Epidemiological Evidence on the Link Between Drug Use and Suicidal Behaviors Among Adolescents

Abstract

Depression, substance abuse, and aggressive behavior disorders have shown to be among the most influential risk factors for suicide in community and clinical studies of adolescents and adults. Increasing evidence suggests that adolescents who use and abuse alcohol and drugs are at increased risk for suicidal ideation, attempted suicide, and completed suicide, although conclusive evidence has not yet been presented suggesting that drug use causes suicidal thoughts and behaviors. This paper reviews existing epidemiological evidence on the link between the use of more commonly used drugs among adolescents (e.g., tobacco, alcohol, marijuana and inhalants) and suicidal thoughts and behaviors.

RESEARCH OVERVIEW

There is a long tradition of quantitatively oriented population studies of suicide, dating back to the work of William Farr and Emile Durkheim in the 19th century. Farr used quantitative methods to test hypotheses about education and suicide; Durkheim’s quantitative work on suicide focused on ‘anomie’ and community contextual influences (Farr, 1885, Durkheim, 1897/1951). More recently, it is recognized that suicidal behavior can be understood as the result of processes that may be amplified by interactions among neurobiological, genetic, psychological, social, cultural, and other environmental influences (PHS, 2001; IOM, 2002). Suicidal behavior commonly occurs in the context of a convergence of multiple predisposing factors and current environmental factors that influence its development.

In recent years, environmental exposures in the form of drug taking have become more prominent in discussions about suicide. This prominence is fuelled in part by media attention to co-occurring drug use and suicide (e.g., Kurt Cobain). Although scientific evidence on this topic is scant, the current evidence base on the link between adolescent drug use and suicidal thoughts and behaviors is from secular trends, psychological autopsy studies, clinical and emergency room studies and some epidemiological reports.

The San Diego Suicide Study was one of the first investigations to report an association between drug use and suicide in its initial findings on 133 consecutive suicides under age 30 years and 150 suicides aged 30 years or older (Rich et al., 1986). One striking finding was the frequent occurrence (66%) of drug abuse and associated problems in the younger subgroup of the sample (Fowler et al., 1986). This excess occurrence relative to published population values led the authors to infer that these disorders might have contributed to a rising suicide rate among young people. This argument rests on three lines of evidence: (1) the apparent excess of drug abuse in the sample of suicides compared to expected values of roughly 11–18% for this age group, and (2) an observed excess suicide rate among treated cases of alcohol and drug abuse for this age group, and (3) frequent occurrence of depression syndromes with drug use. Fowler and colleagues (1986) extended the argument with a suggestion that suicide might be regarded as a late manifestation of alcohol and drug use disorders. This study, as well as other psychological autopsy studies, had yielded observed prevalence values in a range from 37% to 66% (Shafii et al., 1988; Brent et al., 1988; Runeson, 1989; Shaffer et al., 1988; Fowler et al., 1986).

Population based epidemiological research on public health problems like suicide often can constrain some of the biases found in research on samples of convenience and enables findings to be generalized beyond the study sample. The NIMH Epidemiologic Catchment Area Program, (ECA) (Petronis et al., 1990) is a five-site study involving over 18,000 Americans adults. The ECA study reported that the risk of making a suicide attempt was independently associated with being an active case of Major Depression, an active case of alcohol-abuse-dependence, or a user of cocaine (Petronis et al., 1990).

The National Comorbidity Survey (NCS) (Kessler et al., 1999) is the first study with nationally representative data on suicide. The NCS was conduced in 1990–1992 among a sample of almost 9,000 Americans aged 15–54. Using NCS data, Borges et al. (2000) reported that alcohol and drug use predicted subsequent suicide attempts after controlling for sociodemographic variables and comorbid mental disorders. The authors found that abuse and dependence are significant predictors among users for three of the 10 substances considered (alcohol, inhalants, and heroin). The NCS also reported that the number of substances used is more important than the types of substances used in predicting suicidal behavior.

Using data from the Methods for the Epidemiology of Child and Adolescent Mental Disorders Study (MECA), a large-scale community study of 9 to 17 year olds, King and colleagues (2001) found that onset of sexual activity, serious physical fights, and even low levels of tobacco, marijuana, or alcohol use significantly increased the risk of suicidal ideation. This was still apparent after adjusting for sex, ethnicity, socioeconomic status, and any psychiatric diagnosis. The authors stressed the importance of parental monitoring, a lack of which was linked to increased suicide ideation and attempts.

Borowsky and colleagues (2001) using National Longitudinal Study of Adolescent Health data, a nationally representative sample of 13,110 students in grades 7 −12 data, examined the factors that predicted or protected against suicide attempts at individual, family, and community levels. Adolescents were more likely to attempt suicide if they used drugs or alcohol, had academic problems, or had a history of being a perpetrator or victim of violence. Teenagers who had protective factors, such as family support and emotional well-being, were 70 to 85% less likely to attempt suicide. This was true even for those individuals with multiple risk factors.

Depression, substance abuse, and aggressive behavior disorders have been found to distinguish suicide attempters from nonattempters in community and clinical studies of adolescents and adults (Lewinsohn et al., 1996; Kessler et al., 1999; Garrison et al., 1993; Petronis et al., 1990; Brent et al., 1993). Drug and alcohol abuse is associated with greater frequency of suicide attempts, more serious attempts in terms of lethality and intent and increased levels of suicide ideation (Lewinsohn et al., 1996; Crumley, 1990). Garrison and colleagues (1993) found that the relationships between alcohol and illicit drug use and suicidal behaviors are most pronounced with the reported use of the more potentially dangerous or ‘harder’ drugs (e.g., cocaine), but remained even when the substance of interest was nicotine. Below, drug-specific evidence on the link between the drugs most commonly used in adolescence (e.g., tobacco, alcohol, marijuana and inhalants) and suicidal thoughts and behaviors is presented.

Research on Tobacco

Cigarette smoking has been linked to suicidal behavior among populations of adolescents and adults (Angst & Clayton, 1998; Tverdal et al., 1993; Paffenbarger et al., 1994; Rantakallio et al., 1995; Malone et al., 2003; Miller et al., 2000; Makikyro et al., 2004; Hemenway et al., 1993; Hemmingsson & Kriebel, 2003; Doll et al., 1994; Tanskanen et al., 1998). Of the prospective studies that have examined the relation between smoking and suicide (e.g., Hemenway et al., 1993; Smith et al., 1992; Doll et al., 1994) some had too few suicides to control for other covariates that may influence the relationship (Smith et al., 1992; Friberg et al., 1973; Angst & Clayton, 1998) and others with larger samples neglected to control for possible confounders (Hemenway et al., 1993; Doll et al., 1994).

Many studies of suicide and tobacco smoking have found dose response effects, i.e., the more cigarettes smoked the greater the likelihood of suicide (Hemenway et al., 1993; Angst & Clayton, 1998; Miller et al., 2000). Miller and colleagues (2000), in a cohort study of 300,00 male US Army personnel found, after controlling for several factors, a strong, positive, and dose related association between smoking and completed suicide. Hintikka et al. (2001) in a general population study of Finnish adults found that daily smokers reported suicide ideation more often than nonsmokers, and also found that daily smoking was associated with poor recovery from suicidal ideation. Hall and colleagues (1993) noted that nicotine has depressogenic effects that could indicate a causal relationship between smoking and suicidality.

Smoking has also been found to be related to suicidal ideation (Choquet et al., 1993) and to suicidal attempt (Garrison et al., 1993). However, the association between smoking and suicidal thoughts and behaviors may not be independent of other risk factors (Smith et al., 1992), e.g., depression or other psychopathology could predispose people both to smoking and suicidal behaviors, or mediate the association. Hemmingsson & Kriebel (2003) found that when indicators of mental health, substance abuse, and early life stresses were taken into account, there was little evidence for a smoking-suicide association among a large representative, prospective sample of Swedish men. Leistikow (2003) commented that overadjustment is likely in this study, as the authors adjusted for risk factors that smoking may have contributed to. There has been a lively debate in the literature on whether the association between smoking and suicide is causal or can be explained by confounding or mediating influences (Davey Smith et al., 1992, and others).

The three most widely known conceptual models offered to explain the observed association between depression and smoking involve: (1) a self-medication hypothesis (depression leads to smoking), as expressed by Kandel and Davies (1986), and Carmody (1989); (2) smoking leads to depression by altering brain chemistry (a pharmacologic hypothesis either via intoxication or withdrawal), as advanced by Glassman et al (1990) and others, and (3) a common cause hypothesis (either genetic, as in the work of Kendler et al., 1993, or environmental, as in the work of Breslau et al., 1991). The association between cigarette smoking and suicidal behavior may be related to impaired serotonin function in depressed adult patients (Malone et al., 2003). This study suggests a biological explanation for the relationship between suicidal behavior and smoking and implies a possible common diathesis relating smoking, suicidal behavior and aggressive behavior. Although depression may play a role once drug use has started, it has not been shown to be a strong predictor of drug use for adolescents (Glantz & Leshner, 2000). In a study of tobacco smoking and depressed mood in late childhood and early adolescence, Wu and Anthony (1999) found that tobacco smoking signalled a modestly increase risk for the subsequent onset of depressed mood.

Reseazrch on Alcohol

The association between alcohol abuse and suicidal behavior has long been recognized and documented (Hawton et al., 1989; Roy & Linnoila, 1986). Hawton and colleagues (1989) indicated a rise in the proportion of patients who had attempted suicide who received a diagnosis of alcoholism between the late 1970s and the early to mid 1980’s. These authors also found that there was extensive use of alcohol, especially by alcoholics, both immediately before and during suicide attempts.

Alcohol use has been found to increase the risk of suicidal behavior in both alcohol dependent and nonalcohol dependent populations and to be associated with approximately 50% of all suicides (Frances et al., 1987). Acute and chronic alcohol use has repeatedly been identified as a predictor of suicidal behavior across samples and designs (Rossow, 1996). Alcohol could be associated with suicidal behaviors in a few different ways: 1) alcohol intoxication may make someone more likely to act impulsively; 2) alcohol may make an individual more uninhibited; 3) alcohol may cause precipitating events for suicidal behavior such as romantic relationship break ups and other arguments (Beautrais et al., 1997; Grant & Hasin, 1999). Suicidal behaviors are not the only consequences of acute alcohol use. According to the United States Department of Justice, nearly 40% of convicted offenders (about 2 million people) under the supervision of criminal justice authorities in 1996, had been using alcohol at the time of the offence for which they were convicted (Greenfield, 1998).

Alcohol myopia asserts that acute alcohol consumption restricts attention to the immediate salient stimuli (Steele & Joseph, 1990), and consequently may block inhibitory impulses. The disinhibition produced by intoxication may facilitate suicidal ideas and increase the likelihood of suicidal thoughts being put into action, often on impulse. According to Hull (1987) and Baumeister (1990) in escape theory, alcohol and suicidal behavior both can be used to “…escape from meaningful awareness of current life problems and their implications about the self” (p. 91).

Research on Marijuana

There are few research publications on the association between marijuana use and suicide ideation and attempts, which could indicate that marijuana use is not an important risk factor for suicidal thoughts and behaviors. Beautrais and colleagues (1999) compared the relationship between cannabis abuse/dependence and risk of medically serious suicide attempts in individuals and randomly selected comparison subjects in a case-control comparison. They found that much of the association between cannabis abuse/dependence and suicide attempt risk arose because: (a) individuals who develop cannabis abuse/dependence tend to come from disadvantaged sociodemographic and childhood backgrounds which, independently of cannabis abuse, are associated with higher risk of suicide attempt, or (b) because cannabis abuse/dependence is co-morbid with other mental disorders which are independently associated with suicidal behavior. Kung and colleagues (2003) found that, after adjusting for sociodemographic factors, marijuana was associated with suicide deaths in the 1993 National Mortality Followback Survey (NMFS) that included youth and young adults.

Research on Inhalants

Numerous reports suggest that early use of inhalants increases the likelihood of progression to more severe drug-taking behaviors or other problem behaviors (Schutz et al. 1994; Johnson et al., 1995; Dinwiddie et al., 1991; Beauvais, 1992; Compton et al., 1994). In a survey of youths in juvenile correctional institutions, McGarvey et al. (1999) reported that adolescents with past inhalant use were more hopeless and depressed than those with no inhalant use history. In a study of youths on probation, inhalant users reported significantly more lifetime thoughts of suicide and suicide attempts (Howard & Jenson, 1999). In a population of runaway and homeless young adults, Greene & Ringwalt (1996) found that youth who used substances (particularly sedatives, hallucinogens, and inhalants) were more likely than those who had not used substances to have ever attempted suicide. Kirmayer and colleagues found that youth who had used solvents at some time in their past were more likely to attempt suicide (Kirmayer et al., 1996).

SUMMARY

There is accumulating evidence from clinical and epidemiological studies on the link between drug use and suicide among adolescents. Additional research is needed before there is a claim that drug use in adolescence is a cause of suicide ideation or attempts. The influence of genetic susceptibility traits (e.g., possible ‘common cause’ of both drug abuse and suicide) or family process characteristics (e.g., parental monitoring and involvement) could be two fruitful domains of likely influence on both early drug involvement and later mood disturbances and suicidal behaviors. Another interesting research area would be to systematically examine whether large-scale prevention efforts aimed at drug use reduce the risk of suicidal thoughts and behaviors.