Bidirectionality of smoking and depression in adolescents: a systematic review

Abstract Introduction Recently, evidence has been accumulating that both smoking and mental health disorders are continuously increasing among adolescents. This systematic review elucidates the research into evidence of the direction of the association and risk factors influencing the relationship between smoking and depression. We also highlight recent studies on the effects of electronic cigarettes and developments on the association between depression and smoking. Methods A literature search was conducted on databases including PubMed, Ovid Medline, EMBASE, and PsycINFO and in relevant neurology and psychiatry journals. Terms used for electronic searches included smoking, tobacco, cigarettes; depression; adolescent, youth; direction. Relevant information was then utilized to synthesize findings on the association between smoking and depression among adolescent population. Results The initial database searches yielded 2,738 related articles. After screening and cross-referencing, duplicate articles, articles published in languages other than English, and studies on animals, social and lifestyle factors, mood disorders, and substance use were excluded. Of these, a total of 122 publications only focusing on smoking and depression in the adolescent population were selected for synthesis in this qualitative systemic review. These include 110 original research articles, eight meta-analyses and reviews, and four reports and websites. Conclusion The relationship between smoking and depression in the literature does not reflect the cause-effect relationship. The lack of evidence on the direction of the association may reflect futile study designs, confounding factors and/or use of indirect measures of depression and quantification of smoking. Future prospective randomized studies should target elucidation of the causal association.


Introduction
Tobacco use is one of the leading causes of preventable deaths in the United States. According to the Centers for Disease Control and Prevention (CDC), 480,000 Americans die of cigarette smoking each year and smoking related illnesses cost about 300 billion dollars every year. 1 The U.S. Department of Health and Human Services report Health consequences of smoking states that tobacco use and addiction mostly begins in early youth or adulthood. 2 About 87% of smokers begin smoking in their teens and become regular smokers in adulthood. 2 According to the 2012 surgeon general's report The health consequences of smoking-50 years of progress, there were approximately 1.2 million adolescent smokers in the United States. 2 While in 2015, the CDC reported an estimated 4.7 million middle and high school adolescent students were current tobacco product users. 3 Each day more than 3,000 teenagers start smoking and around 2,000 adolescents and young adults who are occasional smokers become daily smokers. 2,3 It is estimated that with the current national smoking trends around 5.6 million current adolescent smokers will die prematurely due to smoking. 2 Mental health of adolescents has been a growing concern in the U.S. population. According to the National Institute of Mental Health (NIMH), in 2017, an estimated 3.2 million adolescents aged 12-17 in the United States had had at least one major depressive episode in the past year. 4 This represents 13.3% of the U.S. population aged 12-17. Adolescent females have a higher prevalence (20%) of depression than adolescent males (6.8%). 4 In the United States, from 2009-2017, rates of depression increased by more than 60% while suicide rates due to depression including suicide ideation, attempt, and/or deaths doubled among young adolescents. 5 An association between smoking and depression has been examined by researchers with emphasis on initiation and progression of smoking and its relationship with depression. The aim of this review is to explore these complex dynamics and discuss the associated factors influencing the relationship with depression among smokers in the adolescent population.

Results
The initial search of the databases yielded 2,738 related articles. After cross-referencing, duplicates were excluded and articles were screened for availability of full text in English. Studies on animals, social and lifestyle factors, mood disorders, and substance use were then excluded from among the 2,323 full-text articles available. Screening of the titles and abstracts yielded 434 pertinent articles. Irrelevant articles and narrative reviews were then excluded, leaving a total of 122 pertinent publications that were used in the synthesis of this qualitative systematic review. These articles comprise 110 original research articles, eight metaanalyses and reviews, and four reports and websites.
The relevant available information from these articles was then used to describe the clinical characteristics of depression, risk factors, and the relationship between smoking and depression in the adolescent population ( Figure 1).

Clinical characteristics of depression
The American Psychiatric Association's (APA) Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) describes depression as a serious mental health disorder. Individuals with depression may feel sad, tired, or worthless and lack motivation to do anything. Some may also have difficulty thinking clearly and may have suicidal thoughts. Diagnostic criteria are met if the symptoms have been present for at least 2 weeks, but the severity of the disorder is based upon the number of symptoms present at the time of presentation 6 ( Table 1).

Genetics
Studies have demonstrated genetic variance of 40-80% as a risk factor for the development of depressive disorder. 7,8 Twins have an overall significantly higher rate of depression than the general population with greater concordance among monozygotic than dizygotic twins. 9 The heritability of brain size, various characteristics of brain shape, and regional brain volume as related to the development of depression is still under scrutiny, but functional and structural imaging studies have identified specific brain circuits which are involved in the development of this disorder. 10 Cognitive Depressive symptoms at an early age are a risk factor for development of depression and incapacitated social behavior at a later period in life. 11 Stressful life, adverse events, and negativity in emotions and lifestyle contribute significantly to development of depression in adolescence. 12 These stressful events in life are linked to serotonin transporter gene polymorphism, resulting in initiation and exacerbation of depressive symptoms; thus demonstrating an interaction between genetic sequences and the development of depression. 13

Social factors
Parental depression, family disintegration, and parent-children conflicts strongly contribute to development of depressive psychopathology. 14

Sociodemographic factors
Factors like age, socioeconomic status, sex, and ethnicity were found to be associated with smoking initiation and continuation. Sex is a significant contributor, although some studies found that females were prone to smoking initiation, whereas others report male sex to have an increased risk. 24,25 Similarly, ethnicity has also been significantly associated with smoking behavior; some researchers suggest whites are more at risk while others identify Hispanics and blacks. 26,27 These relationships between race and ethnicity and smoking behavior are intertwined with and influenced by social and environmental factors like poverty, family environment and smoking status, peer attitudes, school and lifestyle exposure, and other psychological factors. [28][29][30] Psychological factors Researchers have implicated personality traits like impulsivity, sensation seeking, risk taking, and/ or rebelliousness as being associated with smoking onset. [31][32][33] Others report that negative characteristics including academic performance, self-esteem, depression or distress, and coping behavior (including anger and anxiety) impact the relationship of smoking initiation in young adolescents. [34][35][36] Personal and environmental factors The social environment, including smoking by parents, siblings, family members, friends, and peers, has been strongly implicated in smoking initiation among young adolescents. [37][38][39] Some studies concluded that parental smoking has an increased effect on girls compared to boys, while others report that poor relationship among parents significantly affects the prevalence of smoking initiation. 40 Other social factors like physical activity, receptivity to tobacco promotions, marketing, and advertising were also reported to influence smoking initiation. 44,45 Researchers have postulated various mechanisms to elucidate the complex relationship between smoking and depression. Some argue in support of the "selfmedication" theory, suggesting that depressed individuals initiate smoking to control or alleviate their mood symptoms. 46,47 It is postulated that nicotine acts as an antidepressant in these cases by releasing neurotransmitters like dopamine, norepinephrine, and acetylcholine. 46 Other researchers suggested that smoking increases stress levels which in turn contribute to development of depression over time. 48 Many researchers propose that the relationship is bidirectional; stress or adverse events trigger initiation of smoking, resulting in development of depression, which leads to neurochemical changes in the brain, influencing smoking behavior. 49 The association is further complicated by risk factors like genetics and social stressors which may also influence the development of both depression and smoking simultaneously. 50,51

Smoking causing depression
Many studies support smoking as a potential precursor for the development of depression. In their prospective cohort studies, Wu et al. 52  where depressive symptoms were more pronounced in current smokers, followed by former smokers, compared to non-smokers, while female smokers were observed to have more depressive symptoms than male smokers. Steuber et al. 55 (Table 2).
Studies have also explored this relationship while observing the effects of substance abuse. Schuler et al. 72 estimated age-related trends to observe the association  1,731 This study was conducted by the Johns Hopkins University on public school students from the mid-Atlantic area aged 8-9 to 13-14 and consisting of an almost equal percentage of male and female subjects, with 75% being African Americans. It was concluded that smoking behaviors were associated with an increased risk for the onset of depression, but the presence of depressive symptoms did not lead to an increased potential for smoking initiation.

National Longitudinal
Study of Adolescent Health 14,634 Adolescent smokers, quitters, and maintainers were more likely to be depressed as compared to their peers who had never smoked. Females were observed to have increased depressive symptoms around the time of onset of smoking.

Raffetti 56 2019
Sweden Longitudinal study 3,959 Current cigarette smoking, snus use, and tobacco dependence were assessed using questionnaires at baseline and 1-year follow-up. The outcome was onset of depressive symptoms measured with the CES-DC scale. The incidence of depressive symptoms at follow-up was greater in current than never smokers at baseline. Current cigarette smoking at the age of 13 years was strongly associated with onset of depressive symptoms 1 year later, with a significant interaction between tobacco use and sex. Feeling dependent on tobacco was also associated with depressive symptoms in males but not in females. Snus and overall tobacco use were not associated with the onset of depressive symptoms.  (Table 3).

Bidirectionality
Although several of the studies above have discussed the relationship of smoking and depression as antecedents of one another, this complex association is multifaceted and many researchers have tried to explore the idea of bidirectionality. Windle et al. 67   Associations between depressive symptoms and smoking were evaluated among adolescents at baseline and 12 months later. Both current and former smokers had higher depressive symptoms than never smokers.
Smoking at baseline predicted depressive symptoms. Among non-smokers, depressive symptoms at baseline predicted smoking at 12 months. Persistent, past, and new smokers had higher depressive symptoms at both waves than non-smokers. Depressive symptoms promote tobacco use in Asian adolescents by making it more likely that an adolescent will begin smoking and less likely that she or he will quit. in young girls across time. However, they were not able to conclude the reverse association of depressive symptoms leading to an increase in smoking.
The bidirectional relationship was also explored by McGovern et al., 47 who followed an adolescent cohort of 1,093 students for 4 years to assess the relationship between smoking and depression across time. They reported that depressive symptoms in early adolescence are a precursor to smoking behavior during late adolescence, whereas an increase in smoking predicts suppression of these depressive symptoms.
These effects were mediated by peer smoking. In contrast, Brook et al. 100 followed 688 teenagers for 13 years and reported that early cigarette smoking during adolescence is a predictor of depressive symptoms in adulthood, but the reciprocal association with history of depressive symptoms was not statistically significant as a precursor for smoking behavior (Table 4).

Bidirectionality and social factors
The relation between smoking and depression is very complex. Much research has been done to determine the direction of association. However, this intricate connection remains obscure due to its multifactorial effects. Researchers have investigated the social factors that influence this relationship, especially in adolescents. Horton et al. 102 observed an increased effect of smoking on depressive symptoms in adolescent females with either negative or high level of positive religious coping behavior, whereas moderate religious coping decreased the likelihood of smoking.
Romantic breakups, interpersonal stress and family disruption, and decreased church attendance were also found to be associated with mental health issues including depressive episodes, along with substance abuse and smoking among adolescents. 103,104 Peer pressure significantly mediates the association between smoking and depression, especially in adolescent females compared to males. 105 Hostility and rebelliousness were also found to be associated with smoking behavior and the latter was also reported to be a significant predictor of development of depression in adolescents. 66,106 Parental attitude including smoking by either parent or living with a single parent or guardian, low self-esteem, difficulty Data were collected from adolescents in the study at the start of grade 6 and in the 7th grade. Among 6th graders who had never smoked, depressive symptoms and hostility played a part in increasing the risk of smoking behavior by the time they reached 7th grade. Students with prior experience of smoking, increases in depressive symptoms, and hostility were associated with more frequent smoking.

Bares 88 2014 United States
Bivariate autoregressive multigroup structural equation model 6,501 This study demonstrated bidirectionality between depressive symptoms and cigarette use among adolescent females only. The association was not observed for adolescent male participants.

Longitudinal survey 7,885
This nationally representative sample survey was conducted to measure the association of smoking with depression. Depressed adolescents were more likely to start smoking as compared to non-depressed individuals. Increased frequency of depressive symptoms increased the rates of smoking initiation.

Wilkinson 81 2016
United States National Longitudinal Study of Adolescent to Adult Health 12,017 Data from the National Longitudinal Study of Adolescent to Adult Health were used to investigate longitudinal associations between high frequency substance use (alcohol, cigarettes, and marijuana) and depressive symptoms. Increases in depressive symptoms were associated with a later increase in frequency of marijuana use for males and an increase in smoking frequency for females. Equally, increases in smoking frequency were significantly associated with approximately a 0.6-point increase for females and a 0.4-point increase for males in depressive symptoms at a later wave. Results indicate a bidirectional relationship between smoking and depressive symptoms for females Leung 63 2011 Australia Longitudinal survey 10,012 Longitudinal study of adolescent women with follow-up over 13 years showed that incidence of depression was higher among smokers and that depressed individuals were more likely to initiate smoking.

Longitudinal survey 2,460
This study was conducted to demonstrate associations between e-cigarette and combustible cigarette use and mental health symptomatology. Adolescents who had never previously used combustible or e-cigarettes were assessed at baseline and at 6 and 12-month follow-ups. Higher baseline depressive symptoms predicted subsequent onset of cigarette and e-cigarette use and dual use of both products. Sustained use of e-cigarettes over the 12-month observation was associated with a greater rate of increase in depressive symptoms over time. Among those who sustained use of e-cigarettes, higher frequency of use was associated with higher depressive symptoms at the final follow-up. Earlier cigarette smoking in adolescence predicts later depressive symptoms in the late twenties. Depressive symptoms during adolescence predict cigarette smoking in the late twenties but not above and beyond prior smoking. The effects of smoking on dopamine transporter (SLC6A3) and dopamine receptor (DRD2) were evaluated, observing that presence of the DRD2 A1 allele significantly increases the probability of smoking in adolescent smokers.

Gibbons 57 2018
United States Longitudinal study 889 A prospective study was conducted to observe the relationship between PRD and smoking in African American children. PRD was assessed at ages ranging from 10.5 to 24.5. Anger and depressive symptoms were assessed at age 12.5 and at age 24.5. Early PRD predicted smoking at later ages. Negative effects mediated this association while cultural socialization was associated with lower rates of adolescent smoking and buffered the relation between PRD and anger.

Roohafza 61 2017
Isfahan Selfadministered questionnaire 5500 Data were collected with the help of a self-administered questionnaire on background characteristics, smoking status, depression, and risk factors. Lower education attainment of fathers was accompanied by higher depression prevalence in adolescents. Parental smoking and sibling smoking increased the depression likelihood for never-smokers. Positive attitude toward smoking increased the probability of depression among never-smokers. A higher level of self-efficacy was related to lower chance of depression. Risky behavior increased depression likelihood in never-smokers, in experimental smokers, and in current smokers. Family conflict increased depression likelihood in never-smokers, in experimental smokers, and in current smokers. Psychosocial problems (behavioral and emotional) were measured at age 13 and again 2 years later to assess the association with smoking behavior in adolescence. Externalizing problems at baseline predicted the onset of smoking 2 years later.
Internalizing problems only predicted smoking among girls. Reversibly, smoking at baseline was only associated with the onset of externalizing problems 2 years later.

Riehm 112 2019
United States Longitudinal survey 7702 Internalizing and externalizing problems were assessed for associations with initiation of e-cigarette, combustible cigarette, and dual-product use among adolescents. Adolescents with high externalizing problems were more likely to initiate use of e-cigarettes, combustible cigarettes, and both products. Adolescents with high internalizing problems were at increased risk of initiating use of e-cigarettes but not combustible cigarettes or both products. PRD = perceived racial discrimination. coping with stress, low grades in school, social isolation and withdrawal, sense of helplessness, and friends with smoking behavior have also been instrumental in the development of the relationship between smoking and depression in adolescents. 107,108 Smoking in adolescence is also associated with both internalizing and externalizing problems; the observation that young smokers may use smoking behavior to improve the mood symptoms and selfmedicate themselves with nicotine to rescue higher functioning has been proposed as a plausible interpretation of the behavior rationale among the adolescent population 109,110 (Table 5).

Electronic cigarettes and depression
There has been an exponential increase in the utilization of e-cigarettes (electronic nicotine delivery systems) among adolescents, escalating from around 7.5% in 2007 to 20.8% in 2018. 113,114 In 2018, these active users increased to 3.62 million, making e-cigarettes the most common tobacco source amongst adolescents. 114 Moreover, e-cigarette use is also implicated in an estimated six times increased year. 117 In their study based on the Youth Risk Behavior Survey (YRBS), Chadi et al. 118  should be considered when investigating a causal association between the two conditions. Moreover, the literature also provides information on various biological and social confounders like genetics and family and peer influence, which significantly obscure this causal pathway and it is therefore important to recognize and control these factors when examining the association. It is also important to acknowledge that multiple studies used surrogate measures like sadness and affect change to evaluate depression instead of using validated scales based on DSM criteria. Therefore, the results of these studies should be interpreted with caution even when they report a significant association, as they may not reflect a true condition.
While it is essential to understand the limitations, it is also important to realize that there is a need for much research in this area. Of importance is to define a study population, which represents a homogenous sample of adolescents in the community to examine and control for the influencing mediators in the causeeffect pathway. It is also imperative to use validated instruments to measure depression and standardize metrics to quantify smoking behavior. Longitudinal data collection and serial sampling will represent the optimal strategy to ascertain the level of evidence for this causal relationship.
Finally, given the rises in both smoking behavior and depressive symptoms among adolescents, it is important to understand this association between smoking and depression. To do so will not only help educate young individuals to better understand the health consequences, but, more importantly, is also critical for designing effective treatment interventions, targeted clinical and behavioral therapies, and preventative public health strategies and policies.

Conclusion
Although, the relation between smoking and depression is well established, the literature does not reflect the cause-effect relationship. Additionally, researchers have not yet been able to demonstrate concrete evidence on the direction of the association.
This is a reflection of futile study designs, exclusion of confounding factors, and/or use of indirect measure of depression and quantification of smoking. Therefore, the association between depression and smoking should be explored further with rigorous longitudinal studies controlling for influencing factors to identify the causal association between the two conditions.