- Top of page
Purpose: To investigate the relationship between traumatic brain injury (TBI) and psychogenic nonepileptic seizures (PNES). We hypothesized that PNES with TBI would be associated with more psychiatric comorbidities and disability than PNES without TBI.
Methods: In this cross-sectional study comparing patients with PNES with TBI to patients with PNES without TBI, medical records from 255 consecutive patients with electroencephalography (EEG)-confirmed PNES were reviewed to assess variables including demographic, head injury, neurologic, psychiatry, social variables, and quality of life and symptoms scales. Parametric, analysis of covariance (ANCOVA), and logistic regression analyses were performed, to compare psychiatric and function variables between the two study groups while controlling for age and sex.
Key Findings: Of the 92 patients with PNES who fulfilled inclusion/exclusion criteria, 41 (44.6%) had a history of TBI. Of the 41 patients with TBI, 30 (73%) met criteria for mild TBI (mTBI). Patients with TBI had more mood disorder diagnoses, were more likely to receive disability, and had lower global functioning than non-TBI patients with PNES, after adjusting for age and sex. Patients with TBI and PNES had significantly increased odds for having major depression, behavioral impulsivity, posttraumatic stress disorder diagnosis, and a trauma/abuse history.
Significance: TBI is a significant risk factor in patients with PNES, being associated with increased psychiatric diagnostic comorbidity, symptoms severity, poorer functioning, and increased disability. This study reveals the importance of identifying and addressing the impact of TBI in patients with seizure disorders. Addressing the sequelae of TBI in PNES may be a target to improve functioning.
One potential comorbid factor that has not been fully investigated in PNES is traumatic brain injury (TBI). Patients with TBI frequently have comorbid psychiatric disorders both before and after the TBI, the most common being depression and personality disorders (Hibbard et al., 2000; Koponen et al., 2002; Jorge et al., 2004; Kim et al., 2007; Nicholl & LaFrance, 2009; Whelan-Goodinson et al., 2009). Given the psychiatric, cognitive, and seizure disorder comorbidities associated with TBI, TBI may also adversely impact PNES outcomes.
Preliminary investigations have assessed TBI as a risk factor for PNES, but have not address psychiatric and functioning outcome measures (Westbrook et al., 1998; Pakalnis & Paolicchi, 2000; Hudak et al., 2004). The disorders most frequently reported after a TBI or non–TBI-related trauma were depression, anxiety, posttraumatic stress disorder (PTSD), and various phobia or panic disorders (Bryant et al., 2010). The purpose of this study was to examine the psychiatric comorbidities and functioning among patients with TBI and PNES. We hypothesized that TBI in patients with PNES (P + T) would be associated with more psychiatric comorbidities and more disability than in patients with PNES only (P − T).
- Top of page
TBI appears to contribute to significant comorbidity and worsened functioning in patients with PNES. Our study used formal psychiatric evaluation and symptom, quality of life, and functioning outcome measures to assess the relationship between PNES, TBI, and comorbidities. Combinations of mood, anxiety (PTSD), history of abuse, and characterologic traits may be common components of the PNES phenotype, which may be exacerbated by TBI.
Within the PNES sample, we found a prevalence of TBI of 44.6%. This is the first study to the authors’ knowledge of an assessment of TBI and PNES with psychiatric assessment and outcome scales. None of the prior studies describing PNES and TBI described below incorporated formal outcome measures of psychiatric symptoms, quality of life or measures or global functioning, and only one study incorporated a psychiatric evaluation (Westbrook et al., 1998). A study of 102 patients with PNES found that 33 (32%) had a history of TBI, 28 of which were minimal or mild (Westbrook et al., 1998). A study of patients in the Epilepsy Monitoring Unit (EMU) with moderate or severe TBI identified PNES in 34 (32.7%) of 104 patients who had a diagnostic EEG (Hudak et al., 2004). In a pediatric study, 7 (44%) of 16 children diagnosed with PNES were found to have prior history of head injury (Pakalnis & Paolicchi, 2000). In another study, 29 (58%) of 50 veterans thought the cause of their seizures was due to TBI, although the occurrence or extent of TBI was not verified (Salinsky et al., 2011). Our results reveal percentages of patients with PNES and TBI roughly similar to the results of adult and pediatric studies (Westbrook et al., 1998; Pakalnis & Paolicchi, 2000; Hudak et al., 2004), despite different sampling methods. The slightly higher percentage may be explained by the differences in TBI criteria. Our study used the criteria based on the WHO and CDC standards, similar to the Pakalnis criteria, whereas the Westbrook study used the Head Injury Severity Scale and Glasgow Coma Scale scores, and the Hudak study did not include mild TBI (Westbrook et al., 1998; Pakalnis & Paolicchi, 2000). Using the same criteria may result in percentages similar to those of the Westbrook study.
Using formal psychiatric assessment, our study showed that the P + T group was more likely to have MDD, higher numbers of mood disorders, and more cluster B disorders or traits. This is consistent with the rates of comorbidities found in patients with TBI as compared with healthy controls. Patients with TBI frequently have comorbid psychiatric disorders. In a 30-year follow-up study of 60 patients with TBI, 48.3% had an axis I disorder that followed their TBI, the most common of which was MDD. Personality disorders (axis II) were seen in 30.0% of the TBI population (Koponen et al., 2002). Another study found two of the most common disorders post-TBI to be MDD (45%) and anxiety (38%) (Whelan-Goodinson et al., 2009). Patients with TBI often have increased impulsivity (cluster B disorders and traits), PTSD, and MDD (Hibbard et al., 2000; Jorge et al., 2004; Kim et al., 2007; Nicholl & LaFrance, 2009). Patients with PNES also have more mood disorders than control groups (Szaflarski et al., 2003). In one survey of 56 patients with PNES, >50% had diagnoses of depression and almost 40% had a history of suicidal ideation (Ettinger et al., 1999).
Comorbidities such as depression correlate with reduced quality of life (QoL) in patients with PNES (Testa et al., 2007; LaFrance & Syc, 2009). Both the P + T and P − T groups in our study had low QOLIE-31 scores (poorer QoL), although P + T had lower GAF scores (lower global functioning). Patients with P + T had a higher frequency of seizures than those without TBI, but this difference was not statistically significant after controlling for age and sex. A diagnosis of depression and seizure frequency with the TBI may be contributors to the lower functioning and higher disability in the P + T sample.
Depression symptom scores (BDI-II) were in the moderately depressed range for both groups, with a higher rate of MDD diagnosis in the P + T sample. This finding suggests that the overall severity of depressive symptoms between the two samples is similar. This observation may reflect that depressive symptoms can present differently in neuropsychiatric populations, such as occurs in patients with epilepsy (Gilliam et al., 2006) or with TBI (Nicholl & LaFrance, 2009) than in classic unipolar depression. The overlap of somatic symptoms found in TBI and those in neurovegetative signs in major depressive disorder may account for the difference in the presentation of depression in patients with or without TBI.
Our study found patients in the P + T group had higher rates of abuse history. A history of physical or sexual abuse is reported in 30–80% of patients with PNES (Alper et al., 1993; Bowman & Markand, 1996). Patients with PNES reported more severe sexual abuse (Alper et al., 1993) than control subjects with epilepsy. Physical abuse rates in particular were higher for patients in the P + T group, suggesting that patients in that group were exposed to more violent environments, which could have resulted in physical trauma such as head injury. The combination of PNES and TBI resulted in a 2.75 odds increase of having PTSD, and a 3.35 odds increase of having a history of trauma/abuse, underscoring the importance of the “double hit” of emotional and physical traumatic experiences found in the PNES population. This study shows that TBI and PNES are significantly associated with the cluster of MDD diagnosis, PTSD diagnosis, cluster B personality, and/or trauma/abuse history, which could have an impact on functioning.
Both groups of patients with PNES had similarly high levels of unemployment; however, twice as many P + T patients were receiving disability. Increased disability within the P + T sample may reflect a greater severity of symptoms among that group or may reflect the criteria for granting disability. The frequency of disability in this study reflects only current status of disability and does not indicate how many subjects were in the process of applying for disability or were initially rejected. The disability rates in our study are consistent with findings in other studies where up to 60% of patients with PNES receive disability from the government or are financially supported by family (Krawetz et al., 2001). GAF score, a clinical correlate of disability, showed lower average scores (lower functioning) in patients with P + T, supporting the hypothesis that those with both conditions have more significant disability than those with only PNES.
The differences in demographic indicators between the groups may be a result of the interaction between known PNES and TBI demographics. TBI populations are more predominantly male (Bruns & Hauser, 2003), and PNES populations are predominantly female. In our study, males with PNES were more likely to have a history of TBI than females. Although there was a statistically significant difference in current age of patients between the two groups, there was no clinical difference, and the age of PNES onset was not significantly different. After controlling for age and sex, several variables remained significant. This suggests that the differences between TBI and non-TBI populations do not merely reflect the demographic differences, although age was a significant covariate in MDD and impulsivity. Controlling for sex reveals that the differences in psychiatric symptoms and history between the P + T group and the P − T group do not merely reflect the known sex differences (female predominance) in PNES. Due to this size constraint, comparisons between the subgroups of males were not possible.
In our study, 30% had abnormal magnetic resonance imaging (MRI) studies, which is similar to results where 27% of patients with PNES without comorbid epilepsy had abnormal MRI reports (Reuber et al., 2002). Only one MRI abnormality in our study was directly related to TBI. The most common MRI abnormality was periventricular white matter disease. The trend of higher rates of abnormality in subjects with TBI (50% vs. 30%, p = 0.06) may reflect demographic differences or a potential risk factor associated with combinations of TBI and PNES. Male subjects had a greater increase in likelihood of abnormal MRI results if they sustained a TBI than female subjects did. This point is worth further examination given the limited understanding of the neuropathobiology of PNES.
Our retrospective study design does not and cannot prove causation (i.e., that TBI causes PNES); however, our study adds to the literature on contributors to psychiatric comorbidity with TBI. Regarding “causation,” TBI may be one link in the temporal chain of events in the constitutional makeup of some patients with PNES. Our study significantly adds to previous investigations of TBI as a comorbid factor in PNES through comparison of factors previously shown to be significant in either disorder, separately. In the TBI and posttraumatic epilepsy literature, a significant association of TBI and development of posttraumatic epilepsy is also demonstrated (Christensen et al., 2009). When a patient presents with seizures after TBI, frequently the seizures are assumed to be epilepsy and the patient is treated for epilepsy (e.g., numerous antiepileptic drugs [AEDs], repeated laboratory monitoring, and EEG studies, and so on). Even though time from seizure onset to PNES diagnosis was not statistically significant, the mean times to diagnosis were 4.9 years for the no TBI group, versus 7.3 years for the with TBI group, which is clinically significant. The delay in accurate diagnosis results in inappropriate treatment for presumed epilepsy in many cases, while patients are not receiving appropriate for PNES (LaFrance & Benbadis, 2006).
In our cohort of patients with PNES and no epilepsy, those with TBI had more comorbid disorders and more severe symptoms than those without TBI. This may reflect the comorbidities present in TBI, preexisting comorbidities that increased the risk of TBI, or the combined effect of TBI and PNES. Of particular relevance, our finding of a relationship between mild TBI, PTSD, and PNES in a civilian population may be similar to that of the veteran population with PNES (Dworetzky et al., 2005; Salinsky et al., 2011). A prospective study of head injury and its neuropsychiatric sequelae or a cohort study of civilian and veteran participants with TBI and PNES may provide insight into potential causation of PNES.
Limitations of the study included potential for recall bias by the patients in this retrospective chart review. TBI was determined with thorough patient history documented in neuropsychiatric evaluation administered by physicians. Neurology notes and MRI reports were reviewed when available to supplement diagnostic assessment. Limited information was available for some head injury histories reported by patients, and sequelae were sometimes difficult to specifically attribute to TBI. Given that the purpose of this study is to examine the potential for TBI as a risk factor in PNES, a broader set of criteria were used for TBI sequelae. Patients with head injury to which they attributed their cognitive symptoms, those with neurologic symptoms occurring soon after TBI (such as onset of seizures 5 days post-TBI), and those diagnosed with concussion by physician were included in the P + T sample. In addition, some patients reported TBI occurring after NES with their events (e.g., hitting one’s head with a seizure). Selection bias was possible given that patients were from a tertiary care center rather than the general population, which may bias data on severity of symptoms or comorbidities. These factors may limit generalizability of the findings to TBI without PNES. Further prospective investigations of head injury in populations with PNES are warranted to assess TBI’s potential for a causal relationship to PNES.
This study underscores the importance of assessing not only the seizures, but also the somatic and psychiatric comorbidities in patients with PNES. Patients with PNES commonly describe a mixture of problems with mood, anxiety, insomnia, cognition, and pain symptoms. Treating comorbid disorders in PNES may reduce PNES frequency (LaFrance & Devinsky, 2002; LaFrance et al., 2010). Pilot cognitive behavioral therapy trials indicate that cognitive behavioral therapy (CBT) may be effective in reducing seizure frequency, (LaFrance et al., 2009; Goldstein et al., 2010) along with depression, anxiety, and other somatic symptoms while improving QoL (LaFrance et al., 2009). Incorporating treatment for the cognitive, somatic, and psychosocial complaints commonly found in patients with TBI may be important targets for a significant population with PNES and TBI.
In summary, in our study, 45% of patients with PNES had a history of TBI, and one-third had mTBI. The presence of a TBI in patients with PNES plays an important part of their clinical picture. TBI appears to be a significant risk factor in patients, with PNES being associated with increased psychiatric comorbidity, symptoms severity, poorer functioning, and increased disability. Further exploration of the interaction between these two common disorders could help with development of treatments, as new pharmacologic and rehabilitative options become available.