Acute trauma factor associations with suicidality across the first 5 years after traumatic brain injury

Introduction

Individuals with traumatic brain injury (TBI) are 3–4 times more likely to die as a result of suicide as compared to the general population.¹ TBI survivors also have a high rate of suicidal ideation (SI), which is known to carry a 5-fold increase in the risk of suicide attempt (SA), and these risks persist for up to 15 years after TBI.^2,3 Many of the risk factors for suicidality (a broader term used to incorporate both suicidal ideation and attempt) in non-injured populations overlap with chronic sequelae of TBI, including aggression, poor cognitive inhibition, poor problem-solving, reduced ability to benefit from predominantly verbal counseling interventions, impulsive behavior, substance abuse, and psychiatric disorders like major depression.^3–7 Veterans with the polytrauma clinical triad of posttraumatic stress disorder, TBI, and chronic pain have significantly increased odds of suicidality, particularly in the setting of substance abuse.⁸ While chronic pain is common after TBI, its role in both SI and SA seems to be related more to an individual’s perception or acceptance of pain, rather than pain severity,⁹ which is in turn associated with disability,¹⁰ an additional known risk factor for suicidality.^11,12 Severity of TBI has not been linked consistently to depression or risk of suicidality,^7,13,14 but other injury-related factors, such as severity of extra-cranial injury (ECI), that may lead to both increased distress and greater disability, have not been assessed in previous studies.

March et al. (2014) reported that adults with unintentional major traumatic injuries, indicated by total Injury Severity Scores (ISS) >12 (combined head and extra-cranial ISS), have over four times the risk of SA or completed suicide compared to the general population, even after adjusting for psychiatric conditions (anxiety/mood disorders, substance abuse), physical comorbidities, and other psychosocial factors (income, residence).¹⁵ Similarly, those with life-threatening physical illness, such as TBI, stroke, myocardial infarction, and spinal cord injury (SCI) have higher rates of SI (11.3%) developing up to two years after illness onset, compared to the general population.¹⁶ Ryb et al (2006) suggest that higher suicide rates among traumatically injured individuals compared to the general population may be due primarily to a higher prevalence of alcohol abuse in those with traumatic injuries.¹²

The goal of the study was to determine whether severity of head and extra-cranial injuries were associated with SI or SA after TBI. We hypothesized that: (1) severity of head injury would not have a significant association with SI or SA; (2) severity of ECI would be associated with and predictive of SI and SA; and (3) physical disability and substance abuse would be associated with an increased risk of SI and SA.

Methods

Results

Of the 3,575 participants in the final linked data set, 293 (8.2%) endorsed suicidal thoughts in the two weeks prior to scheduled follow-up interviews but did not attempt suicide in the previous year, and 109 (3.0%) attempted suicide at some point during the year prior to the follow-up interview. Baseline and follow-up characteristics are shown in Table 2. Those with SI only, and those who attempted suicide, were more likely to have been using drugs at injury (41.3% vs. 33.2%, p=0.012; and 46.2% vs. 33.2%, p=0.117, respectively) and were generally younger (mean 37 vs. 42, p<0.001; and 33 vs. 42, p<0.001, respectively) than those with no SI or SA. Additionally, a larger percentage of ideators and attempters reported problems with alcohol or drugs at follow-up (all years), and these participants also reported higher depressive symptoms at follow-up (all years) compared to non-ideators and non-attempters. There were statistically significant differences in Functional Independence Measure of motor-based activities of daily living (FIM-mot) between those with and without SI and SA at time of discharge and follow-up and in FIM cognitive disability (FIM-Cog) at follow-up. However, these differences did not represent a meaningful effect size (FIM motor differences <17 points, FIM cognitive differences <3 points²²). Ideators and attempters differed significantly at 1 and 2 years post-injury (20.8% vs 31.2% and 26.0% vs 32.4%, respectively; p<.001) with regard to problem substance use.

Discussion

Recent work suggests that self-injury is the eighth leading cause of death in the United States,²³ and understanding the underlying reasons for suicide remains a significant public health problem. To our knowledge, this is the first study to link acute trauma factors to suicidality across the first 5 years post-TBI, and the analysis revealed a novel finding that those with severe ECI may be at the greatest risk for SI even many years after TBI. While psychological factors, such as higher prevalence of depression after TBI,²⁴ are strong contributors to the elevated risk of suicidality following TBI, there is also evidence that neurobiological factors, such as TBI-induced tissue loss and neuroinflammation, may contribute to suicidality after TBI.^25–27 Consistent with previous research,^13,28,29 we did not find significant associations between SI/SA and multiple indicators of TBI severity (maximum head AIS, GCS, time to follow commands). However, we did find that drug use at time of injury and substance abuse at follow-up, severe ECI, and depressive symptoms were associated with SI after TBI. It was surprising to discover that motor disability over time (FIM-Mot) was not meaningfully associated with SI, contrary to our hypothesis, suggesting that the relationship between ECI and SI was not the result of greater motor disability or less functional independence with basic ADLs. Additionally, ECI was not predictive of depression, suggesting a more direct or alternatively mediated (e.g. through other psychosocial factors over time) relationship between ECI and SI. Perhaps individuals’ overall adjustment to disability, a construct likely determined by a constellation of physical, psychological, and social factors, and not adequately captured by discrete measures of mood or basic functional ability, is a major determinant of post-injury suicidality. Future research is needed to better define and measure adjustment to disability following traumatic injury.

Within the general literature on suicidality, most research focuses on suicidal gestures or attempts, rather than SI in particular. Differences in the significant predictors and associated factors between SI and SA in this study suggest that SI and SA are different, though related, outcomes. Other latent factors not measured in the present study may explain more of the variance in SA. A literature review of studies investigating SI after TBI found SI among those with a range of time post-injury was reported in 3–33% of individuals with TBI.² Those reporting SI have >5 times the risk of SA.² In our sample, participants with SI and participants with SA, shared similar baseline and follow-up characteristics on all factors measured, with the one exception being that attempters were more likely to have substance abuse problems than ideators only. This finding is consistent with recent work suggesting that substance abuse problems contribute significantly to suicide rates nationally.^23,30

Recent studies with non-TBI cohorts report that substance use disorders, second only to depression in some clinical samples,³¹ are highly correlated with both SI and SA.^32–35 These data are also consistent with our findings, as both substance use at injury and post-injury problematic substance use distinguish between those with/without SI in our sample. Interestingly, some studies examining the relationship between substance use and SI or SA in the non-TBI literature suggest that these relationships are bidirectional, and that suicidality may either precede or follow the onset of substance use.^36,37 If drug use at injury in our sample is reflective of premorbid patterns of use, this behavior could be associated with underlying SI prior to the injury, which then persisted through the follow-up period. This idea is consistent with our finding that pre-injury SAs are also associated with long-term SI. These findings reflect the fact that pre-injury psychological and behavioral challenges, including mental health conditions and substance use disorders, can persist post-injury and often complicate recovery,^38–40 highlighting the importance of assessing and accounting for these factors in rehabilitation.

Though not often studied, full injury severity measured by ISS has been informative for some TBI outcomes, like discharge disposition²⁰ and community integration.⁴¹ Our study demonstrates a nearly 3-fold increased likelihood of SI among participants with severe extra-cranial injuries (ISS>15), representing major polytrauma, compared to isolated head injuries. However, the likelihood among those with mild-moderate ECI was not significantly higher. Though the association between TBI and suicidality has been well studied in both civilian and military populations,^1,42,43 there are few studies reporting on the association between ECI severity and SI. March and colleagues reported significantly higher odds of suicide completion or attempt among those with severe traumatic injuries compared to matched controls.¹⁵ That study, however, did not separate cranial and ECI, examine mild-moderate traumatic injuries, or assess SI risk. Individuals sustaining polytraumatic injuries, including TBI, have poorer recovery over time than those with TBI alone or ECI alone.^44,45 Therefore, this is the first study specifically examining ECI severity after TBI as it relates to both SI and SA. Our results may have parallel implications for ECI in the military setting, a population at high risk for polytrauma.

Contrary to our hypothesis, we did not find motor disability to differ to a clinically meaningful degree (e.g. ≥17 points) for those with SI or SA. Extrapolating from spinal cord injury literature, SI/SA was most often associated with control of one’s community activities, spiritual well-being, younger age at injury, and severity of injury.⁴⁶ This literature suggests that restrictions in community-based participation (rather than just basic activities of daily living, as measured with FIM) or differences in age at injury may be factors explaining the association between severity of ECI and SI after TBI. However, these factors require further exploration.