Introduction

Adolescence is a period of changes that identifies the transition from childhood to adulthood. The need for independence and the acquisition of new abilities associated with several physical and brain changes during this critical period prepare the individual to assume adult roles [1]. Adolescence and early adulthood is also a period of increased vulnerability to mental ill health partly because of biologically based changes in brain structures involved in emotional/motivational functions that contribute significantly to risk-taking behaviors and sensation seeking [1, 2]—and a period of increased exposure to adverse life events, which may raise independently the risk of mental ill health [1]. A life event may be generally defined as “a detectable occurrence representing discrete changes in the subject’s social or personal environment that is external and verifiable rather than internal or psychological” [3].

Adverse life experiences during development may induce significant biological changes (biological embedding) and modify the maturation and responsiveness of allostatic systems, thus exerting long-term effects on nervous, endocrine, and immune systems [4]. This is one of the reasons why exposure to adverse life events has been implicated not only in the development of several psychopathological disorders during adolescence and early adulthood—such as major depression, anxiety, disruptive behavior [5], antisocial behavior and substance abuse/dependence [6], psychosis [7], and suicidal behavior [8, 9]—but also in physical ill health [10, 11]. For example, Flaherty and colleagues [10] found that more than 90 % of young adolescents in their sample had experienced some adversity, such as physical abuse, sexual abuse, psychological abuse, neglect, parental substance use, parental depression, or parental criminality, and more than 25 % had at least one health problem.

Suicide is the second most common cause of death during this period of life, the third cause of death in male adolescents (after car accidents and violence) and the first in female adolescents aged 15–19 years [12]. Major risk factors for youth suicidal behavior not only include socio-demographic, educational, psychiatric, and psychological vulnerabilities, but also family adversity, interpersonal difficulties among peers, and adverse life events [12] including specific adverse experiences, such as sexual or physical abuse [13–15] and maltreatment or neglect [8]. The association between these experiences and youth suicidality has received much attention. Several cross-sectional studies have found that sexual abuse is an independent predictor of suicidality in adolescence/early adulthood even after controlling for the presence of risk factors, such as major depression, hopelessness, and other life adversities [16, 17]. There are also longitudinal studies demonstrating a link between physical/sexual abuse and neglect and youth suicidality [18–24]. Some studies also investigated the roles of abuse and neglect relative to other adverse experiences. For example, Thompson et al. [23] reported the existence of a significant link between cumulative lifetime adversities and suicidal ideation. Individually, however, the most predictive adversities of suicidal ideation were childhood physical abuse, childhood neglect, childhood family violence, childhood residential instability, adolescent physical abuse, adolescent sexual abuse, adolescent psychological maltreatment, and adolescent community violence. Nonetheless, the role of specific as opposed to cumulative life adversities in youth suicidality is still poorly understood. With this systematic review, we sought to investigate the association between the type and number of adverse life events and experiences and suicidal behavior in young people.

Methods

Eligibility criteria

To achieve a high standard of reporting, we adopted the ‘Preferred Reporting Items for Systematic Reviews and Meta-Analyses’ (PRISMA) guidelines [25]. Adverse life events were as follows: (1) maltreatment and violence (sexual/physical/emotional abuse, emotional/physical neglect, witnessing home/community aggression); (2) loss events (separations, death of a parent or close friend); (3) intra-familial problems (parental divorce, family instability, social or economic problems); (4) school and interpersonal problems (failure of grade in school or in exam, breaking up with a close friend, and poor social relationships).

We included studies that explicitly mentioned the association between negative/adverse life events and suicidal behavior (OR suicidal ideation OR suicidal thoughts OR suicide attempts, and excluding completed suicides) in clinical and non-clinical samples (for more details see below) aged 10–25 years.

We excluded studies on completed suicides through the psychological autopsy method because assessments of life events relevant to the decedent are frequently dependent on second-hand reporting. Based on the current literature, prospective surveys registering suicide victims have been also excluded as not focused on suicidality among those aged 10–25 years.

When a title or abstract seemed to describe a study eligible for inclusion, the full-text article was obtained and carefully examined to assess its relevance for our review.

Specifically, our exclusion criteria were as follows: (1) studies using adult (>25 years) samples; (2) studies published before 1980; (3) studies without abstracts or with abstracts that did not explicitly mention the association between suicidal behavior in adolescence/early adulthood and negative/adverse life events or life adversities; (4) studies that were not published in English; (5) studies including subjects who died by suicide and using the psychological autopsy method.

Information sources

We conducted a systematic search of 4 major electronic databases comprising medical and social science studies (PubMed, Scopus, Science Direct, and PsycINFO) for titles and abstracts (January 1980–January 2015) relevant to our research question. We additionally hand-searched bibliographies from retrieved articles and from published reviews. We also contacted study authors for further details about the included studies.

Search terms

The following search query was used in Pubmed: adolescent (MeSH) AND Suicide (MeSH) AND (epidemiology OR rates OR trends OR incidence) AND [adverse life events (MeSH) OR adversities OR maltreatment (MeSH) OR abuse (TI) OR neglect (TI) OR parental death (TI)]. In Scopus, the search query was: TITLE (adolescent) AND TITLE-ABS-KEY (suicide) AND TITLE-ABS-KEY (adverse life events). Another search strategy was used about the same topic in Science Direct: (TITLE-ABS-KEY (adolescent) AND TITLE-ABS-KEY (suicide) AND TITLE-ABS-KEY (life events). In PsycInfo, the search query was “adolescent” AND “suicide” OR “ideation” AND “life events” OR “abuse” OR “parental death.”

Selection of studies

Articles were screened and selected in a two-step process to minimize bias. First, two independent researchers (C.M. and G.P.) conducted the literature search. Any discrepancies between the two reviewers who, blind to each other, examined the studies for possible inclusion were resolved by consultation with the senior reviewers (E.F. and M.A.). In the second phase, full-text articles that met our inclusion criteria were retrieved and independently reviewed by G.S. and M.P, who discussed the design and characteristics of the studies to test whether they could be included in the review. If doubts remained, the study was put on the list of those awaiting assessment, pending acquisition of more information, and then was carefully re-analyzed for possible inclusion. Any disagreements in this step were resolved by discussion between reviewers.

Data collection process

A data extraction document was developed [23]. C.M. and G.P. independently extracted the following data elements from the 28 studies included in this review (see ‘Study sample’ below): author/s and publication year, study design, sample size, follow-up, main findings, and main adversities (see Table 1). Reviewers acquired the full text of all 28 articles. The principal reviewers (G.S. and M.P.) analyzed independently all studies. Any disagreements were resolved by discussion with the senior reviewers (E.F., M.A.), who also independently read all articles.

Table 1 Selected cross-sectional studies investigating the association between early adverse life events and adolescent suicidal behavior (N = 15)
Full size table

Summary measures

We assessed the selected 28 studies for quality using the following criteria: (1) representativeness of the sample from the general population (0–2 points), (2) presence and representativeness of a control group (0–2 points), (3) presence of follow-up (0–2 points), (4) evidence-based measures of adverse life events/adversities (e.g., Child Trauma Questionnaire, Life Events Checklist, or other psychometric evaluation) (0–2 points), (5) presence of raters who identified independently the presence of adverse life events (0–2 points), (6) statistical evaluation of inter-rater reliability (0–2 points), and (7) evidence-based measures of suicidal ideation or suicide attempts (e.g., Suicide Risk Scale, Suicidal Ideation Questionnaire, Beck Hopelessness Scale, or other psychometric evaluation) (0–2 points). Quality scores ranged from 0 to 14. Studies were differentiated in quality as follows: (1) good quality (10–14 points) if most or all the criteria were fulfilled, or, where they were not met, the study conclusions were deemed very robust; (2) moderate quality (5–9 points) if some criteria were fulfilled, or, where they were not met, the study conclusions were deemed robust; (3) low quality (0–4 points) if few criteria were fulfilled or the conclusions of the study were not deemed robust. Caution was exercised in interpreting the findings from the low-quality studies (Tables 2, 3, 4).

Table 2 Selected longitudinal studies investigating the association between early adverse life events and adolescent suicidal behavior (N = 7)
Full size table
Table 3 Selected retrospective studies investigating the association between early adverse life events and adolescent suicidal behavior (N = 6)
Full size table
Table 4 Most relevant clinical findings about the association between suicidal behavior and life adversities by type of adversity
Full size table

Results

Study sample

The searches in Pubmed, Scopus, Science Direct, and PsycInfo databases revealed, after the removal of duplicates (17 articles), a total of 235 potentially relevant articles. In particular, the search in Pubmed generated 149 articles, that in Scopus and Science Direct generated 20 and 45 additional articles, respectively, and the search in PsycInfo provided other 38 articles. Of these, 124 were excluded because they were without an abstract or had an abstract that did not explicitly mention suicidal behavior (or suicidal ideation, suicidal thoughts, or suicide attempts) and adverse life events. Four articles were excluded because they were not published in English, and 8 were studies published before 1980. Therefore, 111 full-text articles remained. Of these, 81 were excluded because they did not critically analyze the link between adverse life events and suicidal behavior in adolescence/early adulthood, and 2 were excluded because they were psychological autopsy studies. Thus, 28 articles met our inclusion criteria and were, therefore, used for the present review. Figure 1 summarizes the main results of the search strategy (identification, screening, eligibility, and inclusion process) used for selecting studies.

Fig. 1
figure 1

Flowchart of the search and selection process

Full size image

Study types and sample characteristics

We selected 11 cross-sectional studies including 31,833 individuals, 4 case–control studies including 72,979 subjects and 69,497 controls, 7 longitudinal follow-up studies including 6113 individuals, and 6 retrospective studies including 45,455 subjects and 423,670 controls. Clinical samples included mainly adolescents with major depression or borderline personality disorder, and adolescent inpatients at risk for suicide.

Study quality assessment

According to our quality score system, the mean score of the 28 studies that were included in this review was 5.8. Most of studies (N = 15) were of moderate quality, 3 were of good quality, and 10 of low quality. Below we discuss the main findings from these 28 studies, grouped by life event specification.

Studies on the association between the number of adverse life events and suicidal behavior

In general, it appears that adverse life events cause distress. McKeown et al. [26] showed in a 2-year follow-up longitudinal study that negative life events (such as financial problems, death of a parent or a close friend, parental divorce, and childhood abuse) were significant predictors of subsequent suicide plans (OR 1.10). King et al. [27] confirmed this relationship but also showed differential effects by type of suicidal behavior: suicide attempters were significantly more likely to have experienced stressful life events compared with suicide ideators. Similarly, Liu and Tein [28] showed, in a sample of 1362 Chinese adolescents, that negative life events occurred most frequently in suicide attempters, followed by suicide ideators and finally non-suicidal adolescents, suggesting a dose–response relationship between number of negative life events and suicidal behavior. Importantly, this relationship remained significant (although it was strongly attenuated) even after controlling for the presence of internalizing/externalizing problems. More recently, the Kaplow et al.’s [29] cross-sectional study confirmed the significant positive relationship between the number of adverse life events experienced and risk of suicidal ideation, but also partly explained (38 %), via emotional suppression, the effect of adverse events on suicide attempts. The positive dose–response relationship between the number of adverse life events (such as sexual abuse, drug or alcohol abuse by a family member, running away from home and homelessness) and risk to attempt suicide was also confirmed by Bhatta et al. [30], in another cross-sectional study on 3156 adolescents at a juvenile detention facility. Bhatta and colleagues also reported that the risk to attempt suicide was almost 8 times higher for those who had experienced all of these adversities compared to those with no such experiences.

These relationships were also established with longitudinal data. For example, Thompson et al. [23] confirmed that the number of lifetime adversities was associated with adolescent suicidal behavior, but also showed that the impact of adversities early in life could vary depending on whether they occurred during childhood or adolescence. Psychological maltreatment and sexual abuse had a lower impact if they occurred in childhood, and a higher impact if they were experienced in adolescence. Furthermore, childhood adversities moderated the effects of adolescent adversities on suicidal ideation; the effects of adolescent adversities on adolescent suicidal behavior were stronger at lower (compared to higher) levels of childhood adversities. Nrugham et al. [21] also showed moderation, this time by type rather than timing of adversity. Violent life events were strongly associated (OR 3.85) with suicide attempts but only if experienced, not witnessed.

The association between number of adverse life events experienced and adolescent suicidal behavior was confirmed in clinical samples as well. Stone et al. [24] found that female inpatients with higher rates of dependent events at baseline were at higher risk (42 vs. 21 %) of suicidal behavior during the 34 weeks following their discharge from hospital. Horesh et al. [31] in a case–control study comparing the effect of stressful life events on suicidal behavior in three groups of adolescents (suicide attempters with Major Depressive Disorder, suicide attempters with Borderline Personality Disorder, and healthy controls) reported that suicidal patients experienced a significantly higher number of stressful life events in the year before their suicide attempt compared with healthy controls.

Studies on the association between maltreatment and suicidal behavior

The link between maltreatment—such as sexual, physical, or emotional abuse—and suicidal behavior in young people was investigated in twelve studies. Although maltreatment, in general, was related to suicidal behavior [32], effects appeared to differ by its type. Sexual abuse was the type most consistently and strongly associated with suicidal behavior [17, 30, 33]. For example, Bensley et al. [34] in a cross-sectional study on 4,790 students reported that the association between history of abuse and suicidal behavior (in five levels of severity: “none,” “thoughts,” “plans,” “non-injurious attempts,” and “injurious attempts”) was stronger for combined sexual abuse and molestation compared with non-sexual abuse or sexual molestation alone. In addition, the association was stronger for more severe forms of suicidal behavior, such as injurious suicide attempts (OR 47.1) compared to non-injurious suicide attempts (OR 12.0), suicide plans (OR 6.8) or suicidal thoughts (OR 4.4). Injurious suicide attempts, different from self-injurious behavior (‘self-harm’) [35, 36], may be described as attempts aimed to kill oneself by intentionally cutting, burning, bruising, or otherwise self-injuring.

The role of physical abuse in suicidal behavior in young people is less clear. There are reports of null effects [37], although some studies suggest an independent association, even after accounting for sexual abuse. For example, Johnson and colleagues [18] found that, after controlling for covariates, sexual and physical abuses were significantly associated with risk of suicide attempts during late adolescence/early adulthood (ORs 7.22 and 5.10, respectively). A 6-year follow-up study [38] also suggested that a history of physical abuse increased the risk of suicidal ideation and suicide attempts (ORs 3.6 and 5.6, respectively), even after controlling for gender and other factors. Finally, Brezo et al. [20] showed that the prevalence of lifetime suicidal ideation was higher in their physically abused group (36.6 %) compared to the non-abused group (25.4 %), although those who were sexually abused had higher odds of repeated and late-onset suicide attempts and suicidal thoughts than those who were physically abused. Importantly, the prevalence of lifetime suicidal ideation was higher for young people who experienced both sexual and physical abuse (58.1 %). That study also showed that the impact of abuse frequency on suicide attempts depended on the identity of the abuser, with abuse by a member of the immediate family carrying the greatest risk (RR = 5.0).

The role of emotional abuse or neglect in suicidal behavior was investigated in two studies. Lipschitz and colleagues [39], who conducted a cross-sectional study on 71 adolescent inpatients, found that emotional neglect was a significant predictor of both suicide attempts and self-mutilation. They also reported that emotional neglect was more strongly associated with suicidal ideation and self-mutilation than physical abuse or physical neglect. Tanaka et al. [22] in a 2-year follow-up study found that low self-compassion, which was associated with emotional abuse and neglect, was significantly related to psychological distress and suicidal behavior.

Studies on the correlation between parental death, parental divorce, or family climate and suicidal behavior

The correlation between parental death, parental divorce, or family climate (such as parenting and inter-parental relationship) and suicidal behavior among young people was investigated in six studies. In general, parental death appeared to raise the odds of youth suicidal behavior [40] particularly if the death was a suicide [41]. Parental divorce and also the overall family climate appeared to be associated with this risk, as well. For example, Johnson and colleagues [18], in a longitudinal study conducted on a community sample of 659 families, found that parental separation or divorce was associated with subsequent suicide attempts (OR 1.20). However, maladaptive parenting and harsh parental discipline also raised significantly the odds of suicidal behavior. The role of these and other, related, aspects of the family environment was explored in four studies. King et al. [27] reported significant associations between suicidal behavior and poor family environment (OR 3.6), low parental monitoring (OR 5.0), and parental history of psychiatric disorders (OR 2.0). Liu and Tein [28] found that inter-parental conflict was related to both suicidal ideation (OR 1.94) and suicide attempts (OR 2.67), and Xing et al. [42] confirmed the link between suicidal behavior, harsh parental discipline and maladaptive parenting. By contrast, a supportive and positive family climate appeared to be a protective factor for suicidal behavior in youth. McKeown et al. [26] found that family cohesion was a significant protective factor for suicide attempts although not for suicide plans or suicidal ideation.

Studies on the association between school/interpersonal problems and suicidal behavior

Three studies investigated the association between school/interpersonal problems and suicidal behavior. Baldry et al. [43] found that both direct and, more strongly, relational victimization at school were positively associated with suicidal cognition in youth, and Johnson et al. [18] that a high level of school violence was significantly related to suicide attempts (OR 3.53). Liu and Tein [28] examined, in a sample of rural Chinese adolescents, the role of several school-related problems and adverse experiences in suicidal behavior. Of those, school dissatisfaction had the largest OR (2.34) for suicidal ideation, followed by very high parental expectations (OR 1.99), and change of (or suspension from) school (OR 1.98). The risk of suicide attempts was raised for those who failed in an examination (OR 2.93), felt pressure to enter a better school or college (OR 3.23), and changed or were suspended from school (OR 3.16). However, when all life events and school experiences and other covariates such as age, gender, and family socio-economic status were considered simultaneously, only school dissatisfaction and very high parental expectations remained significant predictors of suicidal ideation (ORs 1.87 and 1.51, respectively). None predicted, independently, the risk of suicide attempts.

Conclusions and discussion

Summary of main findings

The main purpose of this systematic review was to investigate the association between experience of negative life events and suicidal behavior in adolescence and early adulthood. The adversities examined included (1) sexual abuse and molestation (sexual abuse without sexual contact); physical abuse and maltreatment; child abuse and neglect not otherwise specified; (2) family dysfunction and exposure to domestic violence; (3) separation from or death of a biological parent, family member or close friend; parental divorce; (4) poor interpersonal relationships and breaking up with boyfriend/girlfriend; (5) victimization/distress at school. Based on the main findings from our selected studies, experience of adversities or negative life events was significantly related to youth suicidal behavior [17–24, 26–47]. Another important finding was that some adversities are very common, as is the distress associated with experiencing multiple adversities [18, 24]. The third important finding from this review was the strong, positive dose–response relationship between number of events experienced and risk of suicidal behavior [23, 28, 29]. However, it also appears that the relationship between life adversity and suicidal behavior may differ by type of suicidal behavior. For example, young people who had attempted suicide were significantly more likely than those with suicidal ideation to have experienced stressful life events [21, 27, 31, 36]. In turn, the correlation between suicide attempts and adverse life events seems to differ by type of life event. Young people were at higher risk of suicide attempts if they had experienced maltreatment (e.g., abuse or neglect) [22, 39], and, again, this association differed by type of maltreatment, in line with other studies [13]. Our review suggested that sexual abuse, rather than physical abuse or neglect, appears to be more strongly associated with suicidal behavior [20], with sexual abuse being a particularly powerful predictor of several types of suicidal behavior in young people [17, 30, 33, 36, 48, 49]. For example, in the study of Martin and colleagues [17], sexually abused boys had a 10-fold increased risk of making suicidal plans and threats and a 15-fold increased risk of attempting suicide compared to those who were not abused. (By contrast, the findings about the role of non-sexual physical abuse in suicidal behavior were equivocal [37, 38]). Furthermore, it appears that the impact of sexual abuse is particularly severe if the perpetrator was a family member or an intimate partner. For example, Brezo et al. [20] showed that sexual abuse by a member of the immediate family was associated with the highest suicide risk, perhaps because such abuse occurs more frequently in families with multiple difficulties that do not usually guarantee safe conditions after abuse. Also, sexual abuse by a family member can exert long-term consequences on the development of healthy attachment patterns that are needed for mental health [50]. Sexual abuse by an intimate partner also appears to carry significant risk, such as elevated levels of antisocial, violent, and suicidal behavior [19].

As well as the type, the timing of maltreatment seems to matter for suicidal behavior in young people. Earlier onset of maltreatment/abuse is associated with more adverse mental health outcomes, in general [51–53], but also with suicidal behavior. As shown by another recent study [54], those exposed to physical and sexual abuse in childhood or adolescence were more likely to experience high levels of depression and suicidal ideation in young adulthood, compared to those who were not exposed to any maltreatment. Interestingly, among those who experienced maltreatment, first exposure during the early childhood period was associated with the most negative outcome. In particular, those first exposed to physical or sexual abuse during early childhood reported a 77 and 146 % increase in the odds of depression and suicidal ideation, respectively, compared to those individuals first maltreated later in adolescence. Early, compared to later, exposure to maltreatment in childhood may be associated with the most negative consequences because it occurs in a biologically sensitive period. Abuse and neglect early in life impact significantly on brain development, resulting in emotional, social and cognitive impairments, in turn increasing the risk of psychiatric conditions [55–59], psychopathological and attachment disorders, emotional dysregulation, abnormal stress reactivity and executive dysfunction [60, 61], and, as suggested by the studies we reviewed, suicidal behavior.

However, even less severe forms of childhood adversities can impact on suicidal behavior in young people. Our review showed that factors, both in the school and the home context, that were associated with poor mental health outcomes in young people [62, 63] were also related to youth suicidal behavior. Victimization at school [43], school dissatisfaction [28], and experience of school violence [18] were all related to suicidal behavior in young people. Risk factors in the family included poor family environment, low parental monitoring, low family support and cohesion, inter-parental conflict [26–28, 42], and loss of a family member [40, 41]. Early parent loss, especially by suicide, was particularly important. Young people who had lost a parent by suicide early in life were three times more likely to die by suicide themselves than their non-bereaved peers, and more likely than those who had lost a parent as young adults [41].

Of course, not all children exposed to such adversities will show suicidal behavior later in life. It is, therefore, important to consider, albeit briefly given our study aim, the role of protective factors. In general, there has been a study on the role of protective factors in suicidal behavior [64], but few studies have explored their role in buffering the effects of adverse life events, especially in adolescence. A recent review has pointed to the importance of a positive attributional style, higher levels of agency, and greater social support [65], but more research is needed.

Main limitations

Our review should be considered in the light of several limitations. First, we could not carry out a meta-analysis because our studies included different life events and different outcomes. Also, although our review aimed to summarize systematically the most relevant studies in the field, their inclusion and exclusion may reflect our choice, on the basis of our expertise. Moreover, some studies had small sample sizes and small numbers of suicide ideators or attempters, and, as a result, reduced statistical power. In addition, studies did not always distinguish between suicidal ideation and suicide attempts. Also, most of our studies had adopted retrospective designs, and thus findings may have been hampered by recall bias. Finally, some of our studies recruited heterogeneous samples, included a relatively small number of events, or did not include control groups.

Implications and future directions

Most of the studies included in the present review reported a positive, statistically significant association between life adversities, and suicidality in young people. There seemed to be a strong, positive dose–response relationship between number of events experienced and risk of youth suicidal behavior. While the number of events was significant, their type and timing also mattered. Exposure to adversities (in particular sexual abuse/molestation) during vulnerable periods of life may be a critical risk factor for the emergence of suicidal behavior in adolescence and early adulthood. Future studies should elucidate the extent and type of the association between adverse experiences and risk of suicide in youth.