience, Psychiatry, Karolinska Institutet, Karolinska University Hospital, Huddinge, Sweden
Received 6 November 2008; received in revised form 19 December 2008; accepted 23 December 2008 Available online 9 March 2009
Abstract Objective: To investigate the influence of hypnotic usage on all-cause and cause-specific mortality in a middle-aged population. Methods: A cohort of 1750 men and 1773 women aged 30–65 years who responded to a postal questionnaire in 1983. The questionnaire included questions about hypnotic usage, sleep duration, sleep complaints, medical conditions, depression, demographic and life style variables. Mortality data for the period 1983–2003 were collected. Results: Regular hypnotic usage was reported by 1.7% of men and 2.2% of women, and was associated with short sleep, sleeping difficulties, several health problems and depression. During the 20-year follow-up period 379 men (21.5%) and 278 women (15.5%) died. After adjustment for potential risk factors in multivariate analyses regular hypnotic usage was associated with significantly increased risk of all-cause mortality in men (Hazard ratios [HR], 4.54; 95% confidence interval [CI], 2.47–8.37) and in women 2.03 (95% CI, 1.07–3.86). With regard to cause-specific mortality, regular hypnotic usage in men was a risk factor for coronary artery disease death, cancer death, suicide and death from ‘‘all remaining causes.” In women it was a risk factor for suicide. Conclusions: Our results show an increased risk of all-cause mortality and cause-specific mortality in regular users of hypnotics. Ó 2009 Elsevier B.V. All rights reserved. Keywords: Mortality; Hypnotics; General population; Longitudinal
1. Introduction Several population based studies have shown that insomnia complaints as well as short and long sleep duration are associated with increased mortality [1– 12]. A number of studies have, however, failed to detect such associations [13–15]. The inconsistent findings are partly due to the fact that insomnia is often a poorly defined term and its prevalence estimates vary widely. Previous surveys also used different definitions of short and long sleep in follow-up time and considered risk factors that may influence mortality. Similarly, a number of population based surveys have investigated the association between the use of hypnotics and mortality, and the results are somewhat inconclusive.
*
Corresponding author. Tel.: +46 225 494400; fax: +46 225 494413. E-mail address: [email protected] (L. Mallon).
1389-9457/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.sleep.2008.12.004
The first Cancer Prevention Study (CPS) found an increase in 6-year mortality in subjects reporting taking sleeping pills often [1], and the second CPS found an increase in subjects taking prescription sleeping pills in the last month [16]. A 12-year follow-up survey reported increased mortality risk in women taking sleeping pills [6], and in an 18-year follow-up of a middle-aged population daily users of hypnotic or anxiolytic drugs showed a higher mortality than non-users [17]. A 22-year study of twins aged 18 and above observed increased risk of mortality among frequent users of hypnotics/tranquilizers [12]. However, in a 6.3-year study of middle-aged subjects no association between hypnotic use and mortality was found [15], and in a 5-year follow-up of older subjects there was no increase in mortality among those using medication with sedative or hypnotic properties to promote sleep; instead there was increased mortality in subjects using other non-sedative products or over-thecounter preparations [13].
280
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
The association between hypnotic usage and mortality might be causal or confounded by underlying factors. Depression may explain the observed relationship since insomnia is a hallmark of depression, and insomnia is also a predictor for the development of depression [18–20]. In this study, we investigate a possible relationship between hypnotic use and mortality, controlling for poor sleep and depression within a population based cohort of middle-aged Swedes followed over a 20-year period. 2. Methods 2.1. Subjects The aim of the original cohort was to estimate the prevalence of certain physical complaints and depression in relation to sleeping difficulties and hypnotic medication in a middle-aged population. In 1983, 5102 subjects were randomly selected from the general population using the Central Population Registry (2439 men and women aged 30–44 years from the county of Gavleborg, and 2663 men and women aged 45–65 years from the county of Dalarna). The population was 291,500 in Gavleborg and 285,300 in Dalarna at the time of the study, and each group constituted 1/25 of the population in the corresponding age group. The mean age and standard deviation (SD) was 46 ± 10 years. A postal questionnaire was sent and 3550 subjects (69.6%) responded after one reminder (1758 males and 1792 females). The age and sex distribution of the responders were the same as the population at the time of the study. 2.2. Questionnaire The postal questionnaire used in 1983 included questions about sleep habits, sleeping difficulties, hypnotic usage, medical conditions, depression, demographic and life style variables. A detailed description of the questionnaire has been published previously [21]. The questions about hypnotic usage, sleeping difficulties and snoring were adopted from the Uppsala Sleep Inventory (USI) [22], which has previously been used in several epidemiological studies [9,23]. The specific question related to hypnotic usage was ‘‘How often do you use sleep medication?” The five alternatives in the answer were 1 = never, 2 = seldom, 3 = sometimes, 4 = often and 5 = very often. Subjects reporting hypnotic usage ‘‘often or very often” (scores 4–5) were considered regular hypnotic users. Difficulties initiating sleep (DIS) and difficulties maintaining sleep (DMS) were based on the responses ‘‘severe or very severe problems” versus ‘‘no, minor or moderate problems” to the questions, ‘‘How much of a problem do you have with falling asleep at night?” (DIS) and ‘‘How much of a
problem do you have with waking up during the night?” (DMS). Subjects were asked to estimate time to fall asleep (sleep latency, SL) and habitual total sleep time (TST), and the answers were expressed as continuous variables. Long sleep latency was defined as more than 45 min to fall asleep and short sleep was defined as sleeping less than 6 h per night. Subjects were also asked to estimate habitual sleep need, and SSI% (Sleep Sufficiency Index; the ratio of amount of estimated habitual sleep time to the amount of estimated need for sleep expressed in percent) was calculated [24]. Subjects were asked to state the frequency of ‘‘loud and disturbing snoring,” and those reporting symptoms at least often (scores 4–5) were considered habitual snorers. Living conditions were defined as living alone or with someone else, and smoking was defined as currently smoking or not. Heart disease, hypertension, diabetes and asthma were ascertained by asking subjects (in a yes/no question) if they had a history of heart disease, hypertension, diabetes or asthma. Depression was assessed by the question, ‘‘Do you feel depressed?” to be answered by yes or no in the Dalarna questionnaire. In the Gavleborg questionnaire depression was assessed by the question, ‘‘Do you have problems with feelings of depression?” and the alternatives in the answer were 1 = no, 2 = minor, 3 = moderate, 4 = severe and 5 = very severe problems. Responders answering at least severe problems (scores 4–5) were considered to suffer from depression. Questions were asked about body weight and height, and body mass index (BMI, in kg per m2) was calculated. 2.3. Mortality data Mortality data for the period 1983–2003 were collected. Six hundred fifty-seven (18.5%) of the cohort died during follow-up (379 men [21.6%] and 278 women [15.5%]). The mean age at death was 68 ± 11 years for men and 70 ± 10 years for women. Death certificates were obtained from the National Cause of Death Register in Sweden and death certificates were available for all 657 deceased. All-cause mortality and cause-specific mortality was the principal outcome for this analysis. Death certificates were coded using the 9th and 10th revisions of the International Classification of Disease (ICD) and categorised as ischemic heart disease (ICD9 codes 410–414 and ICD-10 codes I20–I25), cancer (ICD-9 codes 140–239 and ICD-10 codes C00–D48), suicide (ICD-9 codes E950–E959 and ICD-10 codes X60–X84), and all remaining causes. 2.4. Statistics Standard methods have been used to calculate means and standard deviations in the baseline survey. For comparison between categorical variables the v2-test was
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
used, and when the comparison involved continuous variables the t-test was used. The Kaplan–Meier survivorship function was used for survival analysis. The influence of possible explanatory variables on mortality hazards ratios (HR) were estimated by using the Cox proportional hazards regression model. Variables were included in the survival analysis if they were considered relevant confounders and sufficiently frequent in the baseline survey. The results are presented as HR with 95% confidence interval (CI). When analysing determinants of cause-specific mortality, subjects who died of other causes were treated as censored observations until 2003. The minimum statistical significance level for all analyses was p < 0.05. Separate analyses were developed for men and women in order to explore potential gender differences in the nature of associations, and estimates were adjusted for age in 5-year strata. The study was approved by the Ethics Committee of the Faculty of Medicine at Uppsala University in Sweden. 3. Results Among the 5102 randomly selected subjects in 1983, 3550 (69.6%) responded after one reminder and 3523 questionnaires (1750 men and 1773 women) were available for analyses after excluding 27 with incomplete data about hypnotic usage. Hypnotic usage never or seldom was reported by 93.2%, hypnotic usage sometimes by 4.8% and regular hypnotic usage (often or very often) by 2.0% of the population. Regular hypnotic usage was reported by 30 men (1.7%) and 39 women (2.2%) (v2 = 1.1, ns). Men using hypnotics regularly were older than nonregular users, 51 ± 11 years vs. 46 ± 10 years (t = 2.45, p = 0.014). They reported longer sleep latencies, 62 ± 50 min vs. 15 ± 24 min (t = 9.71, p < 0.001), shorter TST, 5.8 ± 1.7 h vs. 6.9 ± 1.0 h (t = 5.67, p < 0.001) and a lower SSI, 84 ± 25% vs. 95 ± 13% (t = 4.39, p < 0.0001) compared to men not using hypnotics regularly. Women using hypnotics regularly were older than non-regular users, 55 ± 8 years vs. 46 ± 10 years (t = 5.25, p < 0.001). They also reported longer sleep latencies, 61 ± 53 min vs. 19 ± 29 min (t = 7.74, p < 0.001), shorter TST, 6.3 ± 1.6 h vs. 7.0 ± 1.1 h (t = 4.35, p < 0.001) and a lower SSI, 84 ± 18% vs. 93 ± 14% (t = 4.05, p < 0.001) compared to women not using hypnotics regularly. Regular hypnotic users more often complained of DIS, DMS, long sleep latency (SL > 45 min) and short sleep (TST < 6 h) compared to non-regular users (Table 1). They reported depression more often and were more often living alone compared to non-regular users. Men using hypnotics regularly were more often smoking, overweight (BMI P 30) and had heart disease compared to men not using hypnotics regularly.
281
During the follow-up period 63.3% (19) of men using hypnotics regularly died versus 20.9% (359) of men not using hypnotics regularly. The corresponding figures for women were 46.2% (18) and 14.7% (255). As shown in Figs. 1 and 2 the cumulative survival rate was significantly lower in men (p < 0.001) and in women (p < 0.001) reporting regular hypnotic usage compared to men and women not using hypnotics regularly. Using hypnotics sometimes was not related to lower survival rate in men (p = 0.306) or in women (p = 0.398) compared to men and women not using hypnotics regularly. Table 2 shows that in age-adjusted Cox proportional hazard model regular hypnotic usage, DIS, smoking, heart disease, hypertension, diabetes, depression and living alone were related to all-cause mortality in men and in women. Having long sleep latency was related to allcause mortality in men, and having a BMI P 30 was associated with increased all-cause mortality in women in the age-adjusted model. After adjustment for all potential risk factors simultaneously in a multivariate Cox proportional hazard model, regular hypnotic usage remained significantly related to increased all-cause mortality in men (Table 3) and in women (Table 4). With regard to cause-specific mortality, regular hypnotic usage was a risk factor for suicide in both men and women in the multivariate adjusted model, but the number of cases was small warranting interpretative caution. Of those who died by suicide, 30% reported depression at baseline. Also, in the multivariate adjusted model regular hypnotic usage in men was a risk factor for coronary artery disease death, cancer death and death from ‘‘all remaining causes.” We found no relationship between regular hypnotic usage and coronary artery disease death, cancer death or death from ‘‘all remaining causes” in women. 4. Discussion The foregoing analyses suggest that regular use of hypnotics is associated with an increase in all-cause mortality. The risk associated with hypnotic usage was most pronounced in men. Epidemiological surveys have shown that women generally have a lower risk for morbidity and mortality related to insomnia [9,25], and the present study also suggests a lower mortality risk in women using hypnotics regularly compared to men using hypnotics regularly. The increased mortality risk was not confounded with sleeping difficulties, long sleep latency or short sleep duration. Important confounding factors such as age, depression, physical illness, being overweight, smoking, snoring and living conditions were included and the observed association between regular use of hypnotics and all-cause mortality remained statistically significant. We did not, however, find that using hypnotics sometimes was related to an increase in allcause mortality.
282
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
Table 1 Prevalence of variables in the baseline survey according to usage of hypnotics. Variables in the baseline survey
DIS DMS SL > 45 min TST < 6 h Smoking Habitual snoring BMI P 30 Heart disease Hypertension Diabetes Depression Asthma Living alone
Men
p-Value
Non-regular hypnotic usage (n = 1720)
Regular hypnotic usage (n = 30)
4.3 7.3 9.0 10.3 27.0 18.2 4.9 5.6 9.3 3.2 6.3 2.4 18.1
58.6 37.9 53.1 35.1 52.2 17.9 17.2 30.0 16.7 6.7 46.7 0 33.3
<.0001 <.0001 <.0001 <.0001 .007 ns .003 <.0001 ns ns <.0001 ns .032
Women
p-Value
Non-regular hypnotic usage (n = 1734)
Regular hypnotic usage (n = 39)
5.9 8.3 6.1 10.7 27.8 6.6 7.0 4.3 11.1 2.1 10.7 2.1 15.2
61.5 34.2 55.6 53.6 34.2 11.1 11.1 5.1 10.3 5.1 48.7 0 28.9
<.0001 <.0001 <.0001 <.0001 ns ns ns ns ns ns <.0001 ns .020
Values are in%. DIS, difficulties initiating sleep; DMS, difficulties maintaining sleep; TST, total sleep time; SL, sleep latency; BMI, body mass index; ns, non significant.
The prevalence of regular hypnotic usage in our survey, 1.7% of men and 2.2% of women, is similar or slightly lower to that reported from other surveys. The CPS I found a prevalence of ‘‘sleeping pill usage often” in 1.4% of men and 2.5% of women [1], and another study of North America reported a prevalence of 2.2% for ‘‘users of hypnotics” [15]. In a cohort of subjects aged 20–67 years the prevalence of ‘‘use of drugs for sleep sometimes or always” was 2.1% in men and 3.4% in women [6]. In a study of subjects aged 40–42 years ‘‘usage of hypnotic or anxiolytic drugs daily during the last month” was 1.9% in men and 4.5% in women [17]. In the Finnish Twin Cohort of subjects aged 18 and above 2.9% of men and 3.7% of women frequently used hypnotics and/or anxiolytics [12].
In the present study regular hypnotic usage was associated with sleeping difficulties and several health problems. It has previously been shown that hypnotic users often have somatic and psychiatric illnesses [26,27] and they are often treated with other psychoactive drugs [26,28]. Also, they are more often living alone and have social problems [26,27]. In one study it was revealed that unemployment was associated with increased risk of suicide and death from undetermined causes, and hypnotics was one of the factors that strengthened the association between unemployment and early mortality [29]. It is important to remember that associations do not prove causality and that medication rarely causes but is instead an indicator of disease. Subjects using hypnotics
Fig. 1. Survival in men using hypnotics sometimes or regularly compared to men not using hypnotics. (—) No hypnotic usage; ( – - – - – -) Hypnotic usage sometimes; (– – –) Regular hypnotic usage.
Fig. 2. Survival in women using hypnotics sometimes or regularly compared to women not using hypnotics. (—) No hypnotic usage; ( – - – - – -) Hypnotic usage sometimes; (– – –) Regular hypnotic usage.
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
are probably those with the most severe sleeping difficulties, and subjects with physical diseases or depression are more likely to use hypnotics than healthy persons. The increased mortality found among hypnotic users in this study may in fact be related to underlying diseases and their management rather than the hypnotic medication used. Nevertheless, several epidemiological studies have reported associations between use of hypnotics and mortality. In 1960, the CPS I, with over one million participants in the US, found that subjects taking sleeping pills often had increased mortality over 6 years [1]. In the CPS II of 1982 the wording ‘‘prescription sleeping pills” was used, and using sleeping pills 30 times in the last month was associated with increased mortality [16]. By the time of the first study the main prescription hypnotics were barbiturates, whereas benzodiazepines were preferred at the time of the second study. A survey in Japan [6] found that women, but not men, using sleeping pills had an increased mortality risk during a 12-year follow-up period, but a study of middle-aged men and women in the US reported that hypnotic usage was not associated with increase risk of death during a 3-year follow-up time [15]. A Norwegian study followed a population aged 40–42 years for 18 years and found that daily users of anxiolytic or hypnotic drugs had increased mortality [17]. In a Finnish study of twins aged 18 years and above, frequent use of hypnotics/tranquillizers was associated with increased risk of death in the 22-year follow-up time [12]. In that study, 42.9% of men who were frequent users of hypnotics/tranquillizers died compared to 15.5% of men not using hypnotics/ tranquillizers frequently. The corresponding figures for women were 35.2% and 10.8%. There was an association with frequent use of hypnotics and/or tranquilizers and mortality in all age groups in both genders in the Finn-
283
ish study. In a 5-year survey of elderly there was no increased risk of mortality found among those taking medication with recognized hypnotic or sedative action [13]. Instead, increased mortality was found in users of ‘‘other sleep medication,” including analgesics and over-the-counter medication. However, only few of these longitudinal cohort studies are randomised, general population studies [3,6,15,17], and they have, in varying degrees, controlled for different lifestyle habits and physical health problems. Only a small number of studies have controlled for the presence of depression [3,15]. In the present study the occurrence of depression was based on participants’ responses to a single question about feeling depressed. Brief instruments have been shown to perform as well as longer questionnaires for identifying depression in the general population, and the value of single-item depression screening is established [30,31]. The Dalarna cohort was reinvestigated with a questionnaire survey in 1995, and 75% of those reporting depressed mood in the survey had a mood disorder according to the DSM criteria when interviewed by a psychiatrist [32]. Several studies have indicated that insomnia and depression are intertwined over time [18–20,33], and regular hypnotic users reported sleeping difficulties, long sleep latency, short sleep duration and depression at baseline. Consequently, they are also at risk for becoming depressed during the observed time-period, and it has been shown that depression increases risk of death [34,35]. Thus, one possible pathway for the observed increased mortality may be mediated by depression. It has also been suggested that the incidence of depression is higher in subjects using hypnotics [36,37]. In the cause-specific mortality analysis we found that men and women using hypnotics regularly had an
Table 2 Hazard ratio (95% confidence interval) of age-adjusteda all-cause mortality by using the Cox proportional hazard model. Variables in the baseline survey
Men (n = 379) HR (95% confidence interval)
p-Value
HR (95% confidence interval)
p-Value
Regular hypnotic usage DIS DMS SL > 45 min TST < 6 h Smoking Habitual snoring BMI P 30 Heart disease Hypertension Diabetes Depression Asthma Living alone
3.51 1.84 1.20 1.44 1.01 1.98 1.04 1.42 1.89 1.53 2.75 1.92 1.13 1.72
<.0001 .002 .292 .035 .949 <.0001 .750 .061 <.0001 .001 <.0001 <.0001 .686 <.0001
2.29 1.52 1.27 1.27 0.94 1.92 1.39 1.70 1.52 1.45 5.16 1.41 1.55 1.60
.001 .023 .180 .183 .740 <.0001 .115 .003 .049 .009 <.0001 .023 .224 .001
a
(2.21–5.57) (1.25–2.71) (0.85–1.70) (1.03–2.03) (0.73–1.40) (1.53–2.56) (0.80–1.36) (0.98–2.05) (1.42–2.53) (1.19–1.98) (1.96–3.86) (1.42–2.60) (0.63–2.00) (1.35–2.18)
Women (n = 278)
(1.42–3.71) (1.06–2.17) (0.90–1.81) (0.89–1.80) (0.65–§.35) (1.48–2.53) (0.92–2.11) (1.20–2.39) (1.01–2.29) (1.10–1.93) (3.38–7.89) (1.05–1.91) (0.77–3.13) (1.21–2.10)
Adjusted for age in 5-year strata. All values are hazard ratio (HR) with 95% confidence interval (95% CI). Significant findings (p < 0.05) in boldface. DIS, difficulties initiating sleep; DMS, difficulties maintaining sleep; TST, total sleep time; SL, sleep latency; BMI, body mass index.
284
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
Table 3 Hazard ratio (95% confidence interval) of multivariate adjusteda all-cause and cause-specific mortality in men (n = 1750) by using Cox proportional hazard model, forward stepwise procedure. Cause of death (number deceased)
Variable
HR (95% CI)
p
All-cause mortality (n = 379)
Regular hypnotic use Smoking Diabetes Living alone Hypertension
4.54 1.84 2.97 1.44 1.45
(2.47–8.37) (1.39–2.44) (1.98–4.47) (1.06–1.97) (1.04–2.02)
<.0001 <.0001 <.0001 .019 .028
Diabetes Smoking Regular hypnotic use Heart disease
4.03 2.12 4.55 1.86
(2.15–7.53) (1.29–3.48) (1.65–12.58) (1.04–3.32)
<.0001 .003 .003 .037
Cancer death (n = 110)
Regular hypnotic use Smoking
3.99 (1.25–12.72) 1.70 (1.01–2.86)
.019 .048
Suicide (n = 10)
Regular hypnotic use
21.18 (2.58–173.74)
.004
All other causes (n = 134)
Smoking Diabetes Regular hypnotic use Living alone
2.12 5.14 3.64 1.72
.002 <.0001 .029 .028
Cause-specific mortality Death from CAD (n = 125)
(1.33–3.38) (2.87–9.22) (1.14–11.59) (1.06–2.79)
a Simultaneously adjusted for age in 5-year strata, regular hypnotic usage, DIS (difficulties initiating sleep), DMS (difficulties maintaining sleep), SL (sleep latency) >45 min, TST (total sleep time) <6 h, smoking, habitual snoring, BMI (body mass index) P30, heart disease, hypertension, diabetes, asthma, depression and living alone. All values are hazard ratio (HR) with 95% confidence interval (95% CI). CAD, coronary artery disease.
increase in death by suicide. Epidemiological data suggest that between 59% and 87% of suicide victims suffered from depression [38,39], and it is reasonable to assume that regular hypnotic users who were not depressed at baseline may have developed depression at a later time. Also, hypnotic and anxiolytic drugs may impair judgments and promote violent and risky
behaviours including suicide [40,41]. We also found an increased risk of cancer death in men using hypnotics regularly, and there is some evidence that hypnotics may increase cancer risk possibly through suppression of immune function or promoting viral infections [42]. Some limitations with the present study should be pointed out. First, our study did not include informa-
Table 4 Hazard ratio (95% confidence interval) of multivariate adjusteda all-cause and cause-specific mortality in women (n = 1773) by using Cox proportional hazard model, forward stepwise procedure. Cause of death (number deceased)
Variable
HR (95% CI)
p
All-cause mortality (n = 278)
Regular hypnotic use Smoking BMI P 30 Diabetes Asthma
2.03 2.28 2.22 4.94 2.29
.031 <.0001 <.0001 <.0001 .032
Smoking BMIP30 Diabetes
3.43 (1.67–7.03) 3.06 (1.31–7.12) 5.20 (1.55–17.44)
.001 .010 .008
Cancer death (n = 95)
Smoking Living alone
2.08 (1.23–3.53) 1.92 (1.10–3.34)
.006 .021
Suicide (n = 9)
Regular hypnotic use Heart disease
24.27 (4.36–135.20) 11.26 (2.03–62.48)
<.0001 .006
All other causes (n = 127)
Smoking Depression BMI P 30 Diabetes
1.93 1.76 2.65 7.98
.013 .036 .001 <.0001
Cause-specific mortality Death from CAD (n = 47)
(1.07–3.86) (1.66–3.15) (1.49–3.32) (2.84–8.61) (1.07–4.91)
(1.15–3.27) (1.03–2.96) (1.52–4.63) (3.92–16.27)
a Simultaneously adjusted for age in 5-year strata, regular hypnotic usage, DIS (difficulties initiating sleep), DMS (difficulties maintaining sleep), SL (sleep latency) >45 min, TST (total sleep time) <6 h, smoking, habitual snoring, BMI (body mass index) P30, heart disease, hypertension, diabetes, asthma, depression and living alone. All values are hazard ratio (HR) with 95% confidence interval (95% CI). CAD, coronary artery disease.
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
tion about the types of hypnotics that participants consumed or patterns of usage, particularly duration of use. Swedish official sale statistics show that the majority (about 70%) of hypnotics (group NO5C according to the Anatomical Therapeutic Chemicals classification system) sold in 1983 were benzodiazepines [43]. The second limitation is that certain physical and mental conditions, lifestyle factors such as excessive drinking, and other concomitant medications than hypnotics, which we did not deal with in this study, could influence the results. Therefore, we cannot rule out the possibility that hypnotic usage is a marker for other risk factors and illnesses and acts a confounder or causal intermediate in association with mortality. Also, it has previously been argued that usage of hypnotics might be related to many comorbidities, that epidemiological studies can never control for all possible confounders, and controlling for risk factors might lead to both over- and underestimation of mortality risks [16]. Lastly, the number of cases in some categories was small, warranting caution in interpreting the data. 5. Conclusion The available data today suggest that there is a complex association between use of hypnotics and mortality. The present study suggests that hypnotic usage is associated with sleeping difficulties, depression and physical illnesses, and regular usage increases the risk of mortality. Longitudinal epidemiological data on the consequences of insomnia and its treatment need to be gathered. Detailed information about what types of hypnotics and duration of usage is needed in future surveys. However, epidemiological evidence of association does not prove causality. Proper risk assessment of risk/benefit ratios with hypnotic usage can only be achieved with long term intervention studies and randomized controlled clinical trials. Such studies will allow us to assess the risks and benefits of hypnotic drug usage.
Acknowledgements This study was supported by grants from the Swedish Research Council and the County Council of Dalarna in Sweden.
References [1] Kripke DF, Simons RN, Garfinkel L, Hammond EC. Short and long sleep and sleeping pills. Is increased mortality associated? Arch Gen Psychiatry 1979;36:103–16. [2] Qureshi AI, Giles WH, Croft JB, Bliwise DL. Habitual sleep patterns and risk for stroke and coronary heart disease: a 10-year follow-up from NHANES I. Neurology 1997;48:904–11.
285
[3] Pollak CP, Perlick D, Linsner JP, Wenston J, Hsieh F. Sleep problems in the community elderly as predictors of death and nursing home placement. J Community Health 1990;15:123–35. [4] Kripke D, Garfinkel L, Wingard D, Klauber M, Marler M. Mortality associated with sleep duration and insomnia. Arch Gen Psychiatry 2002;59:131–6. [5] Patel SR, Ayas NT, Malhotra MR, White DP, Schernhammer ES, Speizer FE, et al. A prospective study of sleep duration and mortality risk in women. Sleep 2004;27:440–4. [6] Kojima M, Wakai K, Kawamura T, Tamakoshi A, Aoki R, Lin Y, et al. Sleep patterns and total mortality: a 12-year follow-up study in Japan. J Epidemiol 2000;10:87–93. [7] Heslop P, Smith GD, Metcalfe C, Macleod J, Hart C. Sleep duration and mortality: the effect of short or long sleep duration on cardiovascular and all-cause mortality in working men and women. Sleep Med 2002;3:305–14. [8] Amagai Y, Ishikawa S, Gotoh T, Doi Y, Kayaba K, Nakamura Y, et al. Sleep duration and mortality in Japan: the jichi medical school cohort study. J Epidemiol 2004;14:124–8. [9] Mallon L, Broman JE, Hetta J. Sleep complaints predict coronary artery disease mortality in males: a 12-year follow-up study of a middle-aged Swedish population. J Int Med 2002;251:207–16. [10] Tamakoshi A, Ohno YJACC Study Group. Self-reported sleep duration as a predictor of all-cause mortality: results from the JACC study, Japan. Sleep 2004;27:51–4. [11] Burazeri G, Goflin J, Kark JD. Over 8 h of sleep—marker of increased mortality in Mediterranean population: follow-up population study. Croat Med J 2003;44:193–8. [12] Hublin C, Partinen M, Koskenvuo M, Kaprio J. Sleep and mortality: a population-based 22-year follow-up study. Sleep 2007;30:1245–53. [13] Rumble R, Morgan K. Hypnotics, sleep, and mortality in elderly people. J Am Geriatr Soc 1992;40:787–91. [14] Althuis MD, Fredman L, Langenberg PW, Magaziner J. The relationship between insomnia and mortality among communitydwelling older women. J Am Geriatr Soc 1998;46:1270–3. [15] Phillips B, Mannino D. Does insomnia kill? Sleep 2005;28:965–71. [16] Kripke DF, Klauber MR, Wingard DL, Fell RL, Assmus JD, Garfinkel L. Mortality hazard associated with prescription hypnotics. Biol Psychiatry 1998;43:687–93. [17] Hausken AM, Skurtveit S, Tverdal A. Use of anxiolyic or hypnotic drugs and total mortality in a general middle-aged population. Pharmacoepidemiol Drug Saf 2007;16:913–8. [18] Ford DE, Kamerow DB. Epidemiologic study of sleep disturbances and psychiatric disorders. An opportunity for prevention? JAMA 1989;262:1479–84. [19] Mallon L, Broman JE, Hetta J. Relationship between insomnia, depression and mortality: a 12-year follow-up of older adults in the community. Int Psychogeriatr 2000;12:295–306. [20] Riemann D, Voderholzer U. Primary insomnia: a risk factor to develop depression? J Affect Disord 2003;76:255–9. ˚ gren H. The [21] Liljenberg B, Almqvist M, Hetta J, Roos BE, A prevalence of insomnia: the importance of operationally defined criteria. Ann Clin Res 1988;20:393–8. ˚ gren H, Hambert G, Liljenberg B, Roos [22] Hetta J, Almqvist M, A BE. Prevalence of sleep disturbances and related symptoms in a middle-aged Swedish population. In: Koella W, Ruther E, Schultz H, editors. Sleep´ 84. Stuttgart: Gustav Fischer Verlag; 1985. p. 373–6. ˚ gren H. Affective [23] Liljenberg B, Almqvist M, Hetta J, Roos BE, A disturbances and insomnia. Eur J Psychiatry 1989;3:91–8. [24] Broman JE, Lundh LG, Hetta J. Insufficient sleep in the general population. Neurophysiol Clin 1996;26:30–9. [25] Mallon L, Broman JE, Hetta J. High incidence of diabetes in men with sleep complaints or short sleep duration: a 12-year follow-up study of a middle-aged population. Diabetes Care 2005;28:2762–7.
286
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
[26] Allgulander C, Na¨sman P. Regular hypnotic drug treatment in a sample of 32, 679 Swedes: associations with somatic and mental health, inpatient psychiatric diagnosis and suicide, derived with automatic record-linkage. Psychosom Med 1991;53:101–8. [27] Blennow G, Romelsjo¨ A, Leifman H, Leifman A, Karlsson G. Seatives and hypnotics in Stockholm: social factors and kinds of use. Am J Public Health 1994;84:242–6. [28] Kageyama T, Kabuto M, Nitta H, Kurokawa Y, Taira K, Suzuki S, et al. Prevalence of medically prescribed hypnotics among adult Japanese women in urban residential areas. Psychiatry Clin Neurosci 1998;52:69–74. [29] Voss M, Nylen L, Floderus B, Diderichsen F, Terry PD. Unemployment and early cause-specific mortality: a study based on the Swedish twin registry. Am J Public Health 2004;94:2155–61. [30] Schade C, Jones E, Wittlin B. A 10-year review of the validity and clinical utility of depression screening. Psychiatr Serv 1998;49:55–61. [31] McKenzie N, Marks I. Quick rating of depressed mood in patients with anxiety disorders. Br J Psychiatry 1999;174:266–9. [32] Mallon L, Hetta J. One question approach to detect depression: a comparison between questionnaire and interview data in a community sample of older adults. Eur J Psychiatry 2002;16:135–44. [33] Jansson-Fro¨jmark M, Linton SJ. The course of insomnia over one year: a longitudinal study in the general population in Sweden. Sleep 2008;31:881–6.
[34] Wulsin LR, Evans JC, Vasan RS, Murabito JM, Kelly-Hayes M, Benjamin EJ. Depressive symptoms, coronary heart disease, and overall mortality in the Framingham heart study. Psychosom Med 2005;67:697–702. [35] Mykletun A, Bjerkeset O, Dewey M, Prince M, Overland S, Stewart R. Anxiety, depression, and cause-specific mortality: the HUNT study. Psychosom Med 2007;69:323–31. [36] Kripke DF. Greater incidence of depression with hypnotic use than placebo. BMC 2007;7:42. [37] Neckelmann D, Mykletun A, Dahl MD. Chronic insomnia as a risk factor for developing anxiety and depression. Sleep 2007;30:873–80. [38] Gonda X, Fountoulakis KN, Kaprinis G, Rihmer Z. Prediction and prevention of suicide in patients with unipolar depression and anxiety. Ann Gen Psychiatry 2007;6:23. [39] Rihmer Z. Suicide risk in mood disorders. Curr Opin Psychiatry 2007;20:17–22. [40] Neutel CI, Patten SB. Risk of suicide attempts after benzodiazepine and/or antidepressant use. Ann Epidemiol 1997;7:568–74. [41] Drummer OH. Benzodiazepines—effect on human performance and behaviour. Forensic Sci Rev 2002;14:2–14. [42] Kripke DF. Possibility that certain hypnotics might cause cancer in the skin. J Sleep Res 2008;17:245–50. [43] Drug sales statistics. Internal database at National Corporation of Swedish Pharmacies (Apotekets AB).
Original Article
Is usage of hypnotics associated with mortality? Lena Mallon a,*, Jan-Erik Broman a, Jerker Hetta b b
a Department of Neuroscience, Psychiatry, Uppsala University, University Hospital, SE-75185 Uppsala, Sweden Department of Clinical Neuroscience, Psychiatry, Karolinska Institutet, Karolinska University Hospital, Huddinge, Sweden
Received 6 November 2008; received in revised form 19 December 2008; accepted 23 December 2008 Available online 9 March 2009
Abstract Objective: To investigate the influence of hypnotic usage on all-cause and cause-specific mortality in a middle-aged population. Methods: A cohort of 1750 men and 1773 women aged 30–65 years who responded to a postal questionnaire in 1983. The questionnaire included questions about hypnotic usage, sleep duration, sleep complaints, medical conditions, depression, demographic and life style variables. Mortality data for the period 1983–2003 were collected. Results: Regular hypnotic usage was reported by 1.7% of men and 2.2% of women, and was associated with short sleep, sleeping difficulties, several health problems and depression. During the 20-year follow-up period 379 men (21.5%) and 278 women (15.5%) died. After adjustment for potential risk factors in multivariate analyses regular hypnotic usage was associated with significantly increased risk of all-cause mortality in men (Hazard ratios [HR], 4.54; 95% confidence interval [CI], 2.47–8.37) and in women 2.03 (95% CI, 1.07–3.86). With regard to cause-specific mortality, regular hypnotic usage in men was a risk factor for coronary artery disease death, cancer death, suicide and death from ‘‘all remaining causes.” In women it was a risk factor for suicide. Conclusions: Our results show an increased risk of all-cause mortality and cause-specific mortality in regular users of hypnotics. Ó 2009 Elsevier B.V. All rights reserved. Keywords: Mortality; Hypnotics; General population; Longitudinal
1. Introduction Several population based studies have shown that insomnia complaints as well as short and long sleep duration are associated with increased mortality [1– 12]. A number of studies have, however, failed to detect such associations [13–15]. The inconsistent findings are partly due to the fact that insomnia is often a poorly defined term and its prevalence estimates vary widely. Previous surveys also used different definitions of short and long sleep in follow-up time and considered risk factors that may influence mortality. Similarly, a number of population based surveys have investigated the association between the use of hypnotics and mortality, and the results are somewhat inconclusive.
*
Corresponding author. Tel.: +46 225 494400; fax: +46 225 494413. E-mail address: [email protected] (L. Mallon).
1389-9457/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.sleep.2008.12.004
The first Cancer Prevention Study (CPS) found an increase in 6-year mortality in subjects reporting taking sleeping pills often [1], and the second CPS found an increase in subjects taking prescription sleeping pills in the last month [16]. A 12-year follow-up survey reported increased mortality risk in women taking sleeping pills [6], and in an 18-year follow-up of a middle-aged population daily users of hypnotic or anxiolytic drugs showed a higher mortality than non-users [17]. A 22-year study of twins aged 18 and above observed increased risk of mortality among frequent users of hypnotics/tranquilizers [12]. However, in a 6.3-year study of middle-aged subjects no association between hypnotic use and mortality was found [15], and in a 5-year follow-up of older subjects there was no increase in mortality among those using medication with sedative or hypnotic properties to promote sleep; instead there was increased mortality in subjects using other non-sedative products or over-thecounter preparations [13].
280
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
The association between hypnotic usage and mortality might be causal or confounded by underlying factors. Depression may explain the observed relationship since insomnia is a hallmark of depression, and insomnia is also a predictor for the development of depression [18–20]. In this study, we investigate a possible relationship between hypnotic use and mortality, controlling for poor sleep and depression within a population based cohort of middle-aged Swedes followed over a 20-year period. 2. Methods 2.1. Subjects The aim of the original cohort was to estimate the prevalence of certain physical complaints and depression in relation to sleeping difficulties and hypnotic medication in a middle-aged population. In 1983, 5102 subjects were randomly selected from the general population using the Central Population Registry (2439 men and women aged 30–44 years from the county of Gavleborg, and 2663 men and women aged 45–65 years from the county of Dalarna). The population was 291,500 in Gavleborg and 285,300 in Dalarna at the time of the study, and each group constituted 1/25 of the population in the corresponding age group. The mean age and standard deviation (SD) was 46 ± 10 years. A postal questionnaire was sent and 3550 subjects (69.6%) responded after one reminder (1758 males and 1792 females). The age and sex distribution of the responders were the same as the population at the time of the study. 2.2. Questionnaire The postal questionnaire used in 1983 included questions about sleep habits, sleeping difficulties, hypnotic usage, medical conditions, depression, demographic and life style variables. A detailed description of the questionnaire has been published previously [21]. The questions about hypnotic usage, sleeping difficulties and snoring were adopted from the Uppsala Sleep Inventory (USI) [22], which has previously been used in several epidemiological studies [9,23]. The specific question related to hypnotic usage was ‘‘How often do you use sleep medication?” The five alternatives in the answer were 1 = never, 2 = seldom, 3 = sometimes, 4 = often and 5 = very often. Subjects reporting hypnotic usage ‘‘often or very often” (scores 4–5) were considered regular hypnotic users. Difficulties initiating sleep (DIS) and difficulties maintaining sleep (DMS) were based on the responses ‘‘severe or very severe problems” versus ‘‘no, minor or moderate problems” to the questions, ‘‘How much of a problem do you have with falling asleep at night?” (DIS) and ‘‘How much of a
problem do you have with waking up during the night?” (DMS). Subjects were asked to estimate time to fall asleep (sleep latency, SL) and habitual total sleep time (TST), and the answers were expressed as continuous variables. Long sleep latency was defined as more than 45 min to fall asleep and short sleep was defined as sleeping less than 6 h per night. Subjects were also asked to estimate habitual sleep need, and SSI% (Sleep Sufficiency Index; the ratio of amount of estimated habitual sleep time to the amount of estimated need for sleep expressed in percent) was calculated [24]. Subjects were asked to state the frequency of ‘‘loud and disturbing snoring,” and those reporting symptoms at least often (scores 4–5) were considered habitual snorers. Living conditions were defined as living alone or with someone else, and smoking was defined as currently smoking or not. Heart disease, hypertension, diabetes and asthma were ascertained by asking subjects (in a yes/no question) if they had a history of heart disease, hypertension, diabetes or asthma. Depression was assessed by the question, ‘‘Do you feel depressed?” to be answered by yes or no in the Dalarna questionnaire. In the Gavleborg questionnaire depression was assessed by the question, ‘‘Do you have problems with feelings of depression?” and the alternatives in the answer were 1 = no, 2 = minor, 3 = moderate, 4 = severe and 5 = very severe problems. Responders answering at least severe problems (scores 4–5) were considered to suffer from depression. Questions were asked about body weight and height, and body mass index (BMI, in kg per m2) was calculated. 2.3. Mortality data Mortality data for the period 1983–2003 were collected. Six hundred fifty-seven (18.5%) of the cohort died during follow-up (379 men [21.6%] and 278 women [15.5%]). The mean age at death was 68 ± 11 years for men and 70 ± 10 years for women. Death certificates were obtained from the National Cause of Death Register in Sweden and death certificates were available for all 657 deceased. All-cause mortality and cause-specific mortality was the principal outcome for this analysis. Death certificates were coded using the 9th and 10th revisions of the International Classification of Disease (ICD) and categorised as ischemic heart disease (ICD9 codes 410–414 and ICD-10 codes I20–I25), cancer (ICD-9 codes 140–239 and ICD-10 codes C00–D48), suicide (ICD-9 codes E950–E959 and ICD-10 codes X60–X84), and all remaining causes. 2.4. Statistics Standard methods have been used to calculate means and standard deviations in the baseline survey. For comparison between categorical variables the v2-test was
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
used, and when the comparison involved continuous variables the t-test was used. The Kaplan–Meier survivorship function was used for survival analysis. The influence of possible explanatory variables on mortality hazards ratios (HR) were estimated by using the Cox proportional hazards regression model. Variables were included in the survival analysis if they were considered relevant confounders and sufficiently frequent in the baseline survey. The results are presented as HR with 95% confidence interval (CI). When analysing determinants of cause-specific mortality, subjects who died of other causes were treated as censored observations until 2003. The minimum statistical significance level for all analyses was p < 0.05. Separate analyses were developed for men and women in order to explore potential gender differences in the nature of associations, and estimates were adjusted for age in 5-year strata. The study was approved by the Ethics Committee of the Faculty of Medicine at Uppsala University in Sweden. 3. Results Among the 5102 randomly selected subjects in 1983, 3550 (69.6%) responded after one reminder and 3523 questionnaires (1750 men and 1773 women) were available for analyses after excluding 27 with incomplete data about hypnotic usage. Hypnotic usage never or seldom was reported by 93.2%, hypnotic usage sometimes by 4.8% and regular hypnotic usage (often or very often) by 2.0% of the population. Regular hypnotic usage was reported by 30 men (1.7%) and 39 women (2.2%) (v2 = 1.1, ns). Men using hypnotics regularly were older than nonregular users, 51 ± 11 years vs. 46 ± 10 years (t = 2.45, p = 0.014). They reported longer sleep latencies, 62 ± 50 min vs. 15 ± 24 min (t = 9.71, p < 0.001), shorter TST, 5.8 ± 1.7 h vs. 6.9 ± 1.0 h (t = 5.67, p < 0.001) and a lower SSI, 84 ± 25% vs. 95 ± 13% (t = 4.39, p < 0.0001) compared to men not using hypnotics regularly. Women using hypnotics regularly were older than non-regular users, 55 ± 8 years vs. 46 ± 10 years (t = 5.25, p < 0.001). They also reported longer sleep latencies, 61 ± 53 min vs. 19 ± 29 min (t = 7.74, p < 0.001), shorter TST, 6.3 ± 1.6 h vs. 7.0 ± 1.1 h (t = 4.35, p < 0.001) and a lower SSI, 84 ± 18% vs. 93 ± 14% (t = 4.05, p < 0.001) compared to women not using hypnotics regularly. Regular hypnotic users more often complained of DIS, DMS, long sleep latency (SL > 45 min) and short sleep (TST < 6 h) compared to non-regular users (Table 1). They reported depression more often and were more often living alone compared to non-regular users. Men using hypnotics regularly were more often smoking, overweight (BMI P 30) and had heart disease compared to men not using hypnotics regularly.
281
During the follow-up period 63.3% (19) of men using hypnotics regularly died versus 20.9% (359) of men not using hypnotics regularly. The corresponding figures for women were 46.2% (18) and 14.7% (255). As shown in Figs. 1 and 2 the cumulative survival rate was significantly lower in men (p < 0.001) and in women (p < 0.001) reporting regular hypnotic usage compared to men and women not using hypnotics regularly. Using hypnotics sometimes was not related to lower survival rate in men (p = 0.306) or in women (p = 0.398) compared to men and women not using hypnotics regularly. Table 2 shows that in age-adjusted Cox proportional hazard model regular hypnotic usage, DIS, smoking, heart disease, hypertension, diabetes, depression and living alone were related to all-cause mortality in men and in women. Having long sleep latency was related to allcause mortality in men, and having a BMI P 30 was associated with increased all-cause mortality in women in the age-adjusted model. After adjustment for all potential risk factors simultaneously in a multivariate Cox proportional hazard model, regular hypnotic usage remained significantly related to increased all-cause mortality in men (Table 3) and in women (Table 4). With regard to cause-specific mortality, regular hypnotic usage was a risk factor for suicide in both men and women in the multivariate adjusted model, but the number of cases was small warranting interpretative caution. Of those who died by suicide, 30% reported depression at baseline. Also, in the multivariate adjusted model regular hypnotic usage in men was a risk factor for coronary artery disease death, cancer death and death from ‘‘all remaining causes.” We found no relationship between regular hypnotic usage and coronary artery disease death, cancer death or death from ‘‘all remaining causes” in women. 4. Discussion The foregoing analyses suggest that regular use of hypnotics is associated with an increase in all-cause mortality. The risk associated with hypnotic usage was most pronounced in men. Epidemiological surveys have shown that women generally have a lower risk for morbidity and mortality related to insomnia [9,25], and the present study also suggests a lower mortality risk in women using hypnotics regularly compared to men using hypnotics regularly. The increased mortality risk was not confounded with sleeping difficulties, long sleep latency or short sleep duration. Important confounding factors such as age, depression, physical illness, being overweight, smoking, snoring and living conditions were included and the observed association between regular use of hypnotics and all-cause mortality remained statistically significant. We did not, however, find that using hypnotics sometimes was related to an increase in allcause mortality.
282
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
Table 1 Prevalence of variables in the baseline survey according to usage of hypnotics. Variables in the baseline survey
DIS DMS SL > 45 min TST < 6 h Smoking Habitual snoring BMI P 30 Heart disease Hypertension Diabetes Depression Asthma Living alone
Men
p-Value
Non-regular hypnotic usage (n = 1720)
Regular hypnotic usage (n = 30)
4.3 7.3 9.0 10.3 27.0 18.2 4.9 5.6 9.3 3.2 6.3 2.4 18.1
58.6 37.9 53.1 35.1 52.2 17.9 17.2 30.0 16.7 6.7 46.7 0 33.3
<.0001 <.0001 <.0001 <.0001 .007 ns .003 <.0001 ns ns <.0001 ns .032
Women
p-Value
Non-regular hypnotic usage (n = 1734)
Regular hypnotic usage (n = 39)
5.9 8.3 6.1 10.7 27.8 6.6 7.0 4.3 11.1 2.1 10.7 2.1 15.2
61.5 34.2 55.6 53.6 34.2 11.1 11.1 5.1 10.3 5.1 48.7 0 28.9
<.0001 <.0001 <.0001 <.0001 ns ns ns ns ns ns <.0001 ns .020
Values are in%. DIS, difficulties initiating sleep; DMS, difficulties maintaining sleep; TST, total sleep time; SL, sleep latency; BMI, body mass index; ns, non significant.
The prevalence of regular hypnotic usage in our survey, 1.7% of men and 2.2% of women, is similar or slightly lower to that reported from other surveys. The CPS I found a prevalence of ‘‘sleeping pill usage often” in 1.4% of men and 2.5% of women [1], and another study of North America reported a prevalence of 2.2% for ‘‘users of hypnotics” [15]. In a cohort of subjects aged 20–67 years the prevalence of ‘‘use of drugs for sleep sometimes or always” was 2.1% in men and 3.4% in women [6]. In a study of subjects aged 40–42 years ‘‘usage of hypnotic or anxiolytic drugs daily during the last month” was 1.9% in men and 4.5% in women [17]. In the Finnish Twin Cohort of subjects aged 18 and above 2.9% of men and 3.7% of women frequently used hypnotics and/or anxiolytics [12].
In the present study regular hypnotic usage was associated with sleeping difficulties and several health problems. It has previously been shown that hypnotic users often have somatic and psychiatric illnesses [26,27] and they are often treated with other psychoactive drugs [26,28]. Also, they are more often living alone and have social problems [26,27]. In one study it was revealed that unemployment was associated with increased risk of suicide and death from undetermined causes, and hypnotics was one of the factors that strengthened the association between unemployment and early mortality [29]. It is important to remember that associations do not prove causality and that medication rarely causes but is instead an indicator of disease. Subjects using hypnotics
Fig. 1. Survival in men using hypnotics sometimes or regularly compared to men not using hypnotics. (—) No hypnotic usage; ( – - – - – -) Hypnotic usage sometimes; (– – –) Regular hypnotic usage.
Fig. 2. Survival in women using hypnotics sometimes or regularly compared to women not using hypnotics. (—) No hypnotic usage; ( – - – - – -) Hypnotic usage sometimes; (– – –) Regular hypnotic usage.
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
are probably those with the most severe sleeping difficulties, and subjects with physical diseases or depression are more likely to use hypnotics than healthy persons. The increased mortality found among hypnotic users in this study may in fact be related to underlying diseases and their management rather than the hypnotic medication used. Nevertheless, several epidemiological studies have reported associations between use of hypnotics and mortality. In 1960, the CPS I, with over one million participants in the US, found that subjects taking sleeping pills often had increased mortality over 6 years [1]. In the CPS II of 1982 the wording ‘‘prescription sleeping pills” was used, and using sleeping pills 30 times in the last month was associated with increased mortality [16]. By the time of the first study the main prescription hypnotics were barbiturates, whereas benzodiazepines were preferred at the time of the second study. A survey in Japan [6] found that women, but not men, using sleeping pills had an increased mortality risk during a 12-year follow-up period, but a study of middle-aged men and women in the US reported that hypnotic usage was not associated with increase risk of death during a 3-year follow-up time [15]. A Norwegian study followed a population aged 40–42 years for 18 years and found that daily users of anxiolytic or hypnotic drugs had increased mortality [17]. In a Finnish study of twins aged 18 years and above, frequent use of hypnotics/tranquillizers was associated with increased risk of death in the 22-year follow-up time [12]. In that study, 42.9% of men who were frequent users of hypnotics/tranquillizers died compared to 15.5% of men not using hypnotics/ tranquillizers frequently. The corresponding figures for women were 35.2% and 10.8%. There was an association with frequent use of hypnotics and/or tranquilizers and mortality in all age groups in both genders in the Finn-
283
ish study. In a 5-year survey of elderly there was no increased risk of mortality found among those taking medication with recognized hypnotic or sedative action [13]. Instead, increased mortality was found in users of ‘‘other sleep medication,” including analgesics and over-the-counter medication. However, only few of these longitudinal cohort studies are randomised, general population studies [3,6,15,17], and they have, in varying degrees, controlled for different lifestyle habits and physical health problems. Only a small number of studies have controlled for the presence of depression [3,15]. In the present study the occurrence of depression was based on participants’ responses to a single question about feeling depressed. Brief instruments have been shown to perform as well as longer questionnaires for identifying depression in the general population, and the value of single-item depression screening is established [30,31]. The Dalarna cohort was reinvestigated with a questionnaire survey in 1995, and 75% of those reporting depressed mood in the survey had a mood disorder according to the DSM criteria when interviewed by a psychiatrist [32]. Several studies have indicated that insomnia and depression are intertwined over time [18–20,33], and regular hypnotic users reported sleeping difficulties, long sleep latency, short sleep duration and depression at baseline. Consequently, they are also at risk for becoming depressed during the observed time-period, and it has been shown that depression increases risk of death [34,35]. Thus, one possible pathway for the observed increased mortality may be mediated by depression. It has also been suggested that the incidence of depression is higher in subjects using hypnotics [36,37]. In the cause-specific mortality analysis we found that men and women using hypnotics regularly had an
Table 2 Hazard ratio (95% confidence interval) of age-adjusteda all-cause mortality by using the Cox proportional hazard model. Variables in the baseline survey
Men (n = 379) HR (95% confidence interval)
p-Value
HR (95% confidence interval)
p-Value
Regular hypnotic usage DIS DMS SL > 45 min TST < 6 h Smoking Habitual snoring BMI P 30 Heart disease Hypertension Diabetes Depression Asthma Living alone
3.51 1.84 1.20 1.44 1.01 1.98 1.04 1.42 1.89 1.53 2.75 1.92 1.13 1.72
<.0001 .002 .292 .035 .949 <.0001 .750 .061 <.0001 .001 <.0001 <.0001 .686 <.0001
2.29 1.52 1.27 1.27 0.94 1.92 1.39 1.70 1.52 1.45 5.16 1.41 1.55 1.60
.001 .023 .180 .183 .740 <.0001 .115 .003 .049 .009 <.0001 .023 .224 .001
a
(2.21–5.57) (1.25–2.71) (0.85–1.70) (1.03–2.03) (0.73–1.40) (1.53–2.56) (0.80–1.36) (0.98–2.05) (1.42–2.53) (1.19–1.98) (1.96–3.86) (1.42–2.60) (0.63–2.00) (1.35–2.18)
Women (n = 278)
(1.42–3.71) (1.06–2.17) (0.90–1.81) (0.89–1.80) (0.65–§.35) (1.48–2.53) (0.92–2.11) (1.20–2.39) (1.01–2.29) (1.10–1.93) (3.38–7.89) (1.05–1.91) (0.77–3.13) (1.21–2.10)
Adjusted for age in 5-year strata. All values are hazard ratio (HR) with 95% confidence interval (95% CI). Significant findings (p < 0.05) in boldface. DIS, difficulties initiating sleep; DMS, difficulties maintaining sleep; TST, total sleep time; SL, sleep latency; BMI, body mass index.
284
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
Table 3 Hazard ratio (95% confidence interval) of multivariate adjusteda all-cause and cause-specific mortality in men (n = 1750) by using Cox proportional hazard model, forward stepwise procedure. Cause of death (number deceased)
Variable
HR (95% CI)
p
All-cause mortality (n = 379)
Regular hypnotic use Smoking Diabetes Living alone Hypertension
4.54 1.84 2.97 1.44 1.45
(2.47–8.37) (1.39–2.44) (1.98–4.47) (1.06–1.97) (1.04–2.02)
<.0001 <.0001 <.0001 .019 .028
Diabetes Smoking Regular hypnotic use Heart disease
4.03 2.12 4.55 1.86
(2.15–7.53) (1.29–3.48) (1.65–12.58) (1.04–3.32)
<.0001 .003 .003 .037
Cancer death (n = 110)
Regular hypnotic use Smoking
3.99 (1.25–12.72) 1.70 (1.01–2.86)
.019 .048
Suicide (n = 10)
Regular hypnotic use
21.18 (2.58–173.74)
.004
All other causes (n = 134)
Smoking Diabetes Regular hypnotic use Living alone
2.12 5.14 3.64 1.72
.002 <.0001 .029 .028
Cause-specific mortality Death from CAD (n = 125)
(1.33–3.38) (2.87–9.22) (1.14–11.59) (1.06–2.79)
a Simultaneously adjusted for age in 5-year strata, regular hypnotic usage, DIS (difficulties initiating sleep), DMS (difficulties maintaining sleep), SL (sleep latency) >45 min, TST (total sleep time) <6 h, smoking, habitual snoring, BMI (body mass index) P30, heart disease, hypertension, diabetes, asthma, depression and living alone. All values are hazard ratio (HR) with 95% confidence interval (95% CI). CAD, coronary artery disease.
increase in death by suicide. Epidemiological data suggest that between 59% and 87% of suicide victims suffered from depression [38,39], and it is reasonable to assume that regular hypnotic users who were not depressed at baseline may have developed depression at a later time. Also, hypnotic and anxiolytic drugs may impair judgments and promote violent and risky
behaviours including suicide [40,41]. We also found an increased risk of cancer death in men using hypnotics regularly, and there is some evidence that hypnotics may increase cancer risk possibly through suppression of immune function or promoting viral infections [42]. Some limitations with the present study should be pointed out. First, our study did not include informa-
Table 4 Hazard ratio (95% confidence interval) of multivariate adjusteda all-cause and cause-specific mortality in women (n = 1773) by using Cox proportional hazard model, forward stepwise procedure. Cause of death (number deceased)
Variable
HR (95% CI)
p
All-cause mortality (n = 278)
Regular hypnotic use Smoking BMI P 30 Diabetes Asthma
2.03 2.28 2.22 4.94 2.29
.031 <.0001 <.0001 <.0001 .032
Smoking BMIP30 Diabetes
3.43 (1.67–7.03) 3.06 (1.31–7.12) 5.20 (1.55–17.44)
.001 .010 .008
Cancer death (n = 95)
Smoking Living alone
2.08 (1.23–3.53) 1.92 (1.10–3.34)
.006 .021
Suicide (n = 9)
Regular hypnotic use Heart disease
24.27 (4.36–135.20) 11.26 (2.03–62.48)
<.0001 .006
All other causes (n = 127)
Smoking Depression BMI P 30 Diabetes
1.93 1.76 2.65 7.98
.013 .036 .001 <.0001
Cause-specific mortality Death from CAD (n = 47)
(1.07–3.86) (1.66–3.15) (1.49–3.32) (2.84–8.61) (1.07–4.91)
(1.15–3.27) (1.03–2.96) (1.52–4.63) (3.92–16.27)
a Simultaneously adjusted for age in 5-year strata, regular hypnotic usage, DIS (difficulties initiating sleep), DMS (difficulties maintaining sleep), SL (sleep latency) >45 min, TST (total sleep time) <6 h, smoking, habitual snoring, BMI (body mass index) P30, heart disease, hypertension, diabetes, asthma, depression and living alone. All values are hazard ratio (HR) with 95% confidence interval (95% CI). CAD, coronary artery disease.
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
tion about the types of hypnotics that participants consumed or patterns of usage, particularly duration of use. Swedish official sale statistics show that the majority (about 70%) of hypnotics (group NO5C according to the Anatomical Therapeutic Chemicals classification system) sold in 1983 were benzodiazepines [43]. The second limitation is that certain physical and mental conditions, lifestyle factors such as excessive drinking, and other concomitant medications than hypnotics, which we did not deal with in this study, could influence the results. Therefore, we cannot rule out the possibility that hypnotic usage is a marker for other risk factors and illnesses and acts a confounder or causal intermediate in association with mortality. Also, it has previously been argued that usage of hypnotics might be related to many comorbidities, that epidemiological studies can never control for all possible confounders, and controlling for risk factors might lead to both over- and underestimation of mortality risks [16]. Lastly, the number of cases in some categories was small, warranting caution in interpreting the data. 5. Conclusion The available data today suggest that there is a complex association between use of hypnotics and mortality. The present study suggests that hypnotic usage is associated with sleeping difficulties, depression and physical illnesses, and regular usage increases the risk of mortality. Longitudinal epidemiological data on the consequences of insomnia and its treatment need to be gathered. Detailed information about what types of hypnotics and duration of usage is needed in future surveys. However, epidemiological evidence of association does not prove causality. Proper risk assessment of risk/benefit ratios with hypnotic usage can only be achieved with long term intervention studies and randomized controlled clinical trials. Such studies will allow us to assess the risks and benefits of hypnotic drug usage.
Acknowledgements This study was supported by grants from the Swedish Research Council and the County Council of Dalarna in Sweden.
References [1] Kripke DF, Simons RN, Garfinkel L, Hammond EC. Short and long sleep and sleeping pills. Is increased mortality associated? Arch Gen Psychiatry 1979;36:103–16. [2] Qureshi AI, Giles WH, Croft JB, Bliwise DL. Habitual sleep patterns and risk for stroke and coronary heart disease: a 10-year follow-up from NHANES I. Neurology 1997;48:904–11.
285
[3] Pollak CP, Perlick D, Linsner JP, Wenston J, Hsieh F. Sleep problems in the community elderly as predictors of death and nursing home placement. J Community Health 1990;15:123–35. [4] Kripke D, Garfinkel L, Wingard D, Klauber M, Marler M. Mortality associated with sleep duration and insomnia. Arch Gen Psychiatry 2002;59:131–6. [5] Patel SR, Ayas NT, Malhotra MR, White DP, Schernhammer ES, Speizer FE, et al. A prospective study of sleep duration and mortality risk in women. Sleep 2004;27:440–4. [6] Kojima M, Wakai K, Kawamura T, Tamakoshi A, Aoki R, Lin Y, et al. Sleep patterns and total mortality: a 12-year follow-up study in Japan. J Epidemiol 2000;10:87–93. [7] Heslop P, Smith GD, Metcalfe C, Macleod J, Hart C. Sleep duration and mortality: the effect of short or long sleep duration on cardiovascular and all-cause mortality in working men and women. Sleep Med 2002;3:305–14. [8] Amagai Y, Ishikawa S, Gotoh T, Doi Y, Kayaba K, Nakamura Y, et al. Sleep duration and mortality in Japan: the jichi medical school cohort study. J Epidemiol 2004;14:124–8. [9] Mallon L, Broman JE, Hetta J. Sleep complaints predict coronary artery disease mortality in males: a 12-year follow-up study of a middle-aged Swedish population. J Int Med 2002;251:207–16. [10] Tamakoshi A, Ohno YJACC Study Group. Self-reported sleep duration as a predictor of all-cause mortality: results from the JACC study, Japan. Sleep 2004;27:51–4. [11] Burazeri G, Goflin J, Kark JD. Over 8 h of sleep—marker of increased mortality in Mediterranean population: follow-up population study. Croat Med J 2003;44:193–8. [12] Hublin C, Partinen M, Koskenvuo M, Kaprio J. Sleep and mortality: a population-based 22-year follow-up study. Sleep 2007;30:1245–53. [13] Rumble R, Morgan K. Hypnotics, sleep, and mortality in elderly people. J Am Geriatr Soc 1992;40:787–91. [14] Althuis MD, Fredman L, Langenberg PW, Magaziner J. The relationship between insomnia and mortality among communitydwelling older women. J Am Geriatr Soc 1998;46:1270–3. [15] Phillips B, Mannino D. Does insomnia kill? Sleep 2005;28:965–71. [16] Kripke DF, Klauber MR, Wingard DL, Fell RL, Assmus JD, Garfinkel L. Mortality hazard associated with prescription hypnotics. Biol Psychiatry 1998;43:687–93. [17] Hausken AM, Skurtveit S, Tverdal A. Use of anxiolyic or hypnotic drugs and total mortality in a general middle-aged population. Pharmacoepidemiol Drug Saf 2007;16:913–8. [18] Ford DE, Kamerow DB. Epidemiologic study of sleep disturbances and psychiatric disorders. An opportunity for prevention? JAMA 1989;262:1479–84. [19] Mallon L, Broman JE, Hetta J. Relationship between insomnia, depression and mortality: a 12-year follow-up of older adults in the community. Int Psychogeriatr 2000;12:295–306. [20] Riemann D, Voderholzer U. Primary insomnia: a risk factor to develop depression? J Affect Disord 2003;76:255–9. ˚ gren H. The [21] Liljenberg B, Almqvist M, Hetta J, Roos BE, A prevalence of insomnia: the importance of operationally defined criteria. Ann Clin Res 1988;20:393–8. ˚ gren H, Hambert G, Liljenberg B, Roos [22] Hetta J, Almqvist M, A BE. Prevalence of sleep disturbances and related symptoms in a middle-aged Swedish population. In: Koella W, Ruther E, Schultz H, editors. Sleep´ 84. Stuttgart: Gustav Fischer Verlag; 1985. p. 373–6. ˚ gren H. Affective [23] Liljenberg B, Almqvist M, Hetta J, Roos BE, A disturbances and insomnia. Eur J Psychiatry 1989;3:91–8. [24] Broman JE, Lundh LG, Hetta J. Insufficient sleep in the general population. Neurophysiol Clin 1996;26:30–9. [25] Mallon L, Broman JE, Hetta J. High incidence of diabetes in men with sleep complaints or short sleep duration: a 12-year follow-up study of a middle-aged population. Diabetes Care 2005;28:2762–7.
286
L. Mallon et al. / Sleep Medicine 10 (2009) 279–286
[26] Allgulander C, Na¨sman P. Regular hypnotic drug treatment in a sample of 32, 679 Swedes: associations with somatic and mental health, inpatient psychiatric diagnosis and suicide, derived with automatic record-linkage. Psychosom Med 1991;53:101–8. [27] Blennow G, Romelsjo¨ A, Leifman H, Leifman A, Karlsson G. Seatives and hypnotics in Stockholm: social factors and kinds of use. Am J Public Health 1994;84:242–6. [28] Kageyama T, Kabuto M, Nitta H, Kurokawa Y, Taira K, Suzuki S, et al. Prevalence of medically prescribed hypnotics among adult Japanese women in urban residential areas. Psychiatry Clin Neurosci 1998;52:69–74. [29] Voss M, Nylen L, Floderus B, Diderichsen F, Terry PD. Unemployment and early cause-specific mortality: a study based on the Swedish twin registry. Am J Public Health 2004;94:2155–61. [30] Schade C, Jones E, Wittlin B. A 10-year review of the validity and clinical utility of depression screening. Psychiatr Serv 1998;49:55–61. [31] McKenzie N, Marks I. Quick rating of depressed mood in patients with anxiety disorders. Br J Psychiatry 1999;174:266–9. [32] Mallon L, Hetta J. One question approach to detect depression: a comparison between questionnaire and interview data in a community sample of older adults. Eur J Psychiatry 2002;16:135–44. [33] Jansson-Fro¨jmark M, Linton SJ. The course of insomnia over one year: a longitudinal study in the general population in Sweden. Sleep 2008;31:881–6.
[34] Wulsin LR, Evans JC, Vasan RS, Murabito JM, Kelly-Hayes M, Benjamin EJ. Depressive symptoms, coronary heart disease, and overall mortality in the Framingham heart study. Psychosom Med 2005;67:697–702. [35] Mykletun A, Bjerkeset O, Dewey M, Prince M, Overland S, Stewart R. Anxiety, depression, and cause-specific mortality: the HUNT study. Psychosom Med 2007;69:323–31. [36] Kripke DF. Greater incidence of depression with hypnotic use than placebo. BMC 2007;7:42. [37] Neckelmann D, Mykletun A, Dahl MD. Chronic insomnia as a risk factor for developing anxiety and depression. Sleep 2007;30:873–80. [38] Gonda X, Fountoulakis KN, Kaprinis G, Rihmer Z. Prediction and prevention of suicide in patients with unipolar depression and anxiety. Ann Gen Psychiatry 2007;6:23. [39] Rihmer Z. Suicide risk in mood disorders. Curr Opin Psychiatry 2007;20:17–22. [40] Neutel CI, Patten SB. Risk of suicide attempts after benzodiazepine and/or antidepressant use. Ann Epidemiol 1997;7:568–74. [41] Drummer OH. Benzodiazepines—effect on human performance and behaviour. Forensic Sci Rev 2002;14:2–14. [42] Kripke DF. Possibility that certain hypnotics might cause cancer in the skin. J Sleep Res 2008;17:245–50. [43] Drug sales statistics. Internal database at National Corporation of Swedish Pharmacies (Apotekets AB).