Traumatic Brain Injury and Suicidal Behavior: A Review

Interpersonal theory of suicide

Joiner [109] proposed that the simultaneous presence of desire for suicide and the acquired capability for death by suicide is required for occurrence of lethal or near-lethal suicidal behaviors in individuals [107, 109]. When present simultaneously, two interpersonal constructs, i.e., thwarted belongingness (“I am alone”) and perceived burdensomeness (“I am a burden”), can lead to the desire for death by suicide if an individual develops hopelessness regarding both these states and starts to perceive them as unchanging and stable [107, 109]. Loneliness, absence of reciprocal care, subjective experience of not being part of one’s peer group or family, and social alienation, are all-synonymous with thwarted belongingness [107, 109]. On the other hand, the notion that one’s death is worth more than one’s life to others (liability), self-hate, and one’s perceived conviction of being a burden to family and peer system constitutes perceived burdensomeness [107, 109]. According to this theory, an individual may develop passive suicidal ideation in the presence of either thwarted belongingness or perceived burdensomeness alone [107, 109]. However, transformation of suicidal desire into suicidal intent occurs only when an individual’s fear for death is lowered, together with the presence of desire for death by suicide in that individual [107, 109]. The fear of death, injury, and pain may be attenuated in individuals who have experienced prior pain, violence, or trauma repeatedly, such as by having attempted suicide [110, 111], childhood maltreatment [107], self-injury as part of eating disorders [112, 113], combat exposure [114, 115], experiencing proxy-violence and death (physicians) [116, 117], or conducting animal euthanasia [118]. This process has been referred to as an acquired capability for death by suicide by Joiner [107, 109]. Joiner [109] stated that a higher threshold for tolerating pain and decreased fear of death are the two major characteristics of having a high capability for death by suicide. An individual’s capability for suicide might also be enhanced by personality traits, such as impulsivity [119].

In summary, Joiner’s theory states that a combination of hopelessness, regarding both thwarted belongingness and perceived burdensomeness, may precipitate suicidal desire (ideation), which may further interact with elevated tolerance for physical pain, as well as reduced fear of death by suicide, to increase the likelihood of occurrence of serious suicidal behavior (i.e., lethal or near lethal suicide attempt) [107, 109]. We further interpret Joiner’s model of suicidal behavior as being a consequence of an occurrence of readiness to die and readiness to kill (acquired capability to kill self in Joiner’s model) as explained further in Fig. 1. As further developed later in the article, TBI leads to both readiness to kill and readiness to die which, in turn, elevates the risk of suicide.

OPEN IN VIEWER

Fig.1

Modified model of Joiner’s interpersonal theory of suicide linking cognitive distortions induced by traumatic brain injury (TBI) to suicidal behavior. TBI may lead to physical incapacitation and cognitive distortions of perceived burdensomeness (“I am a burden”) and/or thwarted belongingness (“I am alone”). Both of these cognitive distortions, when present together, may lead to “readiness to die” (modification of the “desire for suicide” in Joiner’s model). “Readiness to kill” (modification of the “acquired capability for suicide” in Joiner’s model) may develop with repeated exposure to pain, violence or trauma, including TBI, leading to a higher threshold for tolerating pain and decreased fear of death. Impulsivity and aggression, endophenotypes of suicidal behavior, may also elevate readiness to kill. In the presence of both “readiness to kill” and “readiness to die”, an individual’s likelihood of engaging in suicidal behavior increases substantially. The model of “readiness to die” and “readiness to kill”, which we are first proposing in this publication, is often characterized by a diminution of the internal conflict between intent to die and intent to live, and of the painful ambivalence that is so common in individuals with suicidal ideation, intent and even plan. Moreover, with each subsequent act of non-fatal suicidal behavior, one’s “readiness to kill” increases due to a concomitant habituation of the fear of death and dying, as well as to an increase in tolerance for pain (dotted line). Although, certain protective factors/deterrents, such as social supports, having children to raise etc., may reduce the risk for suicidal behavior, the availability of means to kill oneself elevate the risk for suicidal behavior, with the availability of lethal means elevating the risk of death by suicide. Also, suicidal behavior may contribute to TBI, due to head trauma resulting from certain suicide attempts, as well as long-term accident proneness, thereby creating a vicious circle.

Stress-diathesis model

Mann et al. (1999) proposed a stress-diathesis model, which states that predisposing vulnerability factors (diathesis or “distal risk factors”) interact with triggering stress factors (“proximal risk factors”) to ultimately escalate death by suicide-risk [108]. Both proximal and distal risk factors are required to reach the threshold for suicidal behavior. Some of the proximal risk factors are psychiatric disorders, including substance abuse, exacerbation of psychiatric illness, adverse psychosocial events, marital or financial issues, and emotional pain [71, 120]. Distal risk factors are comprised of impulsive-aggressive personality traits, TBI, cognitive inflexibility, sensitivity to social stress, pessimism/hopelessness, childhood abuse, and family history of suicide [71, 120].

Despite the heterogeneity of suicidal behavior, the neurobiological, clinical, and demographic characteristics shared by both death by suicide and non-fatal suicide attempt with high lethality suggest the existence of a possible shared diathesis between them [120]. In addition to a psychiatric disorder, diathesis is an important risk factor for death by suicide, as most individuals with a major psychiatric illness do not engage in suicidal behavior [120]. Nearly half of the death by suicide-risk attributable to diathesis is heritable, with men possibly having a lesser heritability than women [121, 122]. Risk of suicidal behavior can be enhanced by the epigenetic effects of childhood adversity, including sexual, physical, or emotional abuse, and subsequent mood-related psychiatric illnesses on an individual’s diathesis [123].

Interestingly, individuals with histories of substance and alcohol abuse, as well as those who are more impulsive and aggressive, are likely to have more frequent head injuries. A possible bidirectional relationship likely exists between these variables. Head injuries, aggressive behaviors, and alcohol and substance use [124–126] may potentiate the diathesis-induced increased risk for death by suicide [71].

In summary, this model attempts to integrate the cognitive and clinical perspectives with the neurobiological phenotypes of suicidal behavior. Diathesis in Mann’s model probably overlaps, at least partly, with the acquired capability of death by suicide in Joiner’s model [127].

Relationship of TBI to death by suicide

Robust data exist supporting a positive association between TBI and death by suicide. Results of a retrospective cohort study based on Danish population registers identified that individuals with TBI had 3–4 times greater risk of death by suicide than those who had not suffered a TBI [33]. A three-times greater risk of death by suicide was identified in adults with a diagnosis of concussion in a recent longitudinal cohort study spanning across a 20-year period [136]. A Swedish study comparing the death records of 200,000 individuals having TBI with those of 2 million control subjects reported that death by suicide was three-times more likely in individuals having TBI compared to the general population control subjects [131]. Even though individuals with TBI have a higher comorbidity of substance use and psychiatric disorders, both of which are independent risk factors that can increase suicide-risk, the elevated rate of death by suicide in individuals with TBI still prevailed, even after adjustment for substance use and psychiatric disorders [131]. Madsen et al. (2018) recently performed a registry-based retrospective cohort study in Denmark and reported that a statistically significant increased risk of death by suicide was present in individuals having medical contact for TBI when compared with the general population without TBI [fully adjusted incidence rate ratio (IRR) = 1.90; 95% CI, 1.83–1.97] [134]. It was speculated that the IRR in this study was lesser than the difference between suicide-rates in individuals with vs. without TBI, measured as hazard rates, standardized mortality rates, or odds ratios, in previous studies [33, 136], probably due to a more extensive adjustment model [134]. The authors reported that, in a fully adjusted model, the absolute rates, as well as the IRRs for death by suicide increased with greater severity of TBI [134]. Also, upon adjustment for multiple confounders, a trend for an increasing risk of death by suicide was reported with increasing number of medical contacts for TBI [134]. Interestingly, even after adjustment for several confounders, a temporal relationship was identified between TBI and increased risk of death by suicide (p < 0.001), with the highest reported risk for death by suicide within the first 6 months (IRR: 3.67; 95% CI, 3.33–4.04) after the last medical contact for TBI. It continued to remain high even 7 years (IRR: 1.76; 95% CI, 1.67–1.86) after the last medical contact related to TBI [134]. Furthermore, the association between TBI and suicide was not merely due to injury proneness or systemic trauma rather than brain trauma, as the suicide-risk was substantially higher after TBI in comparison to that after non–central nervous system-related fractures [134]. Bahraini et al. (2013) completed a systematic review of suicidal ideation/behaviors/death by suicide after TBI and reported that, after adjustment for psychosocial, psychopathological, demographic, and socio-economic factors, which could potentially confound the relationship between these variables, a significantly higher rate of death by suicide after TBI was present [135]. In another systematic review, Simpson and Tate [133] reported that, in comparison to individuals with concussion, individuals with severe TBI had a higher risk of death by suicide (hazard ratio = 1.4; 95% CI, 1.15–1.75). However, the increased risk of suicide in those with concussion is far from negligible. Specifically, based on a meta-analysis of 17 studies examining the risk for death by suicide following a mild TBI/concussion, Fralick et al. [143] reported that, as compared to individuals who had not been diagnosed with TBI/concussion, individuals diagnosed with mild TBI/concussion had about a 2-fold higher relative risk of suicide (2.03 [95% CI, 1.47–2.80]; p < 0.001).

While the risk of suicide is elevated with TBI in both civilian and military/veteran populations, it appears that, against expectations, the magnitude of the association is not as high in the military/veteran studies as in their civilian counterparts. Bryan and Clemans [144] investigated the association of cumulative TBIs with death by suicide-risk in a clinical sample of 161 deployed military personnel referred for a TBI evaluation. They reported that greater numbers of TBIs were associated with greater risk for death by suicide (β [SE] = 0.214 [0.098]; p = 0.03) after controlling for the effects of TBI, depression, and PTSD symptom severity [144]. Brenner et al. [32] reported a greater hazard ratio [1.55 (95% CI, 1.24–1.92)] of death by suicide in veterans with histories of TBI, which resisted adjustment for psychiatric diagnoses, including PTSD. Moreover, the meta-analysis by Fralick et al. (2018) indicated a significant risk for suicide in individuals with history of mild TBI/concussion in civilians (2.36 [95% CI, 1.64–3.40], but not in military/veteran samples (OR 1.46 [95% CI, 0.80–2.58]) [143]. It is possible that non-TBI suicide risk factors in military/veteran populations (such as PTSD) have a greater representation than in the civilian population, and as such, reduce the TBI-control separation in odds ratio for suicide.

Relationship of TBI to suicide attempt

According to a study by Simpson and Tate (2002), over a mean period of 5 years, about 17.4% of the outpatients with TBI had attempted suicide [137]. In another study utilizing the data from the Traumatic Brain Injury Model Systems (TBIMS) funded by National Database National Institute on Disability and Rehabilitation Research (NIDRR), Fisher et al. [138] looked at the rate of suicide attempt in the previous year and reported that 0.8 to 1.7% of the study participants, from a large cohort of patients who had sustained moderate-to-severe TBI, had attempted suicide throughout the 20 years of follow-up [138]. It is intriguing to note that these rates fall barely outside the broad-range of prevalence rates for suicide attempt (2–60%), as reported by Simpson and Tate (2007) in their systematic review of studies spanning various cohorts [133]. However, this wide variation in the prevalence rate of suicide attempt in various studies included in the review by Simpson and Tate (2007) was attributed to significant challenges associated with identifying suicide attempt, particularly in subjects having a history of TBI [133, 137].

Among those who have served in the military, less work has been done to study the relationship between TBI and suicidal ideation/suicide attempt. Gutierrez et al. (2008) looked at a sample of 22 veterans with post-TBI psychiatric hospitalization and reported that 27.3% of the subjects had attempted suicide [139]. In a study by Breshears et al. [140], an interdisciplinary TBI team evaluated a sample of 154 military veterans. In the 2 years following TBI (mild, moderate, or severe), suicide attempt was documented in the VA medical records of 7.1% of the study subjects [140]. However, Gutierrez et al. [139] had reported a much higher rate of post-TBI suicide attempt of 27.3% in a much smaller sample of 22 veterans who had inpatient psychiatric admissions after TBI, or who had a history of TBI (primarily moderate or severe) [139].

Using VA administrative data, Finley et al. [141] identified veterans who had served in OEF/OIF and who had received VA care in the fiscal years 2009–2011. They found a significant increase in the risk for suicidal attempt in veterans who had depression, substance use disorder, or a combination of PTSD and TBI [141]. On the other hand, an increased risk of suicide attempt was not associated with TBI alone [141]. This discrepancy was hypothesized to be due to the means of data collection, i.e., data on both TBI and suicide attempt was based on diagnostic codes documentation in clinical care and not on primary data collection in clinical research [141].

Recently, for the first time, Dreer et al. (2017) examined the participants eligible for and enrolled in one-year, post-injury follow-up in the VA Polytrauma Rehabilitation Centers TBI Model Systems, in order to assess rates of pre-/post-TBI mental health utilization, SIRB, and suicide attempt in these individuals [4]. One year post-injury, about 3% of the surveyed study subjects had attempted suicide [4]. Additionally, higher risk for suicidal behavior was reported independent of the presence or absence of a history of premorbid mental health care utilization or a pre-injury suicide attempt [4].

Relationship of TBI to suicidal ideation

In a systematic review, Simpson and Tate [133] reported that the prevalence rate of suicidal ideation was between 3–33% in individuals who had sustained TBI, with most studies reporting a rate of suicide ideation being more than 10%. A recent study by Mackelprang et al. [142] examined the predictors and rates of SIRB in a cohort of adult patients with a complicated mild to severe TBI, within the first 12 months after they sustained TBI. During the year following TBI, one-quarter of the study subjects reported SIRB during ≥1 assessment period [142]. In an analysis of retrospective data, Tsaousides et al. [34] had also reported similar findings in a community sample. They identified that the prevalence rate for suicidal ideation in individuals with different severities of TBI was 28.3% [34]. In another study utilizing the data from the TBIMS that was funded by NIDRR, Fisher et al. [138] evaluated rates of SIRB and reported that these rates varied from 7.0% to 10.0% over the 20-year follow-up period. Gutierrez et al. (2008) looked at a sample of 22 veterans with post-TBI psychiatric hospitalization and reported that suicidal ideation was present in 72.7% of the subjects [139]. As stated above, Finley et al. [141] had also reported a significant increase in the risk for suicidal ideation in veterans who had depression or substance use disorder, or a combination of PTSD and TBI [141]. Similar to the results for suicide attempt, increased risk of suicidal ideation was not associated with TBI alone [141], a finding attributed to similar limitations of the study, as described above. Furthermore, Dreer et al. (2017) stated that about 13% of their subjects (N = 284) had reported having SIRB one-year post-TBI [4].

Hence, in terms of predicting rates of suicidal ideation and suicidal attempt in individuals with a history of TBI, much less definitive research has been done [135]. This could be related, in part, to limitations, such as reliance on retrospective data (e.g., existing medical records), challenges associated with measuring suicide ideation/suicide attempt, variable follow-up periods, and wide-ranging strategies for measuring outcomes [4].

Facilitation of acts

Disinhibition and impulsivity are characteristic of TBI. Some of the features of the neuropsychological syndrome associated with TBI are impulsivity, environmental dependency, a lack of empathy, aggressive or abusive behavior, selfishness, and lability [148, 153–156]. This syndrome could be understood best as an inability to integrate feedback into one’s behavior in a meaningful way and as a problem with impulse control [157]. Damage to the orbitofrontal region that mediates social comportment and intelligence mediates this behavioral syndrome [155, 159]. Furthermore, facilitation of acts in an emotional context are facilitated by impulsiveness and disinhibition, which could lead to violent behavior directed toward others or self [151]. Research has identified that, in comparison to individuals with other injuries, subjects with a history of TBI have increased aggression toward others [160]. Moreover, individuals with TBI to frontal lobes were found to have greater rates of aggression, compared to the individuals who had TBI located elsewhere in the brain [160]. Additionally, diminished capacity for empathy has also been reported in individuals with TBI [154]. In addition to various premorbid conditions, many other factors might influence violent behavior post-TBI. However, an individual’s ability to inhibit such behaviors might be lowered by an injury to the orbitofrontal cortex [161].

A similar pattern of a decreased ability to inhibit behavior might also follow in SSDV, including death by suicide, via increased aggression [162], increased disinhibition, and altered impulse control [71]. Emphasized by a possible interplay with the deficits in emotional regulation and value attribution (as described below), the orbitofrontal cortex is also activated by social exclusion, as well as other situations involving social rejection and self-blame [163]. Thus, in an individual who may already have a tendency towards increased perseveration about the external environment or other triggers, the diminished ability to inhibit SSDV may enable a suicidal act. Impulsivity and a higher rate of aggression both have been recognized as the endophenotypes for suicidal behavior [71] and can result from TBI-mediated orbitofrontal cortical damage.

Altered modulation of value attribution

Rumination has been found to have a robust association with suicidal behavior and suicidal ideation, as well as with impaired problem-solving ability, depression, and hopelessness [164]. In a study by Homaifar et al. (2012), veterans with a history of TBI and no suicide attempt had significantly fewer perseverative errors on the Wisconsin Card Sorting Test compared to the veterans with a history of TBI and at least one suicide attempt [165]. In fact, the suicide attempters with reduced problem-solving ability have been reported to have ‘cognitive rigidity’ [166]. In turn, the ability to precisely assign value to internal and external events might be affected by increased rumination [151]. Interestingly, the anterior and posterior medial prefrontal cortex (PFC) areas that are known to be frequently damaged in TBI [145] have also been correlated with rumination and self-referential processes that characterize rumination [167, 168]. In individuals having suicidal ideation, there may be difficulties in regulation of rumination, secondary to disruption of these areas in the brain via TBI. Thus, TBI-related perseverative errors might hinder attending to goal-directed behavior, as the individual would be unable to stop ruminating over suicidal thoughts, which may potentially influence the higher risk of suicide in TBI patients. However, more research is needed in this area, as most of the studies highlighted above had small sample sizes and did not effectively take into consideration various preexisting comorbidities, including those that could be a consequence of TBI, including depression.

Impaired regulation of emotional and cognitive responses

One of the identifying features of TBI is impaired regulation of cognitive and emotional responses, which includes depression [169] and emotional instability or lability [170]. Post-TBI, individuals may have increased vulnerability for suicidal behaviors and suicidal ideation, due to their impaired ability to regulate their cognitive and emotional responses. Executive functions, such as problem-solving ability and cognitive flexibility, which also have been implicated in risk for suicidality [152], are further disrupted by impaired emotional regulation induced by TBI. The anterior cingulate cortex, an area that, due to its location, is often damaged in TBI, is primarily believed to be involved in emotional regulation. TBI may also lead to disruption in working memory [172, 173] and prolonged negative emotional states, secondary to increased demands to regulate emotional and cognitive processes. Disruption in working memory in individuals with TBI would require differential and compensatory recruitment of more areas in the brain to complete tasks assigned to them. In fact, in comparison with normal subjects, individuals with moderate to severe TBI have blood flow abnormalities in their frontal lobes on functional magnetic resonance imaging, indicating that, in order to complete the same working memory task, they need to recruit more areas in their brain [173]. Mild TBI also elicits this need for increased activation of frontal lobes during subjects’ task completion [174]. Also, the medial PFC has been reported to have a higher activity related to negative moral feelings, including shame, guilt, and embarrassment [175].

In addition to involvement of areas of the brain located more medially in the frontal lobe, laterally located areas, such as dorsolateral and ventrolateral PFC, may also be damaged in TBI and potentially, the lesions in these areas may lead to dysexecutive syndrome [176]. In this syndrome, impairments in monitoring, organization, set-shifting, planning, and reasoning often co-occur [177]. Researchers have also identified that individuals with TBI have decision-making problems on the Iowa Gambling Test [178]. Compounding the deficits in working memory and emotional regulation due to medial PFC damage, injury to lateral PFC leads to difficulty in ‘set-shifting’ and contributes to increased rumination [167, 177]. Additionally, greater problems with rumination also may be partially exacerbated by impaired working memory. Poor working memory, which likely exacerbates the already impaired regulation of cognitive and emotional responses, and increased rumination in individuals with TBI, likely increases their risk for suicidal behaviors and ideation.

In summary, the person’s inability to regulate cognitive and emotional responses, combined with facilitation of acts and altered modulation of value attribution due to the frontotemporal injuries commonly seen in TBI, might together explain the increased risk for suicidal thoughts and behaviors in individuals with a history of TBI. Hence, in individuals with a history of moderate and severe TBI, all these factors could interplay together and contribute to an increased risk for SSDV [151].

Association between TBI and dementia

A robust association between a single moderate to severe TBI and increased risk of developing progressive cognitive deficits, which may ultimately lead to a later-onset of dementia, has been demonstrated in a number of epidemiological studies [182–193]. However, such an association was not found in several other epidemiological reports, with results possibly explained by recall bias inherent in the retrospective design of these studies [194–202]. Nevertheless, prospective studies support the notion of TBI being a risk factor for dementia [190, 192]. Also, the idea of a ‘dose-response’ relationship between TBI and dementia has emerged, described as increased risk of dementia in individuals with a history of loss of consciousness versus without loss of consciousness [188], as well as increased risk of dementia in survivors of severe versus moderate TBI [190]. Evidence also indicates that onset of dementia may be accelerated by a history of TBI [189, 203–205]. Frequently, dementia associated with TBI has been reported as being of the Alzheimer-type [206].

Histopathological features of TBI and AD

The common pathological features of TBI and AD also support the fact that TBI may precipitate a dementia-like process in the brain. Survival from a single TBI is frequently characterized by generalized brain atrophy, which can be observed 6 months after the injury and may continue to progress for a number of years [207–211]. Histological features of acute TBI, however, are not similar to those of the tau pathology in CTE resulting from chronic repetitive mild TBI [212]. In a study by Johnson et al. (2012), age-matched controls were compared to 39 survivors from a single moderate to severe TBI after ≥1 year of original injury [213]. The authors reported that about 30% of the TBI survivors had a wider distribution and greater density of neurofibrillary tangles, which were observed in a hierarchical distribution similar to that of AD [213]. In addition, acute and chronic neuronal loss and white matter degeneration are also characteristics of all severities of TBI [214].

Increased risk of death by suicide in dementia

As discussed in earlier sections, similar to the increased risk of death by suicide in individuals with TBI, subjects with dementia have also been reported to have a higher risk of death by suicide [215]. In a Danish register-based study, Erlangsen et al. [215] evaluated the records of about 2.5 million participants aged >50 years over a period of 11 years and found that an elevated risk for death by suicide was present in older adults, who were diagnosed with dementia during their hospitalization. The authors reported that death by suicide occurred in 136 persons who had been previously diagnosed with dementia [215]. Also, relative risk of death by suicide was higher in women (10.8 [95% CI, 7.4–15.7]) compared to men (8.5 [95% CI, 6.3–11.3]) aged 50–69 years, both having hospital presentations with dementia [215]. A three-times higher risk of death by suicide was identified in subjects ≥70 years old with dementia, when compared to subjects without dementia of similar age [215]. The risk for death by suicide among subjects with dementia remained significant after controlling for mood disorders. Heightened risk for death by suicide was associated with the time shortly after the diagnosis of dementia, with about 14% of women and 26% of men dying by suicide within the first 3 months after the diagnosis [215]. Later, more than 3 years after the diagnosis, about 41% of women and 38% of men had died by suicide [215]. It has been proposed that alterations in serotonergic function in the brain of patients with dementia may lead, at least in part, to suicidal behavior via impulsive aggression [108, 217]. Also, increased insight into the disorder, may itself lead to increased hopelessness [218] and hence, elevate the risk for suicidal behavior [219]. Interestingly, a bidirectional positive association between insight and depression has also been reported in individuals with TBI and may partly explain the increased risk of suicidal behavior associated with TBI [220]. Likewise, other types of dementia, such as Huntington’s disease [221] and frontotemporal dementia [222], have also been associated with suicidal behavior. As highlighted above, both TBI and AD (along with other types of dementia) are characterized by cognitive deficits, including executive dysfunction [25, 223] and deficits in decision-making capacity [148, 224–227], which in turn have been related to suicidal behavior.

Inflammation and TBI

TBI consistently has been linked with neuroinflammation, in spite of its pathophysiologic heterogeneity [251, 256–258]. Moreover, TBI is used in animal models to induce neuroinflammation [259, 260]. Inflammation itself might play a role, at least partly, in mediating the greater risk of SSDV in TBI, as even after adjusting for depression and its severity, inflammation has been linked predictively to SSDV [256–258]. Even though activation of an immune response is required for healing post-TBI, a poor prognosis is associated with prolonged and intense inflammation after TBI [261–263]. Tardive changes, such as neuronal cell death, delayed neuroinflammation, systemic inflammation, and impaired neurological function are also induced by TBI [263, 264]. Head trauma results in highly complex molecular and cellular responses, with the neuroinflammatory response being responsible for much of this complexity [251, 265]. Injury to the cerebral vascular system [266], a high degree of disruption of the blood-brain barrier [267] (allowing peripheral inflammation to spread easily in the central nervous system), and damage to the white matter tracts [268] are some of the other more specific factors that differentiate inflammation following TBI from other causes of neuroinflammation.

A primed profile of microglia is sometimes established during the post-acute neuroinflammation phase, which is comprised of long-term changes in microglial reactivity, mobility, morphological and biochemical properties [269], and an increased MHC II mRNA and protein expression [270, 271]. TBI, via production of a proinflammatory response, can cause a decrease in the priming threshold for microglial cells [272, 273], or might trigger previously primed microglial cells. A secondary immune challenge, such as allergy, autoimmune disease, or infection, which may be encountered by the already primed microglial cells, could lead to a higher risk for SSDV, as a prolonged and ample state of immune activation might result from these secondary insults that could affect neurons, neurotransmitters and neuronal circuits. Studies in rodents have demonstrated that cognitive deficits [274] and affective dysregulation develop if an immune challenge follows a vulnerable period after TBI [270].

Interestingly, pathophysiology of AD may also involve a significant role of neuroinflammation, especially early in the course of this illness [275–280]. Specifically, clusters of microglia around amyloid plaques, as well as a wide-variety of inflammatory proteins, such as proinflammatory cytokines, acute-phase proteins and complement factors, have been shown to be present in both human subjects with AD and its transgenic models [279]. Also, the pathophysiology of TBI and AD overlap in terms of oxidative stress [281–287], and the double-hit concept of neuronal damage [288–290]. Risk of AD is increased by certain single nucleotide polymorphisms in the genes that encode microglia-specific proteins, which are involved in protein degradation pathways, as well as phagocytosis [291, 292]. Neuroinflammatory response is further aggravated by accumulation of reactive astrocytes around the fibrillary amyloid plaques [293], which operate similarly to microglia, in terms of releasing potentially cytotoxic molecules and cytokines after being exposed to amyloid-β [280]. Increasing age has also been strongly associated with the prevalence of AD, and changes in microglia related to aging have also been postulated to be involved in its pathology [294]. Recent reports have also highlighted a reduction in coverage of brain tissue by microglial cell processes, as well as shortening of microglial cell processes in AD and normal aging [295]. Both experimental and clinical studies have reported a shift towards a pro-inflammatory state in the aged brain [272, 296]. Higher levels of circulating cytokines, like interleukin (IL)-1β, interferon γ and tumor necrosis factor (TNF), in association with an increased microglial and astrocyte reactivity, have been characteristically seen in age-associated neuroinflammation [273, 297].

Inflammation, depression, and suicidal behavior

An increasing amount of evidence has linked inflammation and depressive symptomatology [257, 298–300]. Multiple animal models have also confirmed a causal relationship between depressive-like symptoms and inflammation [301–306]. Recent meta-analyses have concluded that there is also evidence for immune activation in the blood and cerebrospinal fluid of suicide attempters and in postmortem brain samples of individuals who died by suicide [307–309]. The senior author’s (TP) team was the first to identify mRNA transcripts of cytokines in the orbitofrontal cortex of suicide victims [310]. The cytokine profile in this study was consistent with the results from the senior author’s reports on outcomes in rodents sensitized and exposed to aeroallergens, including 1) increased prefrontal cortex cytokine mRNA expression [311], 2) impairment in social interactions [311], and 3) aggressive-behavior after stress [312], mimicking risk factors and endophenotypes of SSDV [71]. Endophenotypes associated with suicidal behavior, such as aggression [81], have been exacerbated (at least partially) by the effects of individual cytokines [301, 313]. A second postmortem study in the same year found pronounced microgliosis in the brains of patients who died by suicide [314]. More recently, increased levels of IL-1β, IL-6, and TNF, cytokines that are excessively produced by primed microglia, were observed at both the mRNA and protein levels in the anterior prefrontal cortex of teenagers who died by suicide [315]. Also, significantly elevated levels of IL-6 in the cerebrospinal fluid of suicide attempters further supports the role of neuroinflammation in contributing to suicidal behavior [316]. Additionally, the protein and mRNA expression of certain Toll-like receptors, which are involved in the production of cytokines, has also been reported to be increased in depressed individuals who died by suicide, as compared to the normal control subjects [317]. Moreover, an increased gene expression of these cytokines in rodents after a peripheral immune challenge has also been reported by a team led by the senior author [312]. In fact, the exaggerated glucocorticoid responses to stress observed after this immune challenge [312] are one of the central and highly replicated pathophysiological features of suicide attempters in comparison to those of non-attempter psychiatric controls [71].

How inflammation triggered by TBI can lead to suicidal behavior?

Transport of tryptophan through the blood-brain barrier occurs via a transporter for large neutral amino acids. Tryptophan is metabolized in the brain through two pathways, which involve synthesis of either kynurenines or serotonin. The rate-limiting enzymes involved in metabolism of tryptophan in the kynurenine pathway are indoleamine-2, 3-dioxygenase (IDO) and tryptophan-2, 3-dioxygenase. Inflammation within the brain induces the expression of IDO. This further increases the production of metabolites downstream along this pathway, including that of kynurenic acid and quinolinic acid [318]. Kynurenic acid serves as a free radical scavenger and antagonizes glutamate to confer neuroprotection. However, quinolinic acid is an agonist at N-methyl-d-aspartate (NMDA) receptors and produces free radicals, which ultimately lead to toxic effects on neurons. The macrophages that infiltrate the brain along with the microglia primarily synthesize quinolinic acid and additional compounds that are toxic. This sequence of events exemplifies the association between inflammation in the nervous system and the degeneration of neurons via activation of the kynurenine pathway [318].

Studies in human and animal models have postulated that pathogenesis of depression may be mediated by NMDA receptors, as well [319–321]. Data from other studies have indicated that disruption of glutamate neurotransmission might be involved in the pathogenesis of depressive symptomatology [322, 323]. A case-control study demonstrated considerably increased levels of glutamate in the occipital cortex of the participants who had depression and who did not receive any medications. These participants underwent comparison with healthy controls who were matched according to gender and age [324]. In fact, a postmortem study that assessed the differences in glutamate receptors in the frontal cortices of human suicide victims versus those of controls who died of sudden death due to causes other than suicide, reported that dysfunction of glutamate receptors may also be implicated in the pathophysiology of suicide [321]. A study performed by Paul et al. (1994) reported that mice that received 17 distinct antidepressant medications/treatments [including citalopram, electroconvulsive shock, imipramine, monoamine oxidase inhibitors, and selective serotonin reuptake inhibitors (SSRI)] over two weeks, as opposed to those that received the medication/treatment for only one day, had adaptive alterations in the NMDA receptor-radioligand binding. These effects were mostly limited to the cerebral cortex [325]. Such constancy of effects from different antidepressant medications and treatments points towards the fact that all antidepressants might act via a similar mechanism that involves adaptive alterations in NMDA receptors [319]. Studies even have shown that Riluzole and Lamotrigine, both of which inhibit release of glutamate, also have effects similar to antidepressants [326, 327]. Also, studies have reported that patients with major depression who had received sub-anesthetic doses of intravenous ketamine, which has antagonistic activity on NMDA receptors, had rapid improvement in their symptoms of depression [328, 329] and suicidal ideation in the setting of major depression [330]. Moreover, Mackay et al. (2006) reported a sustained increase in activation of the kynurenine pathway for a least 1 year or longer after TBI [331]. In essence, TBI may set forth an inflammatory cascade, which upregulates the levels of free radicals and quinolinic acid in the brain, leading to neuronal damage and potentially, to suicidal behavior.

Managing risk for suicidal behavior in general with/without TBI

Clinical trials for managing suicide risk acutely have led to emergence of new treatments. Previously, suicidal patients with severe or treatment-resistant depression usually received a course of electroconvulsive therapy that had a short-term life-saving effect [333]. Acute anti-suicidal effects of a sub-anesthetic dose of ketamine, lasting for at least 1 week, were reported less than 10 years ago [330, 335]. As most antidepressant medications augment the noradrenergic and/or serotonergic functions, and since disturbances in both these neurotransmitter systems have been associated with suicidal behaviors [336–339], it is reasonable to expect an anti-suicidal effect with these medications. However, there is a dearth of direct evidence for such an effect. In a pilot double-blind, randomized, clinical trial, treatment with paroxetine (an SSRI) was reported to be associated with a greater reduction in suicidal ideation and a better antidepressant effect than bupropion (a noradrenergic–dopaminergic drug), with the patients having the most severe suicidal ideation achieving the greatest therapeutic advantage [340]. According to a systematic review [341], antidepressants decrease suicidal ideation, as well as suicidal behavior in depressed patients, and various research strategies together provide reasonable evidence to support this claim. In this review, it was further stated that well-designed epidemiological studies have provided evidence for the prophylactic effect of antidepressants on suicidal behavior, in particular, death by suicide [341]. Randomized control group studies, with some utilizing a placebo arm, provide evidence for the beneficial effects of antidepressants on suicidal ideation [341]. Similarly, Rihmer and Akiskal (2006) reported that, in most countries with traditionally high suicide rates, the widespread use of antidepressants appeared to lead to a highly significant decline in suicide rates, with the decline being more prominent in women compared to men [342].

Counterintuitively, in children and adolescents, a pro-suicidal effect of antidepressant medications has been reported [343]. However, only scarce direct evidence is available that supports a pro-suicidal effect of antidepressants, and later studies have pointed out that this effect is actually less likely [344–346]. Effects of lithium, antipsychotic drugs, and mood stabilizers on suicidal behavior have also been debated by researchers [347]. It is likely that there has been an overestimation of harmful effects on suicide risk by antiepileptic drugs [347]. Reduction of the risk for fatal and non-fatal suicidal behavior by lithium is supported by a sizeable amount of evidence [348]. Protective effect of lithium on suicide may be related to its anti-apoptotic effects [349], serotonin-enhancing properties [350], and its potential to increase the volume of gray matter after its long-term use [351]. In patients with schizophrenia, compared to olanzapine, clozapine has a protective effect on suicidal behavior [352]. Anti-suicidal effects of both clozapine [353] and lithium [354] are independent of their effectiveness as an anti-psychotic and as a mood-stabilizer respectively. These drugs’ specific efficacy profiles possibly indicate that both these medications alter the diathesis, at least partially, to decrease suicide-risk [120].

Managing risk for suicidal behavior specifically in individuals with TBI

It is now well established that individuals with TBI have both an increased risk of death by suicide and greater prevalence rates for suicide attempt and suicidal ideation. Consequently, efforts to screen and prevent suicide should begin at the time of the intake procedure by the rehabilitation clinicians who should also continue with regular follow-ups, especially during phases of transition or stress. Additionally, to identify possible risk factors related to death by suicide, the clinicians engaged in providing psychiatric care should acquire information related to the lifetime history of TBI in their patients [355]. The Rocky Mountain Mental Illness Research Education and Clinical Center TBI Toolkit [356] not only delineates assessment procedures that should be followed by mental health and rehabilitation clinicians, but it also highlights the clinical and research potential of the evidence-based screening tools included in this kit.