Mania Following Traumatic Brain Injury: A Systematic Review

A literature search was performed by using MEDLINE, EMBASE, and PsycINFO to identify all studies describing mania following TBI published through July 15, 2021. Search terms relating to mania, hypomania, or bipolar disorder were combined with terms relating to head injury or brain injury. The full search strategy is summarized in Figure S1 in the online supplement. Reference lists of the included studies were manually searched to identify additional studies. The search methodology was in accordance with PRISMA guidelines (27) (see Figure S2 in the online supplement).

All published studies of mania following TBI among adult patients were included, regardless of temporality and speculation on causality. Case studies needed to include individual primary data and have sufficient information on clinical presentation of at least the first manic episode following TBI or treatment of manic symptoms. Only studies published from 1980 onward were included, because this was the year DSM-III was published, which had criteria with improved reliability and wider acceptability than earlier diagnostic criteria (28). All patients for whom the authors described the presentation as mania or hypomania were included; we did not retrospectively apply current diagnostic criteria to exclude patients, because the criteria used at the time of publication may have differed from the current criteria. Information about individuals who did not acquire TBI in adulthood (≥18 years) was excluded. Other exclusion criteria were studies in which patients had a personal history of mania or bipolar affective disorder prior to sustaining TBI, studies in which manic symptoms were due to delirium or clearly precipitated by medication without speculation as to another cause, and studies published in a language other than English.

Studies identified by the search strategy underwent title and abstract screening by two independent reviewers (A.L., S.L., or M.W.). Conflicts were discussed between two reviewers and resolved. If consensus could not be reached, a third reviewer was consulted to reach a final decision. Studies that met eligibility criteria on title and abstract review and studies that could not be excluded on the basis of information available in the abstract were included in the full-text screening process. Full-text screening was also independently conducted by two reviewers, with a third reviewer consulted to resolve any disagreements. In addition, a manual search of the reference lists in the included studies was conducted to identify additional studies that fulfilled eligibility criteria.

The extracted data included sociodemographic characteristics of patients (age at first manic episode, sex, personal history of psychiatric disorders prior to TBI, family history of psychiatric disorders, and comorbid conditions); country of publication; details of TBI (etiology of TBI, duration of loss of consciousness or coma, lesion location, presence of cognitive impairment, and electroencephalogram [EEG] and neuroimaging findings); details of mania (time interval between TBI and mania onset, initial manic presentation, diagnostic criteria used, clinical course following TBI, and reported symptomatology); and treatment and response to treatment. One reviewer (A.L.) extracted the data from the included case studies. A second reviewer (S.L.) reviewed the data extraction files to verify accuracy. Any discrepancies in the data were resolved through discussion with a clinical neuropsychiatrist and expert in TBI (M.W.). The extracted data are presented in Tables S1 and S2 in the online supplement.

The quality of included studies was assessed by using the Joanna Briggs Institute Critical Appraisal tool for case reports (29), because only case reports and individual cases belonging to case series were found. This tool contained eight questions for evaluation:

Possible ratings for these questions were yes, no, or unclear.

The quality of each study was not determined by cutoff values or scores, because these analyses would add little value to the present review. Case reports and case series have inherent limitations and are already considered to have low quality of evidence. The purpose of this review was to explore future areas of focus on the basis of the data we obtained from these case studies, not to draw conclusions—hence, all case studies meeting eligibility criteria were included regardless of quality or missing data. The quality assessment was performed to assist readers in further interpreting our findings by highlighting the strengths and limitations of each study. The summary of these results (see Figure S3 in the online supplement) provides an overview of whether certain information was reported and whether the information was clearly described. A.L. completed the assessments, which were reviewed by M.W. Any disagreements between the reviewers were resolved by discussion, with input from S.L. when necessary.

Sociodemographic characteristics of the patients are summarized in Table 1. Patients originated from 12 countries, with the majority from the United States (N=27, 54%). All patients sustained TBI in adulthood (≥18 years). The mean±SD age at mania onset was 39.1±14.3 years, with ages ranging from 18 to 70 years. The majority were male (72%). Two patients reported a personal history of psychiatric disorder: one patient had childhood attention-deficit hyperactivity disorder (63), and the other experienced a depressive episode treated successfully with sertraline (48). In studies that reported family history, 95% of patients had no family history of psychiatric disorders. A personal or family history of mania or bipolar affective disorder was not reported in any of the studies. Current or previous comorbid conditions were reported in 13 patients, which included posttraumatic seizures or epilepsy (N=4) (46, 52, 58, 60), alcohol use disorder (N=4) (30, 47, 56, 63), cardiovascular disease (cardiomyopathy and arrhythmias) (N=2) (36, 47), ischemic stroke (N=1) (62), sleep apnea (N=1) (42), fetal alcohol syndrome (N=1) (32), and childhood head injury (N=1) (53). Among the patients who reported posttraumatic seizures or epilepsy, three developed focal seizures, and one developed generalized seizures, with the time interval between TBI and the initial seizure ranging from 1 week to 4 years. A history of childhood trauma, illicit drug use, asthma, migraine, thyroid disease, or obesity—comorbid conditions associated with bipolar affective disorder—was not reported in any of the studies (71).

Details relating to the reported TBI, including etiology of brain trauma, loss of consciousness and cognitive impairment, and neuroimaging findings, are summarized in Table 2. The most frequent etiology of TBI was motor vehicle accidents (N=24, 48%), followed by falls (N=13, 26%); other etiologies included explosions (N=3), assaults (N=2), and accidental head strike (N=1). Etiology was not reported for seven patients (14%). Many patients sustained TBI with loss of consciousness (N=31, 62%). However, information on whether loss of consciousness occurred was not available for 16 patients (32%). Duration of loss of consciousness or coma ranged from several seconds to 2 months. When the location of brain trauma was provided, there was an almost even distribution between the right (N=12, 38%) and left (N=13, 41%) hemispheres, and seven patients had bilateral trauma. Frontal (N=18, 62%) and temporal (N=15, 52%) lobes were most frequently affected. Among the patients with available data on cognitive impairment (N=31), the majority reported the presence of cognitive impairment (N=26, 84%), with executive functioning (N=11, 36%) and memory (N=11, 36%) being the most frequently impaired domains. Additionally, 11 patients (42%) had abnormal EEG findings in the form of slowing or the presence of both spike and slow waves. Lesion pathology was diverse; the most common seen on neuroimaging was hemorrhage or hematoma (N=17, 41%), followed by cerebral atrophy (N=6, 14%), encephalomalacia (N=4, 10%), and diffuse axonal injury (N=2, 5%).

Time intervals between TBI and mania onset were clearly described for 46 patients (Table 3). These latency periods ranged from within a week to 31 years after TBI. One-half of the patients showed a latency period of 2 months or more. The majority of patients (N=34, 74%) experienced mania in the first year after TBI. Of these, 76% (N=26) had a reported latency period under 3 months. In eight patients (17%), a delayed onset of more than 2 years was reported.

Other details pertaining to mania and its clinical presentation are also presented in Table 3. The most frequent initial presentation was mania (N=38, 76%), followed by hypomania (N=6, 12%) and mania with psychosis (N=4, 8%); for two patients, it was unclear whether mania or hypomania was manifested. Only one-half of case studies reported the diagnostic criteria used. Of these, 17 used DSM criteria, two used ICD criteria, and 10 used the Krauthammer and Klerman criteria for secondary mania (72). Clinical course varied across patients: 38% progressed to a disorder with recurrent manic and depressive episodes (N=18), 36% experienced a single manic episode (N=17), 13% had recurrent manic episodes without depressive episodes (N=6), and 13% experienced rapid cycling (N=6). Other possible precipitants for the induction of mania following TBI for five patients included antidepressants (42, 43, 50, 61, 63) and chronic sleep deprivation (42).

Individual data on manic symptoms were reported for 45 patients (for symptoms for each patient, see Table S4 in the online supplement). Of these, increased talkativeness was reported for almost all patients (N=43, 96%) during at least one of the manic episodes after TBI. Elevated or irritable mood was reported for most patients (N=42, 93%), with a higher proportion exhibiting elevated mood (N=29, 64%) over irritable mood (N=26, 56%). Other frequent symptoms included increased goal-directed activity or psychomotor agitation (N=33, 73%), decreased need for sleep (N=32, 71%), and inflated self-esteem or grandiosity (N=23, 51%). Psychotic features during at least one of the patient’s manic episodes were reported for 14 patients and included disordered thoughts (N=6, 13%), persecutory ideation (N=6, 13%), grandiose delusions (N=4, 9%), hallucinations (N=4, 9%), and persecutory delusions (N=3, 7%). Thirteen patients exhibited aggressive or threatening behavior (29%), and nine patients were disinhibited (20%).

Many treatments to improve manic symptoms or achieve remission were identified. Among the 40 patients receiving treatment, 29 (73%) received mood stabilizers (lithium, valproate, or carbamazepine), 14 (35%) received first-generation antipsychotics (haloperidol, chlorpromazine, thioridazine, loxapine, perphenazine, or trifluoperazine), 10 (25%) received second-generation antipsychotics (quetiapine or olanzapine), three (8%) received benzodiazepines (clonazepam or lorazepam), three (8%) received beta blockers (propranolol or pindolol), two (5%) received electroconvulsive therapy, and one (3%) received clonidine. There was minimal reporting on other forms of management, such as psychological interventions. Manic symptoms of four patients resolved without treatment.

It was difficult to determine whether remission was achieved because of insufficient follow-up in many case studies. Many patients also had a prolonged period of receiving various treatments before an efficacious regimen was found, with a few patients requiring triple therapy (59, 70) or quadruple therapy (50) to treat manic symptoms. Electroconvulsive therapy was highly effective among patients who demonstrated an inadequate response to pharmacological treatment (62, 65). Comprehensive information on the various treatments administered to each patient, including specific dosages, is presented in Table S5 in the online supplement.

Patients were more frequently male, <50 years old, and without personal or family history of psychiatric conditions. The wide age range (18–70 years) suggests that mania following TBI may occur at any age and is not concentrated in younger age groups like primary bipolar affective disorder (73). The majority of patients presented with either mania or hypomania, without psychosis. A small number of case studies reported precipitating factors that may have contributed to the development of mania, most notably antidepressants, as reported for five patients. Other potential contributing factors for developing mania following TBI may have included the sensitizing role of TBI (74); the presence of other mood destabilization factors, such as sleep deprivation (75); or an underlying genetic vulnerability to bipolar affective disorder (71) that had not yet been uncovered among these patients.

All classic symptoms of mania, as described in DSM-5 (16), were present across the case studies that reported symptomatology. Symptoms of psychosis, as well as behavioral dyscontrol, such as aggressiveness and disinhibition, were also present. A lower rate of aggression or threatening behavior was seen in the present study (29%), compared with rates observed in the largest case series published by Shukla et al. (25), where assaultive behavior (70%) was frequently observed among the 20 patients who developed mania following closed head trauma. Our sample also had a lower rate of irritability (56%), compared with the 85% rate found in that case series (25). Ascertainment bias may have contributed to these discrepancies, because Shukla et al. recruited patients with primary neurological diagnoses and associated psychiatric symptoms who had been referred to neuropsychiatric outpatient clinics at two hospital centers in New York. Aggression following TBI has been associated with presence of major depression, frontal lobe lesions, poor premorbid social functioning, and a history of alcohol and substance abuse (76), which may have been more prevalent among their patients. The lack of standardized definitions of what constitutes “aggressive behavior” (77) may contribute to variability in reporting; symptoms of aggression and irritability may have been underreported in the patients we reviewed. Aside from these differences, symptomatology profiles in the patients we reviewed were similar to those found in other case series (21, 25) and, in turn, similar to those of primary mania.

Similar to our findings, a full spectrum of manic disorders has been reported in the literature, including brief self-limiting manic episodes, prolonged periods of bipolar or schizoaffective disorder marked by recurrent episodes of mood disturbance, and rapid-cycling variants (25, 78). Clinical course may be complicated by other psychiatric symptoms or conditions following TBI, requiring frequent reevaluation and reformulation over time (46, 49). Patients may follow a peculiar clinical course, such as recurrence of mania in a seasonal pattern (33). Other patients experience poorly controlled mania despite multiple medication trials (44, 47, 56, 66, 70). This was recognized by Satzer and Bond (79), who proposed that secondary mania may be indicated by an unusual illness course, such as a single manic episode with no subsequent mood symptoms (including depression), unremitting mania, or poor response to antimanic treatments. However, we found that recurrent mood episodes comprised the most common clinical course, which will have treatment implications (see the Special Considerations for Treatment in This Population section below).

Clinical outcomes were difficult to determine, because many of the case studies did not provide adequate follow-up or clear information about remission. It is known that major depression following TBI is associated with poorer health-related quality of life (80). On the basis of descriptions provided, functional impairment remained for many patients despite the resolution of mania, indicating that the course of associated manic symptoms may be independent of cognitive impairment and recovery (44, 47, 49, 56). Many patients were unable to return to work due to persisting emotional deficits and disruptive behavioral symptoms (44, 49, 57, 67), highlighting the profound long-standing psychosocial impact that TBI can have on an individual.

Additional brain injury, unrelated to TBI or perhaps because of TBI, makes determining causality more difficult. One patient in our review had a reported history of stroke prior to both TBI and mania onset, while four patients experienced posttraumatic seizures or epilepsy. Patients with secondary causes of mania, such as that occurring after stroke and as a component of epilepsy, were not included in this review, but these causes of mania are important and can be conceptualized as sequelae of TBI (79, 81, 82). Similar to the rates of developing mania following TBI, the estimated prevalence rate of poststroke mania is less than 2% (83). Although comorbid bipolar symptoms were reported among 12% of patients with epilepsy in a U.S. community-based study (84), when confounding variables are accounted for in other studies, the rate of postepilepsy mania is almost equal to the rates of bipolar affective disorder in the general population (82).

The relationship between certain aspects of the brain injury, such as severity and location, and the development of secondary mania has not been delineated. In our study, severity could not be accurately determined retrospectively as a result of incomplete data on the Glasgow Coma Scale and other relevant descriptors. However, a spectrum of severity (mild, moderate, and severe) can be assumed from the variation in Glasgow Coma Scale scores at presentation and the duration of loss of consciousness and posttraumatic amnesia reported among the reviewed patients. Of the patients with reported loss of consciousness and posttraumatic amnesia, the majority acquired some degree of cognitive impairment and abnormal neuroimaging findings. In a case series by Jorge et al. (21), development of mania did not appear to have positive associations with the severity of brain injury, the degree of physical impairment, or the degree of cognitive impairment. Instead, the literature suggests that lesion location may be a more important etiological factor (33, 74, 85).

Our study found an almost even distribution of damage between the right (38%) and left (41%) hemispheres, with a smaller proportion of patients showing bilateral distribution (22%), reflecting heterogeneity in TBI lesion locations. In contrast, a right-sided predominance of lesions (61%) among 201 patients with mania secondary to vascular, tumor, TBI, and other nonvascular etiologies was reported in a recent review (86). However, only 24 individuals (12%) in that sample sustained TBI. The most frequently reported lesion areas in our study (frontal and temporal lobes) are supported by findings from a study by Cotovio et al. (87), in which a mania network was derived from brain lesions reported in the literature. Lesion locations associated with mania were found to be heterogenous. However, the investigators showed a unique pattern of functional connectivity to the right orbitofrontal cortex, right inferior temporal gyrus, and right frontal pole (87). The lack of laterality in our findings does not support this finding however. Satzer and Bond (79) similarly reported frontal, temporal, and subcortical limbic brain regions as the most common lesion locations. Moreover, brain activity appears to shift toward the left hemisphere, resulting in hyperactivity of left-hemisphere reward-processing areas and hypoactivity of right-hemisphere limbic brain regions, which aligns with current models of bipolar affective disorder (79). Due to the heterogeneity in pathology, with TBI often being a diffuse or multifocal injury, conclusions regarding the relationship between laterality and secondary mania in TBI cannot be made.

Diagnostic criteria for secondary mania have been inconsistently applied, and different diagnostic criteria have been used—none of these criteria have confirmed diagnostic validity or reliability (88). The diagnostic criteria used in our case studies included ICD-10 (33, 51), DSM-III (57, 65, 66, 69, 70), DSM-III-R (52, 56), DSM-IV (37, 55), DSM-IV-TR (38), DSM-5 (36, 48), and Krauthammer and Klerman criteria (31, 35, 39, 52, 56, 62, 64). No uniform diagnostic criteria exist for mania secondary to TBI. This may be because clearly attributing the development of mania to TBI presents a variety of challenges. In DSM-5, guidance on determining causality is vague and is at the discretion of the clinician, taking temporality, as well as plausibility, of a causal relationship into consideration.

As a result of the lack of clear guidance, it has been proposed that a close temporal relationship between TBI and mania onset, in the absence of other etiology, may be the best indicator of causality (88). However, mania has been observed up to 31 years after TBI (46), with 26% of case studies reporting a time interval of more than 1 year. In their case series on 20 patients with mania following TBI, Shukla et al. (25) reported a mean latency of 2.8±3.4 years. Moreover, Chi et al. (8) found that although an increase in the risk for bipolar affective disorder in the first year following TBI was observed, the risk intensified in the 2 to 4 years after TBI and eventually decreased after 5 years. A delayed increase in the risk for bipolar affective disorder (>1 year) was also found by Orlovska et al. (12). Furthermore, Sachdev et al. (89) reported a mean latency of 4.6±4.6 years among patients developing schizophrenia-like psychosis following TBI, with latency durations that ranged from 2 weeks to 17 years. Even if causality cannot be proven in these patients showing long-latency durations, TBI may still be a contributing factor to mania development due to its pathophysiological effects on the brain (90).

Neuropsychiatric features add complexity in this population with mania secondary to TBI, and because presenting symptoms appear phenotypically similar to those of primary mania, using symptomatology alone to distinguish between primary and secondary mania may be challenging (79). Additionally, patients with brain injury frequently exhibit impulsivity, affective lability, and behavioral irritability, making it difficult to distinguish from mania (91). Symptoms of behavioral dyscontrol, such as aggression and disinhibition, are commonly associated with moderate-severe TBI and may cooccur with the classic symptoms of mania following TBI or may be manifestations of the mania itself (92). Personality changes, characterized by the aforementioned symptoms, can be difficult to differentiate from mania or hypomania (49) and frequently fulfill the criteria for secondary mania (93). However, a persistent and continuous pattern of these symptoms and less pervasive alteration of mood may suggest a personality change due to TBI rather than mania (94). These overlapping features make reaching a definitive diagnosis more difficult.

To our knowledge, no randomized controlled trials on pharmacological treatments for mania following TBI have been conducted. There is still insufficient evidence to develop formal treatment guidelines for patients affected by TBI (95). A variety of monotherapies and combinations of therapies have been successful in treating symptoms. However, it can be difficult to discern which specific medications were responsible when combinations are used. This is further complicated by inadequate follow-up data to confirm whether true remission has been achieved.

Mood stabilizers (lithium, valproate, and carbamazepine) were used successfully with 73% of the patients we reviewed. Valproate has the most evidence available to support its use among patients with TBI, demonstrating good tolerability and good efficacy, in addition to its rapid onset (96). However, such research has been centered on the treatment of behavioral symptoms rather than mood symptoms. Studies have suggested that valproate has less adverse effects on cognition and motor performance than carbamazepine and lithium, particularly in the early postinjury period (97, 98). In contrast, emerging evidence from preclinical studies found that lithium may provide neuroprotective and neuroregenerative effects after injury, aiding recovery (99). Lithium may, however, lower seizure threshold, which can be significant given that patients with TBI may be vulnerable to posttraumatic seizures (25). Given that lithium may be associated with increased seizure risk and that valproate is well tolerated among patients with TBI, it could be suggested that valproate is a first-line choice among mood stabilizers for TBI.

The second most widely used drug class was antipsychotics. Despite the success of haloperidol in six patients, second-generation antipsychotics are preferred, because patients with TBI are more susceptible to adverse effects, such as extrapyramidal side effects (79), as demonstrated in two patients (50, 51). It has also been suggested that antipsychotics may interfere with neuronal recovery after brain injury (100–102). Of the second-generation antipsychotics, olanzapine and quetiapine were most frequently used to manage manic symptoms in the reviewed studies. Olanzapine may be one of the most efficacious second-generation antipsychotics in treating acute mania and is at least as effective as lithium or valproate (103). However, its use may be limited for some patients due to side effects of weight gain, development of diabetes mellitus, and hyperlipidemia (104). For these patients, aripiprazole may be a novel alternative; there is a published case study of its successful use for an adult presenting with mania after sustaining TBI in childhood (105).

In addition to the importance of following the start low and go slow principle, there are other considerations when initiating and maintaining treatment in this patient population. Mania emerged within the first year after TBI in 74% of the patients we reviewed, which is a percentage similar to that for other psychiatric disorders (22), hence this early period is critical. The initiation of psychotropic medication administration was associated with increased odds of late mortality among 1-year survivors who were treated in an intensive care unit, even after adjusting for TBI severity (106). However, the causes of death among these patients were not available. Moreover, administration of psychotropic medications was used as a proxy for psychiatric comorbidity in that study. Hence, the findings could indicate that psychiatric illness itself, rather than psychotropic medication, is more causally related to mortality, which further emphasizes the importance of early identification of psychiatric comorbid conditions among TBI survivors. Therefore, early comprehensive assessment, including the timing and pattern of symptoms, full psychiatric history, and exploration of other behavioral disturbances, should always be conducted to reduce misdiagnoses and enhance treatment decision making.

Once the diagnosis of mania has been made, combination therapy may be required, especially for patients with severe mania. Although most clinicians will likely start with a monotherapy, early escalation to a second agent may be required for patients who show limited early response. Almost one-half of the patients in our review required two or more medications to adequately control manic symptoms, and because mania after TBI may follow a complicated clinical course, having an efficacious treatment regimen early may prevent mania recurrence. This notion was supported by Glue and Herbison (107) in their network meta-analysis of pharmacological treatments for acute mania, which showed that combination treatment with a mood stabilizer and second-generation antipsychotic was significantly more effective than monotherapy. Moreover, in addition to their antimanic effect, both second-generation antipsychotics and mood stabilizers may have added benefit in reducing behavioral dyscontrol symptoms; carbamazepine, valproic acid, quetiapine, and olanzapine may reduce irritability, agitation, aggressiveness, or insomnia (108–110).

Furthermore, approximately one-half (51%) of the patients in the case studies we reviewed experienced recurrent manic and depressive episodes, including those with rapid cycling, highlighting the need for greater vigilance when selecting treatments. Consistent with our study, individuals who experience recurrent mood episodes or who are later diagnosed with bipolar affective disorder typically report a depressive episode first (111). Because antidepressants may have precipitated or contributed to the development of mania in five of the reviewed patients, careful consideration should be taken when choosing which antidepressant to prescribe. Serotonin and noradrenaline reuptake inhibitors, namely, venlafaxine, and tricyclic antidepressants seem to have higher switch rates than selective serotonin reuptake inhibitors (112–116). Moreover, it is advisable that mood stabilizers be added in the treatment of bipolar affective disorder, because antidepressant monotherapy should be avoided due to increased risk of manic switching (117). The need for close monitoring for signs of antidepressant-induced mania is necessary, especially among patients with brain injury, for whom comorbid conditions or vulnerabilities may confer greater risk.

This study contains several limitations and potential biases. The inclusion of case reports and case series alone renders the sample not representative of the population and does not allow for statistical conclusions. There is potential for publication bias, because the vast majority of case studies were published between the late 1980s and early 2000s and reflected the prevailing ideas (i.e., on lesion laterality and treatments) during that time. Differences in reporting resulted in missing data points. Differing diagnostic criteria further limited comparability of the data. Given that approaches to diagnosis have changed over time, some patients may not have fulfilled present-day criteria for secondary mania. Additionally, a level of interpretability was occasionally used when extracting data for symptoms and clinical course due to vague descriptions. Furthermore, the presence of other risk factors or comorbid conditions associated with mania may not have been reported when they were in fact present. Lastly, without a search of the gray literature and with the exclusion of non-English case studies, this review is not exhaustive. Non-English studies may have contained information about cultural differences in clinical presentation and treatment not considered in this study.