Increased levels of smoking

People with mental health problems are more likely to smoke and to smoke more heavily than the general population. However, smoking is associated not only with the prevalence but also with first-ever incidence of mental disorders (Reference Cuijpers, Smit and ten HaveCuijpers et al, 2007). In a prospective, population-based epidemiological study, people who smoked but had no history of mental disorder had an increased risk of developing such a disorder; this risk remained significant after correcting for major risk indicators of mental disorders (incidence rate ratio RR = 1.56, P<0.05). Smoking tobacco is significantly associated with increased prevalence of all major psychiatric disorders (Reference Farrell, Howes and BebbingtonFarrell et al, 2001) and the probability of any mental health problem in smokers is double that in non-smokers (22.4 v. 12.2%). In a large population survey of psychiatric morbidity in the UK, 64% of those with probable psychosis were smokers compared with 29% without psychosis (Reference Coultard, Farrell and SingletonCoultard et al, 2000). Smokers with mental health problems smoke more heavily, are more nicotine-dependent and have smoked for longer than smokers in the general population (Reference Kumari and PostmaKumari & Postma, 2005). The highest levels of smoking are among patients in psychiatric units, where up to 70% smoke, 50% heavily (Reference Jochelson and MajrowskiJochelson & Majrowski, 2006); rates are lower for people with mental illnesses living at home. Indeed, psychiatric hospitalisation can also lead non-smokers to become smokers (Reference Lawn, Pols and BarberLawn et al, 2002). Heavy smoking is associated with increased severity of psychotic illness, poorer outcomes and more frequent hospital admissions (Reference Aguilar, Gurpegui and DiazAguilar et al, 2005).

Although smokers with schizophrenia or depression may recognise the risks of smoking, there is evidence that they find cigarettes more appealing than do the general population (Reference Spring, Pingitore and McChargueSpring et al, 2003). In 2005, tobacco companies in the USA spent more than $13 billion on advertising and promotion (Federal Trade Commission, 2007) and people with mental health problems may be more vulnerable to such influence. A recent study analysed previously secret tobacco industry documents which indicated that the tobacco industry monitored or directly funded research supporting the idea that individuals with schizophrenia were less susceptible to the harms of tobacco and that they needed tobacco as self-medication (Reference Prochaska, Hall and BeroProchaska et al, 2008). The authors reported that the tobacco industry promoted smoking in psychiatric settings by providing cigarettes, supporting efforts to block hospital smoking bans and engaging in a variety of activities that slowed development of treatment for nicotine dependence treatment for this population. Box 1 summarises other explanations for the higher levels of smoking in people with mental health problems.

Physical health and mortality rates

Mortality rates for people with schizophrenia show a life expectancy roughly 20% shorter than that of the general population (Reference Hennekens, Hennekens and HollarHennekens et al, 2005). More than two-thirds of people with schizophrenia, compared with about a half in the general population, die of coronary heart disease. The main factors for this excess mortality are cigarette smoking, obesity, diabetes and hypertension. Mental illness also predicts higher morbidity from respiratory disease, with a diagnosis of schizophrenia increasing risk of death from respiratory disease to 10 times that for the general population (Reference Joukamaa, Heliövaara and KnektJoukamaa et al, 2001). Smoking cessation before middle age can avoid more than 90% of the lung cancer mortality risk attributable to tobacco (Reference Peto, Darby and DeoPeto et al, 2000). It is likely that high rates of smoking result in this increased risk of many smoking-related diseases and that this is responsible for most of the excess natural mortality of people with mental health problems (Reference Brown, Barraclough and InskipBrown et al, 2000).

The risk of many diseases, in particular respiratory and cardiovascular disease, in people with mental health problems could therefore be preventable by appropriate support to quit smoking. This would significantly improve physical health, life expectancy and well-being. Moreover, results of a population survey showed that people whose mental illness had remitted were not at increased risk for subsequent smoking, in contrast with individuals with active disorders (Reference Breslau, Novak and KesslerBreslau et al, 2004). There are therefore additional and previously unrecognised public health benefits of early treatment of mental disorders.

Smoking and suicide as causes of mortality

Since one in two regular smokers dies prematurely (Reference Doll, Peto and BorehamDoll et al, 2004), at least 50% of people with serious mental illness who persist in smoking regularly will die prematurely from smoking-related diseases. This compares with a 10% lifetime risk of suicide, a cause of death that has received a great deal of attention from government and the psychiatric profession. In addition, there is a link between smoking and suicide, as studies have indicated that smokers are more likely to think about suicide (Reference Wilhelm, Arnold and NivenWilhelm et al, 2004) and have higher suicide rates (Reference Malone, Waternaux and HaasMalone et al, 2003).

Smoking as a psychiatric disorder

The relationship between mental illness and smoking is complex, particularly if smoking itself is considered as a psychiatric disorder. Nicotine dependence fulfils DSM–IV core criteria for mental disorder, being a ‘clinically significant behavioural or psychological syndrome or pattern that occurs in an individual and associated with distress, disability and a significantly increased risk of suffering death and disability’ (American Psychiatric Association, 1995). Both nicotine dependence (305.10) and nicotine withdrawal are included as disorders in DSM–IV. Chapter V of ICD–10 includes mental and behavioural disorders due to use of tobacco and defines both dependence (F17.2) and withdrawal (F17.3) (World Health Organization, 1992). Nicotine dependence is therefore the most prevalent mental disorder and typically persists for decades once established, with very low rates of permanent remission: less than 2–3% a year in the UK (Reference Taylor, Lader and BryantTaylor et al, 2006).

Effects of smoking on neurotransmitters

Nicotine crosses the blood–brain barrier several seconds after inhalation and binds to nicotinic cholinergic receptors. In the central nervous system this causes the release of neurotransmitters such as noradrenaline, serotonin, dopamine, acetylcholine, γ-aminobutyric acid and glutamate. Chronic smoking causes up-regulation of nicotinic receptors (Reference Gotti, Fornasari and ClementiGotti et al, 1997), in contrast to the down-regulation of most neurotransmitter receptors on exposure to agonists.

Smoking inhibits the enzyme monoamine oxidase B (MAO–B), which is responsible for the catabolism of several brain neurotransmitters, including dopamine, serotonin and noradrenaline (Reference Haustein, Haffner and WoodcockHaustein et al, 2002). Smokers therefore have lower brain MAO–B levels than non-smokers, although levels return to normal on quitting (Reference Khalil, Steyn and CastagnoliKhalil et al, 2000). Inhibition of MAO–B causes dopamine levels to increase, thereby making smoking potentially rewarding. Long-term smokers have been found to have neurochemical abnormalities in the locus cerulus similar to those seen in animals treated with antidepressant drugs and opposite to those observed in people with major depression (Reference Klimek, Zhu and DilleyKlimek et al, 2001). Since MAO inhibitors are effective antidepressants, these findings suggest that smoking might have some antidepressant effects (Reference HughesHughes, 1999). The association between cigarette smoking and the presence and severity of suicidal behaviour across major psychiatric disorders has been postulated to be related to lower brain serotonin function in smokers with depression (Reference Malone, Waternaux and HaasMalone et al, 2003).

Nicotine may be particularly reinforcing in schizophrenia because it stimulates the subcortical reward system and the prefrontal cortex, which both appear to be hypofunctional in schizophrenia (Reference Chambers, Krystal and SelfChambers et al, 2001). Reference Watkins, Koob and MarkouWatkins et al (2000) suggested that the ability of nicotine to trigger dopamine release in mesolimbic reward centres might make smoking one of the few reinforcers able to overcome the anhedonia that characterises both schizophrenia and depression. In addition, noradrenergic effects of nicotine may account for increased attentiveness and improved performance of repetitive tasks (Reference Sullivan and CoveySullivan & Covey, 2002).

Smoking and depression

The national survey of psychiatric morbidity (Reference Coultard, Farrell and SingletonCoultard et al, 2000) found that people with depressive or anxiety disorders are twice as likely to smoke as those without. A clear relationship was noted between smoking prevalence and number of depressive/anxiety symptoms as measured by the revised version of the Clinical Interview Schedule (CIS–R). General population surveys in both Australia and the USA have also found that smokers have higher rates of depression than non-smokers (Reference Wilhelm, Mitchell and SladeWilhelm et al, 2003). As previously noted, smokers are more likely to think about suicide and have higher suicide rates.

Evidence suggests that smoking increases risk of depression (Reference Klungsoyr, Nygard and SorensenKlungsoyr et al, 2006). A general population study of more than 3000 people revealed that ex-smokers reported less depression than current smokers (Reference Haukkala, Uutela and VartiainenHaukkala et al, 2000). Studies also suggest that smoking increases risk of depression in adolescents. In a prospective study of 2000 young people (8–14 years old) without a history of depression, initiation of smoking was not associated with current depressed mood, although smoking modestly increased the risk of developing depressed mood, suggesting that smoking led to depression (Reference Wu and AnthonyWu & Anthony, 1999). This finding was borne out in a larger prospective study of adolescents in which smoking was shown to predispose to the development of depression (Reference Goodman and CapitmanGoodman & Capitman, 2000). Another US study of over 13 000 individuals, which used data from the National Household Survey on Drug Abuse, revealed that smoking had a dose relationship with depression in adolescents: the greatest number of depressive symptoms were reported in current smokers and the lowest number in those who had never smoked; in former smokers, depressive symptoms correlated inversely with time since quitting (Reference Martini, Wagner and AnthonyMartini et al, 2002).

There is evidence of shared genetic factors predisposing individuals to both smoking and depression: higher rates of smoking and major depression have been found in co-twins of monozygotic probands than in co-twins of dizygotic probands (Reference Kendler, Neale and MacleanKendler et al, 1993). It has also been suggested that cigarette smoking and depression are linked by common parameters of deprivation, rather than cigarette smoking causing depression or the converse (Reference Roy, Parker and MitchellRoy et al, 2001). Furthermore, previous studies of associations between depression and smoking might have overestimated the association by ignoring comorbid anxiety and other cofounders. In a large population-based health survey, the association with smoking was strongest for comorbid anxiety and depression, less strong for anxiety alone and only marginal for depression alone (Reference Mykletun, Overland and AarøMykletun et al, 2008). Potential confounders suggested as contributors to the association were somatic symptoms, socio-demographics, alcohol problems and low physical activity. Anxiety and depression were most common in current smokers, followed by ex-smokers and then never-smokers, although no decline in anxiety or depression was found with time since cessation.

Smoking and anxiety

A US study of 688 youths reported that cigarette smoking increases the risk of certain anxiety disorders during late adolescence and early adulthood (Reference Johnson, Cohen and PineJohnson et al, 2000). Daily smoking has also been associated with an increased risk of the first occurrence of panic attack, with evidence of reduced risk of a panic disorder after smoking cessation in two epidemiological studies of more than 5000 people (Reference Breslau and KleinBreslau & Klein, 1999). Perceived stress levels were found to reduce on stopping smoking and increase again on return to smoking (Reference Cohen and LichtensteinCohen & Lichtenstein, 1990). In another study, anxiety decreased following the first week of abstinence (Reference West and HajekWest & Hajek, 1997), which suggests that smoking is chronically anxiogenic rather than anxiolytic. This is consistent with the finding that anxiety is increased among smokers in general population studies (Reference Coultard, Farrell and SingletonCoultard et al, 2000).

Aside from smoking's apparent anxiogenic action, depressive and anxiety symptoms may increase the risk that an individual will take up smoking because of increased susceptibility to peer influences (results from a prospective study of more than 2000 students; Reference Patton, Carlin and CoffeyPatton et al, 1998). Similarly, social anxiety disorder seems to predict uptake of smoking (more than 3000 adolescents; Reference Sonntag, Wittchen and HoflerSonntag et al, 2000).

A systematic review revealed high levels of smoking (respectively 40–86% and 34–61%) in both clinical and non-clinical populations with post-traumatic stress disorder (Reference Fu, McFall and SaxonFu et al, 2007).

Summary

There appears to be a clear relationship between smoking and depressive and anxiety symptoms, and symptoms reduce after smoking cessation. Smokers are also more likely to think about suicide and have higher suicide rates.

Smoking and schizophrenia

Smoking and other substance misuse

Smokers are more likely to have problems with drugs and alcohol, and the link with alcohol misuse and dependence is particularly strong (Reference Miller and GoldMiller & Gold, 1998). Alcohol use disorders have been significantly associated with regular heavy smoking (Reference Dierker, Avenevoli and StolarDierker et al, 2002). Smoking at an early age is also associated with substance misuse (Reference Williams and ZiedonisWilliams & Ziedonis, 2004) and current daily cigarette smokers are more likely to have a comorbid substance use disorder than people who have never smoked (Reference John, Meyer and RumpfJohn et al, 2004). Use of cannabis has also been found to reduce rates of smoking cessation (Reference Ford, Vua and AnthonyFord et al, 2002).

It has been estimated that about half of people with schizophrenia misuse illicit substances and alcohol at some time during their illness (Reference Reiger, Meyers and KramerReiger et al, 1984) and those that do have poorer symptomatic and functional outcomes than those that do not (Reference Kavanagh, McGrath and SaundersKavanagh et al, 2002). In a 15-month follow-up study examining the course of substance misuse and daily tobacco use in 103 individuals treated for first-episode psychosis (Reference Wade, Harrigan and EdwardsWade et al, 2006), a significant reduction occurred in the rate of any substance misuse but not of daily tobacco use.

Depression and anxiety

Nicotine withdrawal symptoms related to decreased dopaminergic activity cause significant distress and impairment in at least half of smokers when they quit (Reference HughesHughes, 2007a ). Anger, anxiety, depression, difficulty concentrating, impatience and insomnia peak in the first week and last up to 4 weeks (Reference HughesHughes, 2007b ). Appetite can be increased for more than 10 weeks (Reference BenowitzBenowitz, 1998) and weight gain following cessation is significant (Reference O'Hara, Connett and LeeO'Hara et al, 1998).

Although smokers report that smoking makes them feel calm, this may simply reflect relief of withdrawal from nicotine. The frequent experience of unpleasant withdrawal symptoms in between cigarettes has been suggested as a possible explanation for the association between nicotine dependence and higher rates of mental health problems (Reference ParrottParrott, 2006). Withdrawal symptoms significantly overlap with symptoms of mental illness and it is sometimes difficult to distinguish the two when assessing the impact of smoking cessation. Staff and patients may interpret signs of nicotine withdrawal as a worsening of mental health symptoms that requires relief with cigarettes (Reference Lawn and PolsLawn & Pols, 2003). However, a review showed that smoke-free policies in psychiatric clinics had no negative effect on mental illness even when it led to withdrawal symptoms (Reference Haustein, Haffner and WoodcockHaustein et al, 2002).

A number of studies have shown that quitting smoking exacerbates mental health problems in people with depression. The most recent review (Reference HughesHughes, 2007c ) identified seven empirical trials which excluded smokers with major depression or who were taking antidepressants at the start of the trial but all had methodological limitations. Incidence of current major depressive disorder was 0–14% among all smokers who tried to quit, 3–24% in those with previous major depressive disorder who tried to quit and 1–31% among smokers who became abstinent for several weeks or more. Hughes concluded that smokers with a past history of major depressive disorder are more likely to have a post-cessation relapse of depression. However, the evidence is unclear whether abstinence increases the incidence of the disorder or whether treatment with antidepressants prevents post-cessation relapse.

Schizophrenia

There is little evidence to suggest worsening of symptoms after people with schizophrenia stop smoking. For example, a 3-day study of 60 patients on acute psychiatric wards, 44 of whom were smokers, found no clear evidence that stopping smoking exacerbated psychotic illness (Reference Smith, Pristach and CartagenaSmith et al, 1999). In another study, of 256 in-patients with chronic schizophrenia (Reference Kitabayashi, Narumoto and ShibataKitabayashi et al, 2006), a smoke-free policy had only minor unfavourable effects on psychiatric status and body mass index. There was no indication of associated deterioration in symptoms or functioning during a smoking cessation trial involving 298 smokers with psychosis who received NRT, motivational interviewing and CBT (Reference Baker, Richmond and HaileBaker et al, 2006). However, akathisia can occur after smoking cessation (Reference Barnes, Lawford and BurtonBarnes et al, 2006).

In summary, given that some biochemical and physiological evidence supports the benefits of nicotine to people with schizophrenia, provision of alternative, less-harmful methods of nicotine intake such as NRT is important for very heavy smokers. There is also opportunity for lower doses of antipsychotic medication after cessation and therefore fewer long-term side-effects.

Aggression

Reviews of smoke-free policy in mental health and addiction settings (Reference el-Guebaly, Cathcart and Currieel-Guebaly et al, 2002; Reference Lawn and PolsLawn & Pols, 2005) indicate that total or partial smoking bans had no major untoward effects on behaviour or the frequency of aggression. On one in-patient unit a smoke-free policy had significantly fewer adverse effects than staff anticipated (Reference Patten, Bruce and HurtPatten et al, 1995). In a study in a maximum-security setting, total disruptive behaviour and verbal aggression significantly reduced following a ban (Reference Hempel, Kownacki and MalinHempel et al, 2002). Another study, also in a maximum-security setting, noted that a temporary increase in physical aggression towards staff may have occurred with the introduction of a smoking ban, although overall the policy was associated with clear benefits and almost no detectable ill effect (Reference Harris, Parle and GagneHarris et al, 2007).

Weight

The increase in weight that smokers experience deters many from attempting to stop and has a potentially significant health impact for those with mental illness. In a large cohort study of the general population (Reference O'Hara, Connett and LeeO'Hara et al, 1998), weight gain was often progressive for a period of at least a year; average weight gain in the first year of smoking cessation was about 5 kg and 33% of ex-smokers gained more than 10 kg compared with 6% of continuing smokers. This weight gain was more than just a return to the weight that smokers would have achieved had they never smoked, and it indicates the potential value of proactively combining healthy eating messages and exercise with smoking cessation.

Effect of medication on smoking behaviour

The choice of antipsychotic medication influences smoking behaviour. Haloperidol has been associated with an increase in smoking (Reference Dawe, Gerada and RussellDawe et al, 1995) and clozapine with a decrease (Reference George, Sernyak and ZiedonisGeorge et al, 1995). However, in comparison with typical antipsychotics, atypical antipsychotics are associated with reduced smoking (Reference McEvoy, Freudenreich and WilsonMcEvoy et al, 1999; Reference Hempel, Kownacki and MalinHempel et al, 2002; Reference Barnes, Lawford and BurtonBarnes et al, 2006), reduced desire to smoke (Reference Hempel, Kownacki and MalinHempel et al, 2002) and with higher rates of smoking cessation (Reference George, Ziedonis and FeingoldGeorge et al, 2000). Switching people with schizophrenia from typicals to atypicals may help them to give up smoking (Reference Barnes, Lawford and BurtonBarnes et al, 2006). This might be explained by the greater capacity of typical antipsychotics to block dopamine D₂ receptors. Another possible explanation is the higher level of anticholinergic side-effects of the typicals, which are reduced by smoking (Reference Minzenberg, Poole and BentonMinzenberg et al, 2004).

Effect of smoking on psychiatric medication

As noted above, smokers are generally prescribed higher doses of antipsychotic medication than non-smokers. This is because the toxic products of tobacco combustion increase the metabolism of antipsychotics by inducing the cytochrome P450 (CYP) enzyme system (Reference Ziedonis, Kosten and GlazerZiedonis et al, 1994). The enzyme CYP1A2 is responsible for the metabolism of many medications as well as activation of some carcinogens. Smoking can thereby lower the serum levels of medication by as much as 50% (Reference Wilhelm, Arnold and NivenWilhelm et al, 2004). Medication metabolised by CYP1A2 includes diazepam, haloperidol (partial), olanzapine (partial), clozapine, mirtazapine (partial) and tricyclics (Reference BazireBazire, 2003) as well as opiates, barbiturates and benzodiazepines (Reference Schall, Pries and KattaSchall et al, 1996).

Reduction of inequality

It is estimated that half of the difference in survival to 70 years of age between social classes I and V in the UK is due to the higher prevalence of smoking in class V (Reference WanlessWanless, 2004). People with mental illness already experience health inequality, and smoking exacerbates it. Mental health services committed to a smoke-free policy could help to redress the balance.

Health promotion and addressing unmet needs

Smoke-free mental health services can help to address physical health issues and health promotion for people with mental health problems. Such work can coordinate with primary care in addressing other physical public health issues such as obesity.

Improved well-being

Psychological problems decline significantly in smokers who stop smoking for 6 months (Reference Mino, Shigemi and OtsuMino et al, 2000). Levels of anxiety reduce following cessation, and quitting smoking is followed by enduring mood improvements (Reference ParrottParrott, 2006). Patients may also experience increased self-esteem following cessation (Reference CookeCooke, 1991).

Long-term cessation

A review of smoking bans in the workplace (Reference Fichtenberg and GlantzFichtenberg & Glantz, 2002) found 15% of smokers stopped completely and others reduced their consumption. Bans in mental health settings have had little effect on smoking cessation in the longer term, although this may be partly explained by poor coordination between in-patient, out-patient and smoking cessation services (Reference el-Guebaly, Cathcart and Currieel-Guebaly et al, 2002; Reference Lawn and PolsLawn & Pols, 2005). Specialist services for smokers with mental illness have achieved abstinence rates as high as those for the general population (Reference Foulds, Steinberg and RichardsonFoulds et al, 2006). With improved availability and coordination of smoking cessation treatment, longer-term cessation is likely to increase.