Completed suicides have been shown to be associated with smoking cigarettes in cohort studies of mainly female registered nurses in the U.S.A. (Hemenway et al., Reference Hemenway, Solnick and Colditz1993), both sexes in the Finnish general population (Tanskanen et al., Reference Tanskanen, Tuomihehto, Viinamaki, Vartiainen, Lehtonen and Puska2000), male army recruits in the U.S.A. (Miller et al., Reference Miller, Hemenway, Bell, Yore and Amoroso2000a), males in the U.S. general population (Davey Smith et al., Reference Davey Smith, Phillips and Neaton1992), army recruits in Sweden (Hemmingsson and Kriebel, Reference Hemmingsson and Kriebel2003), males aged 40–69 years in the general population in Japan (Iwasaki et al., Reference Iwasaki, Akechi, Uchitomi and Tsugane2005), and the elderly in a retirement community in the U.S.A. (Ross et al., Reference Ross, Bernstein, Trent, Henderson and Paganini-Hill1990). Also, a Swiss case-control study of army recruits reported similar findings (Angst and Clayton, Reference Angst and Clayton1998). Additionally, positive correlation between prevalence rates of smoking and both general population (Shah and Bhandarkar, Reference Shah and Bhandarkar2008) and elderly (Shah, Reference Shah2008) male suicide rates has been observed in large cross-national ecological studies. This relationship between completed suicides and smoking was “dose-dependent” (Hemenway et al., Reference Hemenway, Solnick and Colditz1993; Tanskanen et al., Reference Tanskanen, Tuomihehto, Viinamaki, Vartiainen, Lehtonen and Puska2000; Miller et al., Reference Miller, Hemenway, Bell, Yore and Amoroso2000a; Davey Smith et al., Reference Davey Smith, Phillips and Neaton1992; Hemmingsson and Kriebel, Reference Hemmingsson and Kriebel2003; Iwasaki et al., Reference Iwasaki, Akechi, Uchitomi and Tsugane2005). Moreover, this relationship was maintained in some studies after controlling for confounding variables including demographic characteristics, socioeconomic characteristics, levels of alcohol consumption, psychiatric symptoms and physical health (Tanskanen et al., Reference Tanskanen, Tuomihehto, Viinamaki, Vartiainen, Lehtonen and Puska2000; Miller et al., Reference Miller, Hemenway, Bell, Yore and Amoroso2000a), but disappeared in other studies (Hemmingsson and Kriebel, Reference Hemmingsson and Kriebel2003; Shah, Reference Shah2008; Shah and Bhandarkar, Reference Shah and Bhandarkar2008).

The relationship between smoking and completed suicides has mainly been examined in cohort or case-control studies at an individual-level in younger subjects, and there is a paucity of studies examining this relationship in the elderly. Therefore, the relationship between elderly suicide rates and the prevalence of smoking in England and Wales was examined.

Data on elderly suicide rates for males and females in the age-bands 65–74 years and 75+ years for England and Wales were ascertained from the World Health Organization website (http://www.who.int/whosis/database/mort/table1.cfm) for the 23-year period 1979 to 2001.

Data on the prevalence of smoking in England for the general population and those over the age of 60 years in males and females were ascertained from the Office of National Statistics for every alternate year between 1978 and 2000 and for the year 2001. These data on the prevalence of smoking were derived from serial General Household Surveys in England between 1978 and 2002.

The relationship between suicide rates in both elderly age-bands in both sexes and the prevalence of smoking in the general population and in those over the age of 60 years in the general population for males, females and both sexes combined was examined using Spearman's correlation coefficient (ρ).

Data on elderly suicide rates and the prevalence of smoking were available for 10 separate years (nine of the alternate years between 1978 and 2000 and for the year 2001). Table 1 illustrates that there were significant positive correlations between suicide rates in both elderly age-bands in both sexes and the prevalence of smoking in the general population and in those over the age of 60 years in the general population for males, females and both sexes combined. There may be several possible explanations for these findings including methodological issues. Data on suicide rates were for England and Wales, whereas data on the prevalence of smoking were for only England. However, it is unlikely that the suicide rates for England and Wales were significantly influenced by the suicide rates for Wales as the population size of Wales is significantly smaller than that of England. The positive correlations may have been spurious due to type 2 statistical errors because there were only a small number of data points (N = 10). Other confounding variables may independently influence the trends over time in both elderly suicide rates and the prevalence of smoking, leading to a spurious correlation between these two variables (epiphenomena). However, as noted above, this relationship was maintained in some previous studies after controlling for confounding variables including demographic characteristics, socioeconomic characteristics, levels of alcohol consumption, psychiatric symptoms and physical health (Tanskanen et al., Reference Tanskanen, Tuomihehto, Viinamaki, Vartiainen, Lehtonen and Puska2000; Miller et al., Reference Miller, Hemenway, Bell, Yore and Amoroso2000a). Such confounding variables were not examined in this study.

Table 1. Correlations between elderly suicide rates and prevalence of smoking

The positive correlations between suicide rates in both elderly age-bands in both sexes and the prevalence of smoking in the general population and in those over the age of 60 years in the general population for males, females and both sexes combined were consistent with the previously observed relationship between smoking and completed suicides in cohort and case-control studies at an individual-level in younger subjects, a cohort study of the elderly in a retirement community in the U.S.A. (Ross et al., Reference Ross, Bernstein, Trent, Henderson and Paganini-Hill1990), and general population (Shah and Bhandarkar, Reference Shah and Bhandarkar2008) and elderly (Shah, Reference Shah2008) male suicide rates in large cross-national ecological studies. Thus, the findings of the current study may be genuine.

There has been much debate about the possible causal link between smoking and completed suicides (Davey Smith et al., Reference Davey Smith, Phillips and Neaton1992; Sheikh, Reference Sheikh2000; Miller et al., Reference Miller, Hemenway, Bell, Yore and Amoroso2000b). A number of potential explanations have been proposed. First, there may be a causal link between smoking and depressive illness (Hemenway et al., Reference Hemenway, Solnick and Colditz1993; Tanskanen et al., Reference Tanskanen, Tuomihehto, Viinamaki, Vartiainen, Lehtonen and Puska2000). This may be mediated through the central effects of nicotine (Tanskanen et al., Reference Tanskanen, Tuomihehto, Viinamaki, Vartiainen, Lehtonen and Puska2000). Japanese suicide victims with a history of smoking had higher levels of nicotine and cotinine in their blood and urine (Moriya and Hashimoto, Reference Moriya and Hashimoto2005; Moriya et al., Reference Moriya, Furumiya and Hashimoto2006); the authors speculated that there may be a marked increase in smoking among habitual smokers before committing suicide. Second, depressed smokers may find it particularly difficult to quit smoking (Hemenway et al., Reference Hemenway, Solnick and Colditz1993; Sheikh, Reference Sheikh2000). Third, other factors, including low self-esteem (Hemenway et al., Reference Hemenway, Solnick and Colditz1993) and alcohol and substance misuse (Phillips, Reference Phillips1992), may predispose to both smoking and suicide. Fourth, depressed patients may have a higher prevalence of smoking (Sheikh, Reference Sheikh2000). Fifth, smoking and suicide may be linked to other disorders like cancer (Hemenway et al., Reference Hemenway, Solnick and Colditz1993; Sheikh, Reference Sheikh2000); occult carcinomas may clinically present with depressive illness (Shah and De, Reference Shah and De1998), and occult carcinomas have been discovered at post mortem in elderly suicide victims (Catell, Reference Cattell1988). Finally, smoking may directly enhance suicidal behavior. There is evidence of serotonergic hypofunction in psychiatric patients with lethal suicidal behavior (Mann and Malone, Reference Mann and Malone1997), and an inverse relationship between smoking and indices of serotonin function has been observed in psychiatric patients with depressive illness (Malone et al., Reference Malone, Waternaux, Haas, Cooper, Li and Mann2003).

Caution should be exercised in attributing a causal relationship and the direction of causality from this cross-sectional ecological study due to ecological fallacy. Nevertheless, the observed relationship between elderly suicide rates and smoking suggests that there is a need to confirm this relationship in the elderly in individual-level case-control or cohort studies. If such a relationship is confirmed then there will be a need to examine the explanatory mechanisms for this relationship.

Conflict of interest

None.