Introduction
Of the 1.5 million traumatic brain injuries occurring annually in the United States, approximately 500,000 result in emergency department visits for children under 15 years of age, and 80%−90% of these cases are considered mild.
1 Even if a small proportion of mild traumatic brain injury (MTBI) incidents in children result in behavioral problems, the large number of children affected represents a major public health concern. It is important, therefore, to consider whether MTBI is associated with new-onset or novel psychiatric disorder (NPD). A clear answer to this question has been hampered by a number of factors, including the lack of prospective (rather than retrospective) investigations, a large sample size sufficient to explore a range of injury and social variables, and a persisting assumption based on earlier investigations, that new psychiatric disorder after mild TBI may be unrelated to the brain injury per se.
2 A better understanding of the genesis of NPDs in TBI would allow them to be predicted and perhaps treated early so as to limit morbidity.
NPD, by definition, can manifest in two ways.
3 First, it could emerge after TBI in an individual with no pre-injury lifetime psychiatric disorder. Second, it could represent a novel disorder after TBI in an individual with another form of pre-injury lifetime psychiatric disorder (e.g., a patient with a pre-injury lifetime history of attention-deficit/hyperactivity disorder (ADHD) who develops a major depressive disorder after the TBI). To date, two prospectively-studied cohorts have been used to examine the relationship between MTBI and NPD.
4 We found that children with mild/moderate TBI and a history of psychiatric disorder before the injury were at significantly higher risk for developing NPD in the first 3 months post-injury.
3 Furthermore, after mild/moderate TBI, NPD occurred at rates of 8/37 (22%), 3/30 (10%), 7/31 (23%), and 6/30 (20%), at assessments 3, 6, 12, and 24 months post-injury, respectively.
4 We recently reported that NPD in the first 6 months after injury occurred in 25/70 (36%) of children consecutively hospitalized for MTBI and that NPD was significantly associated with the presence of frontal white matter lesions.
5Additional investigations focused on specific symptom ratings, such as post-concussion symptoms (PCS), attention, and conduct, documented typically from brief parent and child interviews and/or questionnaires administered to parents, children, and teachers, as contrasted with psychiatric disorders derived from standardized psychiatric interviews.
6–12 Results compatible with our findings
3 were obtained in an earlier study of PCS, where children with MTBI whose symptoms increased had poorer pre-injury behavioral adjustment than those whose symptoms did not increase.
11 Another study found that significant ongoing behavioral difficulty 3 months after injury in children with MTBI was related to higher incidence of previous TBI, premorbid stressors, pre-existing psychiatric or neurological problems, and learning difficulties.
9 Findings from a recently-studied cohort of children with MTBI extended these PCS-related findings.
6,7,10,12 For example, a high acute level of PCS was likely among children with MTBI whose acute clinical presentation reflected more severe injury.
10 PCS was significantly higher in an MTBI group than in an orthopedic injury (OI) control group 2 weeks after injury (51% versus 30%), but not thereafter in the first year post-injury (19%−27% versus 19%−21%).
12 Results from a birth cohort study indicated that MTBI resulting in inpatient, rather than outpatient, treatment was associated with increased ratings of hyperactivity/inattention and conduct disorder, especially if the MTBI occurred before age 5.
8 Additional prospective
13 and retrospective
14 studies suggest varying neurobehavioral morbidity after MTBI.
In addition to behavioral and emotional morbidity that characterizes NPD, neurocognitive sequelae of TBI are important clinically;
15 yet the relationship between NPD and neurocognitive deficits in an MTBI population is largely unstudied. Narrowing this knowledge gap is critical because it is clinically relevant to address pre-injury or post-injury general and specific cognitive deficits if they are present, especially within the context of NPD. We reported that NPD was significantly associated with concurrent deficits in processing speed, expressive language, and intellectual functioning 6 months after MTBI.
5 We have found that NPD is significantly related to neurocognitive outcome in hospitalized children with a broader range of TBI severity (severe TBI and MTBI), and OI.
16 Specifically, intellectual functioning and memory outcome were independently significantly related to a “neuropsychiatric factor” (encompassing severity of injury and NPD) and a “psychosocial disadvantage factor” (including socioeconomic status, family psychiatric history, and family functioning). NPD added significantly to severity of injury indices, and both family functioning and family psychiatric history added significantly to socioeconomic status (SES) in explaining several specific cognitive outcomes.
Aside from the few studies investigating the relationship of MTBI, NPD, and neurocognitive outcome, there are several other reports of MTBI, behavioral symptomatology, and neurocognitive outcome. One study found that children with MTBI whose “postconcussive symptoms” increased versus those whose did not increase from before injury to 3 months post-injury, performed significantly more poorly on tests of processing speed, visual memory, attention, and executive functioning.
11 Another study found that children with MTBI and ongoing behavioral difficulties were not differentiated by their performance on verbal memory, visual memory, processing speed, attention, or executive function tests.
9 A recent study found that measures of executive function were not significantly poorer in an MTBI versus OI group in the first year after injury.
7 These negative findings are consistent with a recent study and reviews suggesting benign cognitive outcomes after MTBI.
17–19 However, MTBI was more likely to result in PCS than was OI, among children of lower versus higher cognitive ability. This was especially evident for children with complicated MTBI (lesion detected on MRI).
6It is clear that children with MTBI do not necessarily escape psychiatric morbidity and that this morbidity may be related to neurocognitive morbidity. In this article, we explore each of these forms of morbidity in a large MTBI cohort studied prospectively and, in addition, investigate the relations between NPD and concurrent neuropsychological functioning. Considering the literature reviewed, we hypothesized that NPD in the 6–12-month post-injury interval in children with MTBI would be predicted by indices of injury severity (e.g., presence of a lesion; frontal white matter lesion), estimate of pre-injury academic functioning, and indices of psychosocial adversity (e.g., socioeconomic status, family psychiatric history, pre-injury family functioning). We further hypothesized that NPD (a broad category of psychopathology) at 12 months would be associated with significant deficits in concurrent adaptive and academic functioning, and a broad range (implying nonspecific rather than specific association) of neurocognitive measures, including processing speed, memory, and expressive language.
Discussion
The main findings of this study are that novel psychiatric disorders (NPDs) 6–12 months after mild traumatic brain injury (MTBI) are common, are associated with significant deficits in adaptive functioning, and are significantly associated with pre-injury psychosocial risk factors, including lower SES and psychosocial adversity. Our findings also suggest that NPD is significantly related to estimated pre-injury academic functioning. Finally, NPD appears to be associated with deficits in concurrent neuropsychological functioning across multiple cognitive and academic domains.
NPD after MTBI occurred at a relatively high rate, appearing in 28% (17/60) of the children who returned for the 12-month assessment. This rate is slightly higher than the rate reported by a previous study, which found that 7 of 31 children and adolescents (23%) expressed NPD 6–12 months after mild and moderate brain injuries.
40 Either result has significant public health implications, because of the high incidence of MTBI.
1 The specific NPDs expressed by the children were heterogeneous, a result similar to those of other pediatric TBI studies.
2,40 The most frequently occurring NPDs were secondary ADHD and personality change due to TBI, followed by mood, anxiety, and behavioral disorders. The secondary ADHD finding is particularly striking because 25% (20/79) of the participants enrolled at baseline had pre-injury ADHD and were thus not eligible to develop secondary ADHD. The high rate of pre-injury ADHD, a two-to-three-fold increase, as compared with community prevalence, is consistent with other studies of pediatric TBI and may be related to impulsivity as a contributing causal factor to injury.
41NPD was significantly associated with SES and psychosocial adversity. NPD was also clearly related to other pre-injury psychosocial measures, including family psychiatric history and family functioning, with small-to-moderate effect sizes, respectively. These results contrast with those of the same cohort with respect to NPD in the first 6 months after injury, in that the significant association with frontal lobe white-matter damage is no longer apparent.
5 As time-since-injury increases, the relationship of brain injury indices and psychiatric outcome may decrease,
2,42,43 and there may be a closer connection with psychosocial factors that influence daily life. This pattern is by no means uniform, such that the reverse may be true (e.g., regarding oppositional defiant disorder symptomatology)
44 or changing injury, and psychosocial correlates may occur with the passage of time (e.g., in the case of personality change due to TBI).
45,46 However, the association between psychosocial adversity and NPD/neurobehavioral deficits is one of the most consistent findings across cohorts internationally in pediatric TBI research.
2,4,41,47–49Estimated pre-injury academic functioning was significantly related to NPD. Not surprisingly, estimated pre-injury academic functioning was related to SES (r=0.377; p=0.004). Together, these findings suggest that children with pre-injury psychosocial adversity and associated pre-injury lower academic functioning are at increased risk for the development of NPD. The fact that authoritative reviews
18,19 and a recently published study
17 find benign neurocognitive outcome after pediatric MTBI, makes it unlikely that the early post-injury deficits in academic functioning associated with NPD in this study represent a decline in functioning or are due to brain damage. However, the latter-referenced study had participants with milder injuries than those described in this investigation.
Neurocognitive deficits present 12 months post-injury were associated with NPD across multiple domains of functioning with small-to-large effect sizes. Neurocognitive functioning was so closely tied to SES that an independent relationship of neurocognitive outcome (except for memory function) or SES to NPD could not be consistently demonstrated in regression analyses. This suggests that the neurocognitive deficits were already present before the MTBI in the children who went on to develop NPD. Therefore, children with neurocognitive deficits or lower neurocognitive reserve
50 may be at increased risk for NPD or other behavioral problems.
5 Nevertheless, it remains at least possible that the MTBI caused the significant neurocognitive deficits.
In contrast to the—at least possible—argument about the sequential connection between MTBI and neurocognitive deficits, the methods of this psychiatric study allowed confident documentation of new psychopathology (i.e., NPD, after MTBI). The children with NPD at 12 months post-injury had significantly lower adaptive functioning than counterparts with no NPD. Interestingly, this relationship remained significant even when pre-injury adaptive functioning was controlled. There are several implications of this finding. First, the construct of NPD is clinically meaningful and relates not simply to psychopathology but also to adaptive dysfunction. Second, behavioral change in the form of a psychiatric disorder clearly occurs after MTBI and is associated with a decline in adaptive functioning. These findings, in juxtaposition to the only possible likelihood of significant changes in neurocognitive functioning after MTBI, suggest that NPD may be a more sensitive index for the effects of MTBI. However, an injured control group (e.g., with OI and no TBI), might clarify to what extent NPD at 12 months post-injury may be due to brain damage per se.
The results of this study must be considered within its limitations. First, the MTBI sample was limited to hospitalized children, and there is a growing trend for children with MTBI to be discharged from emergency rooms.
51 Thus, our sample may differ from the entire population of children who have experienced MTBI. As a result, the sample may also reflect negative injury or psychosocial factors that would influence whether the child is hospitalized. An example of a negative injury factor is the high rate of children with an abnormal MRI (48%), compared with a cohort recruited regardless of hospitalization status (17%).
10 Second, interrater reliability assessments for the diagnosis of NPD were not directly evaluated on the basis of videotaped interviews. However, the assessments at each site were closely supervised by the child psychiatrists or psychologist. Furthermore, frequent telephone conferences were held, and transmission of written psychiatric assessment summaries were critiqued by the first author and the other interviewers, which helped maintain fidelity in diagnosis across sites. Third, attrition in terms of participation of enrolled and eligible children was 24%, although there were no differences in injury, demographic, or psychosocial variables between the children who returned and those who did not, except for lower psychosocial adversity in children who did not return. Thus, the 28% rate of NPD may be an overestimate. However, even if none of the participants lost to attrition developed an NPD, the rate would still be high (17/79; 22%). Fourth, the image analysis did not use volumetric measurements or diffusion tensor imaging, which could have more clearly delineated imaging correlates of NPD. Fifth, although our hypotheses did not include an orthopedic injury (OI) comparison group, such a group could serve as a control for NPD in children predisposed to and exposed to injuries.
The strengths of this study should also be appreciated. This is the largest prospective psychiatric interview study of a consecutively-hospitalized population of pediatric MTBI. The breadth and depth of assessments were extensive and included interview assessments of psychopathology, adaptive functioning, and family psychiatric history, in addition to rating scales representing injury and other psychosocial risk factors for NPD, as well as assessments of several domains of neurocognitive functioning. Furthermore, expert neuroradiologists performed the lesion analysis.