- Email: [email protected]
Psychiatric comorbidity in veterans with psychogenic seizures
Epilepsy & Behavior 25 (2012) 345–349
Contents lists available at SciVerse ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Psychiatric comorbidity in veterans with psychogenic seizures Martin Salinsky a, b,⁎, Collette Evrard a, Daniel Storzbach a, Mary Jo Pugh c a b c
Portland Veterans Affairs Medical Center, Portland, OR, USA Oregon Health & Science University, Portland, OR, USA San Antonio Veterans Affairs Medical Center, San Antonio, TX, USA
a r t i c l e
i n f o
Article history: Received 24 April 2012 Revised 9 July 2012 Accepted 14 July 2012 Available online 24 October 2012 Keywords: Epilepsy Psychogenic seizures Veterans Post-traumatic stress disorder
a b s t r a c t Psychogenic non-epileptic seizures (PNES) are frequently encountered in epilepsy monitoring units (EMU) at Veterans Affairs Medical Centers (VAMCs) and cause significant long-term disability. An understanding of psychiatric factors associated with PNES could aid in earlier diagnosis and treatment. We studied 50 consecutive veterans diagnosed with PNES and 37 veterans diagnosed with epileptic seizures (ES), evaluated at a VAMC EMU. We reviewed all available mental health evaluations prior to EMU evaluation. Univariate comparisons included axis I diagnoses, axis II diagnoses, and psychiatric hospitalizations. Predictive models of seizure classification were evaluated by logistic regression. A diagnosis of post-traumatic stress disorder (PTSD) preceded the diagnosis of PNES in 58% of patients and the diagnosis of ES in 13.5% (pb 0.001). On logistic regression, PTSD was the only significant psychiatric diagnosis (odds ratio 9.2). Major depression and alcohol abuse were common diagnoses but did not differentiate PNES and ES groups. Published by Elsevier Inc.
1. Introduction Psychogenic non-epileptic seizures (PNES) are clinical events that resemble epileptic seizures (ES) but that are not associated with abnormal brain electrical discharges or other physiological problems. These neuropsychiatric events are common, accounting for 20–50% of epilepsy monitoring unit (EMU) discharge diagnoses [1–5]. Most PNES patients are mistakenly thought to have medication‐resistant epilepsy and are chronically treated with one or more antiepileptic drugs (AEDs), while the psychological nature of the disorder remains hidden [3,6–9]. Short‐ and long-term disabilities are common [7,10]. Psychogenic non‐epileptic seizures are also common in veterans treated at Veterans Affairs Medical Centers (VAMCs). Patients diagnosed with PNES accounted for 25% of all EMU discharge diagnoses in a recent VAMC study and were more common than patients diagnosed with ES [8]. In the same study, 50 veterans with EMU‐verified PNES (without ES) were compared with 50 civilians with PNES studied in the same EMU and by the same personnel. Veterans had a considerably longer delay from onset of symptoms to the diagnosis of PNES as compared to civilians, with a median of 5 years. In one third of cases the delay exceeded a decade. The delay in diagnosis was associated with greater cumulative exposure to AEDs. Prognosis and long-term disability may also be negatively affected by the delay in diagnosis [5,7,11].
⁎ Corresponding author at: 3181 SW Sam Jackson Park Road, CR-120, Portland, OR 97239, USA. Fax: +1 503 494 6658. E-mail address: [email protected] (M. Salinsky). 1525-5050/$ – see front matter. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.yebeh.2012.07.013
Psychogenic non‐epileptic seizures are generally believed to be symptoms of comorbid psychiatric disorders. These disorders have been comprehensively studied in civilians but not in veterans [12–15]. An understanding of the psychiatric disorders associated with PNES is essential for designing effective treatment strategies and could aid in early recognition of PNES. Comorbid psychiatric disorders in veterans with PNES are likely to differ from those of civilians due to (a) differences in the age and sex of the populations; (b) risk of traumatic brain injury (TBI) and psychological effects of combat exposure including post-traumatic stress disorder (PTSD); and (c) potential financial motivations in the VA health care system fostering illness behavior. We studied psychiatric diagnoses preceding documentation of PNES in a carefully evaluated group of veterans with PNES and a comparison group of veterans with documented ES. 2. Methods 2.1. Patients We reviewed records of all veterans admitted to the Portland VAMC EMU from January 2000 to July 2010. Details for this group are provided elsewhere [8]. We excluded patients who had a previously documented EMU evaluation at the Portland VAMC or elsewhere, except (a) when these evaluations were non-diagnostic or (b) when the diagnosis was modified by the repeat monitoring at the Portland VAMC. We eliminated 4 patients with missing records. There were 203 unique patients in the final group. The mean age at admission was 50.6 years. Eighty-seven percent were men.
346
M. Salinsky et al. / Epilepsy & Behavior 25 (2012) 345–349
2.2. Classification Each EMU admission was reviewed, and every patient was classified into one of five diagnostic categories: (1) ES only; captured ES (diagnostic criteria below) with no current or historical verification of non-epileptic seizures. (2) PNES only; captured PNES (diagnostic criteria below) with no current or historical verification of ES and no interictal epileptiform activity (IIEA). (3) ‘Mixed’ ES and PNES; captured ES and PNES or PNES and IIEA. (4) Other non-epileptic seizures; captured physiologic events confused with ES, including syncope, sleep disorders, etc. No current or historical verification of ES or PNES. (5) Nondiagnostic; no diagnostic events captured during EMU admission (including patients with subjective spells without EEG correlate, patients with IIEA and no diagnostic spells, and patients with spells that could not be confidently classified). A total of 50 patients met criteria for PNES only (mean age 47.0; 80% men), and 37 patients met criteria for ES only (mean age 50.8; 95% men). These patients comprised the study groups. The diagnosis of PNES used the following criteria: (1) Captured events during the EMU admission were typical for the patient's habitual spells and demonstrated alteration of consciousness and/ or bilateral motor activity. (2) At least two diagnostic events or one prolonged event (> 5‐minute ictal phase). We accepted one event provided that it was endorsed by a friend or family member as typical, based on video review (3 patients). (3) No epileptiform discharges or ictal rhythmic EEG changes associated with any event. (4) Events had a presentation consistent with PNES. (5) No other physiologic abnormalities that might reasonably explain the spells. (6) No IIEA. (7) No previous verifiable documentation of ES or IIEA (from EMU or other EEG studies). The diagnosis of ES required (1) captured events during the EMU admission that were typical for the patient's habitual spells; (2) at least two diagnostic events or one prolonged diagnostic event (> 5‐minute ictal phase) as mentioned above; (3) epileptiform discharges or ictal rhythmic EEG changes associated with each event; (4) events that had a presentation consistent with ES; and (5) no verifiable documentation of PNES (from prior EMU or other EEG studies). All diagnoses were made by fellowship‐ trained epileptologists and confirmed for this study by one investigator (MS).
was included in the medical records. For patients with a diagnosis of PTSD, we classified the cause as ‘civilian’, ‘military combat’, ‘military sexual trauma’, ‘military non-combat’, or ‘not specified’ based on the qualifying mental health evaluations only. Patient age at admission, sex, and duration of seizures prior to EMU diagnosis (determined from all available sources including patient/family history, paper and electronic charts) were also recorded. To simplify analysis, certain axis I diagnoses were combined. The category ‘other depression’ included all depression not categorized as major depression (including dysthymia, depression NOS, mood disorder associated with medical illness). ‘Other depression’ was only counted once even if there were multiple types of ‘other depression’ diagnosed. Patients with major depression, who were also diagnosed with ‘other depression’, were only counted as having major depression. Bipolar disorder and adjustment disorder with depressed mood were included as separate axis I categories. The category ‘other substance abuse’ included all substance abuse not categorized as alcohol‐related. Other substance abuse was only counted once (even if multiple substances were listed separately); however, patients could have diagnoses of both other substance abuse and alcohol abuse. 2.4. Statistical analysis Two-group (PNES vs. ES) univariate comparisons used either the Wilcoxon test or Fisher's Exact Test. Significance level was set at p ≤ 0.05 (two-tailed). Predictive models of seizure type classification (PNES vs. ES) were evaluated by logistic regression. Independent variables included age at EMU admission, duration of seizures, sex (M/F), all axis I diagnoses recorded in ≥ 10% of PNES or ES patients, the total number of documented axis I diagnoses, and presence of any documented axis II diagnosis. Bipolar disorder was excluded from logistic regression analysis as no ES patients had this diagnosis. Forward and backward models were determined using an inclusion criterion of p ≤ 0.1, and a criterion of p ≤ 0.05 to remain in the model. Statistical analyses were performed on SAS software (Cary, NC). This study was approved by the Institutional Review Board and the ethical standards committee of the Portland VAMC.
2.3. Review of psychiatric diagnoses
3. Results
We performed an extensive medical record review of all patients with a diagnostic classification of PNES or ES. The review included (a) the Portland VAMC electronic medical record (EMR); (b) linked EMRs from other VAMCs; (c) scanned records from other facilities that were included in the VA EMR; and (d) paper charts. We did not review military medical records or claims files except where notes were included in the EMR. The review began with the earliest available records and continued through the day prior to the diagnostic EMU admission (defined as the first EMU admission resulting in diagnostic classification). The time interval between the first available medical record and the date of EMU admission was recorded. For each patient, we abstracted all DSM psychiatric diagnoses made by a physician (MD or DO), psychologist (PhD), or psychiatric nurse practitioner (MHNP). We only included diagnoses for which the primary psychiatry/psychology evaluation was available (a ‘qualifying’ mental health evaluation). We did not include psychiatric diagnoses made by primary care physicians, social workers, or other care providers outside of mental health nor did we include psychiatric diagnoses referred to by these providers unless the primary psychiatry/psychology documentation was available. ‘Rule out’ diagnoses were not included. Psychiatric diagnoses that were discussed in mental health notes but not included as a diagnosis were not included. We also recorded the number of psychiatric hospitalizations for each patient, only including admissions for which primary documentation
A median of 76 months of prior medical records was available for review of PNES patients and 116 months for review of ES patients (p = 0.01, Wilcoxon). The difference reflected the longer duration of illness for the ES group (Table 1). The mean age at admission was similar for the PNES and ES groups (47.0 vs. 50.8), and for men and women (48.7 vs. 48.0). Women were over-represented in the PNES group (p b 0.05). Thirty-nine of 50 PNES patients (78%) had at least one qualifying mental health evaluation prior to EMU diagnosis, vs. 23 (62%) of ES patients (ns; Table 1). The presence or absence of a qualifying mental health evaluation was not explained by the months of available records, the duration of the seizure disorder, or patient age at the time of EMU admission (ns; Wilcoxon tests). Only two patients (one PNES, one ES) had a qualifying mental health evaluation but did not have at least one axis 1 diagnosis. Psychogenic non‐epileptic seizure patients had a median of three axis 1 diagnoses (cumulative) prior to EMU admission whereas ES patients had a median of one (p = 0.02). All axis 1 diagnoses recorded in ≥ 10% of PNES or ES patients are graphed in Fig. 1. Post‐traumatic stress disorder was the most common axis I diagnosis in veterans with PNES, documented in 29 patients (58%) vs. 5 ES patients (13.5%; p b 0.001). Major depression was diagnosed in 23 PNES patients (46%) and 11 ES patients (29.7%; ns). Combining major depression and other depression did not result in a significant difference between groups.
M. Salinsky et al. / Epilepsy & Behavior 25 (2012) 345–349 Table 1 Demographics and mental health evaluation summary for veterans with psychogenic non-epileptic seizures and epileptic seizures.
N Age at diagnostic admission (median years; range) Duration of seizures (median, months) Sex (% women) Any qualifying mental health evaluation (%)a Any axis I diagnosis (%) Total axis I diagnoses (median, range) Any axis II diagnosis (%) Any psychiatric hospitalization (%) Total psychiatric hospitalizations (median, range)
PNES
ES
50 49 (24–66)
37 52 (24–88)
p ns (W)
60.5 22.0 78.0
132 5.4 62.2
b0.01 (W) 0.04 (F) ns (F)
76.0 3 (0–8) 36.0 36.0 0 (0–7)
59.5 1 (0–6) 10.8 21.6 0 (0–16)
ns (F) 0.02 (W) 0.01 (F) ns (F) ns (W)
W — Wilcoxon test; F — Fisher's Exact Test; PNES — psychogenic non-epileptic seizures; ES — epileptic seizures. a Qualifying evaluations were by psychiatry (MD,DO, MHNP) or psychology (PhD). Significant p values are bolded.
Similarly, combining alcohol abuse and other substance abuse did not result in a significant difference between groups. Post‐traumatic stress disorder, substance abuse, and depression were often seen in the same PNES patients; 17 of 50 patients (34%) had all three diagnoses, and 29 (58%) had at least two. Ten other axis I diagnoses were recorded in b10% of PNES or ES patients (including attention deficit hyperactivity disorder, explosive disorder, schizoaffective disorder, delusional disorder, conversion disorder, and somatoform disorder). A personality disorder (axis II) was diagnosed in 18 veterans with PNES (36%; Table 1). The most common diagnoses were personality disorder NOS (8 patients) and borderline personality disorder (7 patients). The number of psychiatric hospitalizations per patient did not differ between PNES and ES groups (Table 1). Most patients in each group had no psychiatric hospitalizations (median of zero for both groups). Post‐traumatic stress disorder was attributed to civilian trauma in 15 PNES patients (11 with notation of childhood physical or sexual abuse; 7 men and 4 women), military combat related trauma in 10, military sexual trauma in 4, military non-combat trauma in 1, and the cause was not specified in 3. Post‐traumatic stress disorder was associated with both civilian and military traumas in 4 PNES patients. Nine of 11 women with PNES (82%) had been diagnosed with PTSD (including 4 with military sexual trauma) vs. 19 of 39 men (49%; ns). Univariate analyses (Fisher's Exact Test) revealed significant differences between the PNES and ES groups for diagnoses of PTSD
Percent of Patients
60
*** Psychogenic
50
Epileptic
40 30 20
*
347
Table 2 Logistic regression results for all measures and for the final model (below line). Significant p values are bolded. All variables
Odds ratio
95% confidence interval
Total axis I diagnoses Any axis II diagnosis PTSD Major depression Alcohol abuse Substance abuse Adjustment d/o Anxiety d/o Other depression Age at admission Sex (% female) Duration of illness
0.673 2.375 20.522 1.010 0.718 3.600 0.790 2.293 0.376 0.989 1.664 0.995
0.302–1.497 0.286–19.713 2.625–160.441 0.188–5.418 0.142–3.619 0.310–41.770 0.107–5.816 0.343–15.330 0.057–2.462 0.943–1.038 0.257–10.772 0.991–1.000
Final model PTSD Duration of illness
9.151 0.995
2.878–29.091 0.991–0.999
p > chi-square 0.332 0.423 0.004 0.991 0.688 0.306 0.817 0.392 0.307 0.652 0.593 0.041
b0.001 b0.01
(p b 0.001); any personality disorder (p = 0.01); bipolar disorder (p = 0.02); and other substance abuse (p = 0.03), all with higher frequency in the PNES group (Table 1; Fig. 1). Logistic regression analysis is presented in Table 2. Post‐traumatic stress disorder was the only significant psychiatric classification variable (at p ≤ 0.05). The large confidence interval in the model with all variables is in part due to the large number of variables and relatively small number of subjects. The final model consisted of PTSD and duration of seizures. Overall classification accuracy of the model was 71.3% (at 0.5 probability). However, classification was predominately related to the diagnosis of PTSD. Table 3 is a 2 × 2 classification for seizure type diagnosis (PNES vs. ES), based on a prior diagnosis of PTSD only. Overall classification accuracy was 70.1%. Sensitivity was 58%, specificity 87%, and positive predictive value 85%.
4. Discussion The results demonstrate that in veterans admitted for EMU evaluation, a prior diagnosis of PTSD is predictive of PNES rather than ES. Psychogenic non‐epileptic seizure patients also had a significantly greater overall burden of comorbid psychiatric illness as compared to those with ES, including a greater number of axis I diagnoses and a greater prevalence of axis II diagnoses. However, PTSD was the only diagnosis that significantly contributed to seizure type classification in the logistic regression model. This observation has direct implications for earlier identification of veterans with PNES, a known problem within the VA medical system [8]. An association between PTSD and PNES has been reported previously, and studies comparing PTSD rates in patients with PNES and ES have generally reported higher rates in patients with PNES [14,16,17]. Civilian studies using the Structured Clinical Interview for DSM III or IV have reported PTSD rates of 29–58%, and a study utilizing the PTSD checklist diagnosed PTSD in all 8 patients with
*
10 0
Fig. 1. Axis I diagnoses preceding diagnostic EMU admission for veterans with psychogenic vs. epileptic seizures. PTSD — post traumatic stress disorder. ***p b 0.001; *p b 0.05.
Table 3 2 × 2 classification table for the diagnosis of psychogenic non-epileptic seizures (PNES) vs. epileptic seizures (ES) based on a preceding diagnosis of post-traumatic stress disorder. Observed
Predicted PNES
ES
PNES ES % correct
29 5 85.3%
21 32 60.4% 70.1% (overall correct)
348
M. Salinsky et al. / Epilepsy & Behavior 25 (2012) 345–349
PNES [12–14,16]. In the only previous study of veterans, PTSD was diagnosed in 14 of 22 men with PNES as compared to 3 of 34 with ES [3]. Our results support and extend these observations. Post‐traumatic stress disorder was the most common axis I diagnosis in veterans with PNES, diagnosed in 58% prior to EMU admission. The association between PTSD and PNES contrasts to the relatively weak association with ES. In the present study, PTSD was diagnosed in 13.5% of veterans with ES, only mildly elevated as compared to the general population rate of 8% estimated by the National Co-morbidity Survey, and similar to the estimated prevalence of 15% for Vietnam theater veterans and 17% for veterans of Iraq [18–20]. The contrast between PTSD rates in veterans with PNES and those with ES resulted in a positive predictive value of >85%. Psychogenic non‐epileptic seizures were more frequently comorbid with axis I and II psychiatric diagnoses as compared to ES. This observation is not explained by availability of medical records, as these were substantially more available for the ES group. Excluding PTSD, major depression was the most common axis I diagnosis, recorded in nearly 50% of the PNES group. However, the difference in the incidence of major depression (or all types of depression combined) was not statistically significant, and depression did not contribute to the classification of ES vs. PNES. Similar findings have been reported in both civilian and veteran comparison studies [3,14,21]. Rates of major depression are increased in ES patients as compared to the general population, particularly in patients with poorly controlled seizures, the group likely to be admitted to an EMU [22]. Differences between PNES and ES groups may, therefore, be reduced. In the same way, a diagnosis of substance abuse (including alcohol abuse) did not separate veterans with PNES and ES due to the high prevalence in both groups. Personality disorders, though more common in PNES patients, similarly did not contribute to seizure type classification on multivariate analysis. The roles of physical and psychological traumas in the development of PNES and PTSD have been discussed elsewhere [12,14, 16,17,23]. The relatively high rates of both disorders in veterans reflect the burden of combat and other military traumas, added to those traumas typical in civilian populations. Half of the veterans with PNES and PTSD in the present study were diagnosed with PTSD related to military experiences. These were primarily combat experiences and also included military sexual trauma and other traumas. Others were thought to have PTSD on the basis of civilian traumas, primarily physical and sexual abuse in childhood as is commonly reported in civilian PNES series [12,16,23]. Exposure to military traumas is an issue specific to veterans and may result in unique psychological effects. As such, the strong association between PNES and PTSD found in this study may not be generalizable to civilians. We have previously shown a relatively high ratio of PNES to ES diagnoses in veterans (as compared to civilians) admitted to the EMU [8]. The percent of patients discharged from EMU evaluation with a diagnosis of PNES was identical for veterans and civilians (~ 25%), but fewer veterans were diagnosed with ES (18% of discharges vs. 40% in civilians). The reasons for this high ratio are not known but may in part be mediated through PTSD and through awareness of the relationship between TBI and the development of epileptic seizures. Post‐traumatic stress disorder is relatively common in veterans of recent US military conflicts and is associated with an increased risk of somatic symptoms, including neurological symptoms [20,24]. The current findings suggest that PTSD is also strongly associated with the development of PNES. It can be speculated that the development of seizures (as opposed to other somatic symptoms) is related to awareness of the association between TBI, particularly military TBI, and the development of epileptic seizures [25]. This awareness and the emergence of TBI as the signal injury of veterans of recent military conflicts may prime the development of seizure-like symptoms in veterans predisposed to somatization. Veterans with TBI (self-reported in 58% of our PNES patients) and PTSD may, therefore, be at particularly
high risk [8]. In general, PTSD may mediate negative health outcomes in veterans with mild TBI [26,27]. This retrospective study has several limitations. We were able to take advantage of the linked EMR systems within the VAMCs. However, we could not be certain that all psychiatric evaluations had been reviewed, and it is likely that the diagnoses abstracted are an underestimate of psychiatric diagnoses prior to EMU evaluation. We accepted only DSM III‐ or IV‐based diagnoses from psychiatrists, psychologists, or mental health nurse practitioners so as to optimize consistency. However, care providers did not use a structured inventory and may have used somewhat different criteria in reaching their diagnoses. Also, the patients reviewed in this study represent those admitted to an EMU, usually for the evaluation of poorly controlled seizures. As such, they may not represent the broader population of seizure patients served by VAMCs. Our observations are only valid for those veterans seeking care within the VA medical system, a minority of veterans overall. Many of these limitations will be addressed in ongoing prospective studies. Early diagnosis of PNES may be the most effective way of reducing long-term disability. Civilian studies have shown a reduction in seizures, AED usage, and health care utilization following diagnosis and conventional care [7,9–11,28]. Delay in diagnosis may also be associated with worse prognosis [29,30]. In a previous study, we documented a nearly 5-fold delay in the diagnosis of PNES in veterans as compared to civilians [8]. The present study identifies an easily elicited historical risk factor for PNES. A prior diagnosis of PTSD in veterans with poorly controlled seizures should alert clinicians to the possibility of PNES and lead to early referral for diagnostic monitoring. Acknowledgments The authors thank J. Cereghino MD and D. Smith MD for review of this manuscript. This material is the result of work supported with resources and the use of facilities at the Portland VA Medical Center. References [1] Martin R, Burneo JG, Prasad A, et al. Frequency of epilepsy in patients with psychogenic seizures monitored by video-EEG. Neurology 2003;61:1791-2. [2] Benbadis SR, O'Neill E, Tatum WO, Heriaud L. Outcome of prolonged video-EEG monitoring at a typical referral epilepsy center. Epilepsia 2004;45:1150-3. [3] Dworetsky B, Strahonja-Packard A, Shanahan C, Paz J, Schauble B, Bromfield E. Characteristics of male veterans with psychogenic nonepileptic seizures. Epilepsia 2005;46:1418-22. [4] Pierelli F, Chatrian G, Erdly W, Swanson P. Long-term EEG-video-audio monitoring: detection of partial epileptic seizures and psychogenic episodes by 24-hour EEG review. Epilepsia 1989;30:512-23. [5] Selwa LM, Geyer J, Nikakhtar N, Brown MB, Schuh LA, Drury I. Nonepileptic seizure outcome varies by type of spell and duration of illness. Epilepsia 2000;41:1330-4. [6] Benbadis SR, Agrawal V, Tatum WO. How many patients with psychogenic nonepileptic seizures also have epilepsy? Neurology 2001;57:915-7. [7] McKenzie P, Oto M, Russell A, Pelosi A, Duncan R. Early outcomes and predictors in 260 patients with psychogenic nonepileptic attacks. Neurology 2010;74:64-9. [8] Salinsky MC, Spencer D, Ferguson F, Boudreau E. Psychogenic seizures in U.S. veterans. Neurology 2011;77:945-50. [9] Ettinger AB, Devinsky O, Weisbrot DM, Ramakrishna RK, Goyal A. A comprehensive profile of clinical, psychiatric, and psychosocial characteristics of patients with psychogenic nonepileptic seizures. Epilepsia 1999;40:1292-8. [10] Reuber M, Pukrop R, Bauer J, Helmstaedter C, Tessendorf N, Elger CE. Outcome in psychogenic nonepileptic seizures: 1 to 10 year follow-up in 164 patients. Ann Neurol 2003;53:305-11. [11] Walczak T, Papacostas S, Williams DT, Scheuer ML, Lebowitz N, Notarfrancesco A. Outcome after diagnosis of psychogenic nonepileptic seizures. Epilepsia 1995;36: 1131-7. [12] Bowman E, Markand O. Psychodynamics and psychiatric diagnoses of pseudoseizure patients. Am J Psychiatry 1996;153:57-63. [13] D'Alessio L, Giagante B, Oddo S, et al. Psychiatric disorders in patients with psychogenic nonepileptic seizures, with and without comorbid epilepsy. Seizure 2006;15:333-9. [14] Arnold L, Privatera M. Psychopathology and trauma in epileptic and psychogenic seizure patients. Psychosomatics 1996;37:438-43. [15] Marchetti R, Kurcgant D, Neto J, Bismark M, Marchetti L, Fiore L. Psychiatric diagnoses in psychogenic non-epileptic seizures. Seizure 2010;17:247-53.
M. Salinsky et al. / Epilepsy & Behavior 25 (2012) 345–349 [16] Rosenberg HJ, Rosenberg SD, Williamson PD, Wolford GL. A comparative study of trauma and posttraumatic stress disorder in prevalence in epilepsy patients and psychogenic nonepileptic seizure patients. Epilepsia 2000;41:447-52. [17] Dikel TN, Fennell EB, Gilmore RL. Posttraumatic stress disorder, dissociation, and sexual abuse history in epileptic and nonepileptic seizure patients. Epilepsy Behav 2003;4:644-50. [18] Kessler R, Sonnega A, Bromet E, Hughes M, Nelson C. Post-traumatic stress disorder in the national comorbidity study. Arch Gen Psychiatry 1995;52:1048-59. [19] Schlenger W, Kulka R, Fairbank J, et al. The prevalence of post-traumatic stress disorder in the Vietnam generation: a multimethod, multisource assessment of psychiatric disorder. J Trauma Stress 1992;5:333-63. [20] Hoge CW, Terhakopian A, Castro CA, Messer SC, Engel C. Association of post-traumatic stress disorder with somatic symptoms, health care visits, and absenteeism among Iraq war veterans. Am J Psychiatry 2007;164:150-3. [21] Asmussen SB, Kirlin KA, Gale SD, Chung SS. Differences in self-reported depressive symptoms between patients with epileptic and psychogenic nonepileptic seizures. Seizure 2009;18:564-6. [22] La France W, Kanner A, Hermann B. Psychiatric comorbidities in epilepsy. Int Rev Neurobiol 2008;83:347-83.
349
[23] Alper K, Devinsky O, Perrine K, Vazquez B, Luciano D. Nonepileptic seizures and childhood sexual and physical abuse. Neurology 1993;43:1950-3. [24] Andreski P, Chilcoat H, Breslau N. Post-traumatic stress disorder and somatization symptoms: a prospective study. Psychiatry Res 1998;79:131-8. [25] Raymont V, Salizar AM, Lipsky R, Goldman D, Tasick G, Grafman J. Correlates of posttraumatic epilepsy 35 years after combat head injury. Neurology 2010;75:224-9. [26] Hoge CW, McGurk D, Thomas JL, Cox AL, Engel CC, Castro CA. Mild traumatic brain injury in U.S. soldiers returning from Iraq. N Engl J Med 2008;358:453-63. [27] Pietrzak RH, Johnson DC, Goldstein MB, Malley JC, Southwick SM. Posttraumatic stress disorder mediates the relationship between mild traumatic brain injury and health and psychosocial functioning in veterans of Operations Enduring Freedom and Iraqi Freedom. J Nerv Ment Dis 2009;197:748-53. [28] Martin R, Gilliam FG, Kilgore M, Faught E, Kuzniecky R. Improved health care resource utilization following video-EEG confirmed diagnosis of nonepileptic seizures. Seizure 1998;7:385-90. [29] Lempert T, Schmidt D. Natural history and outcome of psychogenic seizures: a clinical study in 50 patients. J Neurol 1990;237:35-8. [30] Buchanan N, Snars J. Pseudoseizures: clinical management and outcome in 50 patients. Seizure 1993;2:141-6.
Contents lists available at SciVerse ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Psychiatric comorbidity in veterans with psychogenic seizures Martin Salinsky a, b,⁎, Collette Evrard a, Daniel Storzbach a, Mary Jo Pugh c a b c
Portland Veterans Affairs Medical Center, Portland, OR, USA Oregon Health & Science University, Portland, OR, USA San Antonio Veterans Affairs Medical Center, San Antonio, TX, USA
a r t i c l e
i n f o
Article history: Received 24 April 2012 Revised 9 July 2012 Accepted 14 July 2012 Available online 24 October 2012 Keywords: Epilepsy Psychogenic seizures Veterans Post-traumatic stress disorder
a b s t r a c t Psychogenic non-epileptic seizures (PNES) are frequently encountered in epilepsy monitoring units (EMU) at Veterans Affairs Medical Centers (VAMCs) and cause significant long-term disability. An understanding of psychiatric factors associated with PNES could aid in earlier diagnosis and treatment. We studied 50 consecutive veterans diagnosed with PNES and 37 veterans diagnosed with epileptic seizures (ES), evaluated at a VAMC EMU. We reviewed all available mental health evaluations prior to EMU evaluation. Univariate comparisons included axis I diagnoses, axis II diagnoses, and psychiatric hospitalizations. Predictive models of seizure classification were evaluated by logistic regression. A diagnosis of post-traumatic stress disorder (PTSD) preceded the diagnosis of PNES in 58% of patients and the diagnosis of ES in 13.5% (pb 0.001). On logistic regression, PTSD was the only significant psychiatric diagnosis (odds ratio 9.2). Major depression and alcohol abuse were common diagnoses but did not differentiate PNES and ES groups. Published by Elsevier Inc.
1. Introduction Psychogenic non-epileptic seizures (PNES) are clinical events that resemble epileptic seizures (ES) but that are not associated with abnormal brain electrical discharges or other physiological problems. These neuropsychiatric events are common, accounting for 20–50% of epilepsy monitoring unit (EMU) discharge diagnoses [1–5]. Most PNES patients are mistakenly thought to have medication‐resistant epilepsy and are chronically treated with one or more antiepileptic drugs (AEDs), while the psychological nature of the disorder remains hidden [3,6–9]. Short‐ and long-term disabilities are common [7,10]. Psychogenic non‐epileptic seizures are also common in veterans treated at Veterans Affairs Medical Centers (VAMCs). Patients diagnosed with PNES accounted for 25% of all EMU discharge diagnoses in a recent VAMC study and were more common than patients diagnosed with ES [8]. In the same study, 50 veterans with EMU‐verified PNES (without ES) were compared with 50 civilians with PNES studied in the same EMU and by the same personnel. Veterans had a considerably longer delay from onset of symptoms to the diagnosis of PNES as compared to civilians, with a median of 5 years. In one third of cases the delay exceeded a decade. The delay in diagnosis was associated with greater cumulative exposure to AEDs. Prognosis and long-term disability may also be negatively affected by the delay in diagnosis [5,7,11].
⁎ Corresponding author at: 3181 SW Sam Jackson Park Road, CR-120, Portland, OR 97239, USA. Fax: +1 503 494 6658. E-mail address: [email protected] (M. Salinsky). 1525-5050/$ – see front matter. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.yebeh.2012.07.013
Psychogenic non‐epileptic seizures are generally believed to be symptoms of comorbid psychiatric disorders. These disorders have been comprehensively studied in civilians but not in veterans [12–15]. An understanding of the psychiatric disorders associated with PNES is essential for designing effective treatment strategies and could aid in early recognition of PNES. Comorbid psychiatric disorders in veterans with PNES are likely to differ from those of civilians due to (a) differences in the age and sex of the populations; (b) risk of traumatic brain injury (TBI) and psychological effects of combat exposure including post-traumatic stress disorder (PTSD); and (c) potential financial motivations in the VA health care system fostering illness behavior. We studied psychiatric diagnoses preceding documentation of PNES in a carefully evaluated group of veterans with PNES and a comparison group of veterans with documented ES. 2. Methods 2.1. Patients We reviewed records of all veterans admitted to the Portland VAMC EMU from January 2000 to July 2010. Details for this group are provided elsewhere [8]. We excluded patients who had a previously documented EMU evaluation at the Portland VAMC or elsewhere, except (a) when these evaluations were non-diagnostic or (b) when the diagnosis was modified by the repeat monitoring at the Portland VAMC. We eliminated 4 patients with missing records. There were 203 unique patients in the final group. The mean age at admission was 50.6 years. Eighty-seven percent were men.
346
M. Salinsky et al. / Epilepsy & Behavior 25 (2012) 345–349
2.2. Classification Each EMU admission was reviewed, and every patient was classified into one of five diagnostic categories: (1) ES only; captured ES (diagnostic criteria below) with no current or historical verification of non-epileptic seizures. (2) PNES only; captured PNES (diagnostic criteria below) with no current or historical verification of ES and no interictal epileptiform activity (IIEA). (3) ‘Mixed’ ES and PNES; captured ES and PNES or PNES and IIEA. (4) Other non-epileptic seizures; captured physiologic events confused with ES, including syncope, sleep disorders, etc. No current or historical verification of ES or PNES. (5) Nondiagnostic; no diagnostic events captured during EMU admission (including patients with subjective spells without EEG correlate, patients with IIEA and no diagnostic spells, and patients with spells that could not be confidently classified). A total of 50 patients met criteria for PNES only (mean age 47.0; 80% men), and 37 patients met criteria for ES only (mean age 50.8; 95% men). These patients comprised the study groups. The diagnosis of PNES used the following criteria: (1) Captured events during the EMU admission were typical for the patient's habitual spells and demonstrated alteration of consciousness and/ or bilateral motor activity. (2) At least two diagnostic events or one prolonged event (> 5‐minute ictal phase). We accepted one event provided that it was endorsed by a friend or family member as typical, based on video review (3 patients). (3) No epileptiform discharges or ictal rhythmic EEG changes associated with any event. (4) Events had a presentation consistent with PNES. (5) No other physiologic abnormalities that might reasonably explain the spells. (6) No IIEA. (7) No previous verifiable documentation of ES or IIEA (from EMU or other EEG studies). The diagnosis of ES required (1) captured events during the EMU admission that were typical for the patient's habitual spells; (2) at least two diagnostic events or one prolonged diagnostic event (> 5‐minute ictal phase) as mentioned above; (3) epileptiform discharges or ictal rhythmic EEG changes associated with each event; (4) events that had a presentation consistent with ES; and (5) no verifiable documentation of PNES (from prior EMU or other EEG studies). All diagnoses were made by fellowship‐ trained epileptologists and confirmed for this study by one investigator (MS).
was included in the medical records. For patients with a diagnosis of PTSD, we classified the cause as ‘civilian’, ‘military combat’, ‘military sexual trauma’, ‘military non-combat’, or ‘not specified’ based on the qualifying mental health evaluations only. Patient age at admission, sex, and duration of seizures prior to EMU diagnosis (determined from all available sources including patient/family history, paper and electronic charts) were also recorded. To simplify analysis, certain axis I diagnoses were combined. The category ‘other depression’ included all depression not categorized as major depression (including dysthymia, depression NOS, mood disorder associated with medical illness). ‘Other depression’ was only counted once even if there were multiple types of ‘other depression’ diagnosed. Patients with major depression, who were also diagnosed with ‘other depression’, were only counted as having major depression. Bipolar disorder and adjustment disorder with depressed mood were included as separate axis I categories. The category ‘other substance abuse’ included all substance abuse not categorized as alcohol‐related. Other substance abuse was only counted once (even if multiple substances were listed separately); however, patients could have diagnoses of both other substance abuse and alcohol abuse. 2.4. Statistical analysis Two-group (PNES vs. ES) univariate comparisons used either the Wilcoxon test or Fisher's Exact Test. Significance level was set at p ≤ 0.05 (two-tailed). Predictive models of seizure type classification (PNES vs. ES) were evaluated by logistic regression. Independent variables included age at EMU admission, duration of seizures, sex (M/F), all axis I diagnoses recorded in ≥ 10% of PNES or ES patients, the total number of documented axis I diagnoses, and presence of any documented axis II diagnosis. Bipolar disorder was excluded from logistic regression analysis as no ES patients had this diagnosis. Forward and backward models were determined using an inclusion criterion of p ≤ 0.1, and a criterion of p ≤ 0.05 to remain in the model. Statistical analyses were performed on SAS software (Cary, NC). This study was approved by the Institutional Review Board and the ethical standards committee of the Portland VAMC.
2.3. Review of psychiatric diagnoses
3. Results
We performed an extensive medical record review of all patients with a diagnostic classification of PNES or ES. The review included (a) the Portland VAMC electronic medical record (EMR); (b) linked EMRs from other VAMCs; (c) scanned records from other facilities that were included in the VA EMR; and (d) paper charts. We did not review military medical records or claims files except where notes were included in the EMR. The review began with the earliest available records and continued through the day prior to the diagnostic EMU admission (defined as the first EMU admission resulting in diagnostic classification). The time interval between the first available medical record and the date of EMU admission was recorded. For each patient, we abstracted all DSM psychiatric diagnoses made by a physician (MD or DO), psychologist (PhD), or psychiatric nurse practitioner (MHNP). We only included diagnoses for which the primary psychiatry/psychology evaluation was available (a ‘qualifying’ mental health evaluation). We did not include psychiatric diagnoses made by primary care physicians, social workers, or other care providers outside of mental health nor did we include psychiatric diagnoses referred to by these providers unless the primary psychiatry/psychology documentation was available. ‘Rule out’ diagnoses were not included. Psychiatric diagnoses that were discussed in mental health notes but not included as a diagnosis were not included. We also recorded the number of psychiatric hospitalizations for each patient, only including admissions for which primary documentation
A median of 76 months of prior medical records was available for review of PNES patients and 116 months for review of ES patients (p = 0.01, Wilcoxon). The difference reflected the longer duration of illness for the ES group (Table 1). The mean age at admission was similar for the PNES and ES groups (47.0 vs. 50.8), and for men and women (48.7 vs. 48.0). Women were over-represented in the PNES group (p b 0.05). Thirty-nine of 50 PNES patients (78%) had at least one qualifying mental health evaluation prior to EMU diagnosis, vs. 23 (62%) of ES patients (ns; Table 1). The presence or absence of a qualifying mental health evaluation was not explained by the months of available records, the duration of the seizure disorder, or patient age at the time of EMU admission (ns; Wilcoxon tests). Only two patients (one PNES, one ES) had a qualifying mental health evaluation but did not have at least one axis 1 diagnosis. Psychogenic non‐epileptic seizure patients had a median of three axis 1 diagnoses (cumulative) prior to EMU admission whereas ES patients had a median of one (p = 0.02). All axis 1 diagnoses recorded in ≥ 10% of PNES or ES patients are graphed in Fig. 1. Post‐traumatic stress disorder was the most common axis I diagnosis in veterans with PNES, documented in 29 patients (58%) vs. 5 ES patients (13.5%; p b 0.001). Major depression was diagnosed in 23 PNES patients (46%) and 11 ES patients (29.7%; ns). Combining major depression and other depression did not result in a significant difference between groups.
M. Salinsky et al. / Epilepsy & Behavior 25 (2012) 345–349 Table 1 Demographics and mental health evaluation summary for veterans with psychogenic non-epileptic seizures and epileptic seizures.
N Age at diagnostic admission (median years; range) Duration of seizures (median, months) Sex (% women) Any qualifying mental health evaluation (%)a Any axis I diagnosis (%) Total axis I diagnoses (median, range) Any axis II diagnosis (%) Any psychiatric hospitalization (%) Total psychiatric hospitalizations (median, range)
PNES
ES
50 49 (24–66)
37 52 (24–88)
p ns (W)
60.5 22.0 78.0
132 5.4 62.2
b0.01 (W) 0.04 (F) ns (F)
76.0 3 (0–8) 36.0 36.0 0 (0–7)
59.5 1 (0–6) 10.8 21.6 0 (0–16)
ns (F) 0.02 (W) 0.01 (F) ns (F) ns (W)
W — Wilcoxon test; F — Fisher's Exact Test; PNES — psychogenic non-epileptic seizures; ES — epileptic seizures. a Qualifying evaluations were by psychiatry (MD,DO, MHNP) or psychology (PhD). Significant p values are bolded.
Similarly, combining alcohol abuse and other substance abuse did not result in a significant difference between groups. Post‐traumatic stress disorder, substance abuse, and depression were often seen in the same PNES patients; 17 of 50 patients (34%) had all three diagnoses, and 29 (58%) had at least two. Ten other axis I diagnoses were recorded in b10% of PNES or ES patients (including attention deficit hyperactivity disorder, explosive disorder, schizoaffective disorder, delusional disorder, conversion disorder, and somatoform disorder). A personality disorder (axis II) was diagnosed in 18 veterans with PNES (36%; Table 1). The most common diagnoses were personality disorder NOS (8 patients) and borderline personality disorder (7 patients). The number of psychiatric hospitalizations per patient did not differ between PNES and ES groups (Table 1). Most patients in each group had no psychiatric hospitalizations (median of zero for both groups). Post‐traumatic stress disorder was attributed to civilian trauma in 15 PNES patients (11 with notation of childhood physical or sexual abuse; 7 men and 4 women), military combat related trauma in 10, military sexual trauma in 4, military non-combat trauma in 1, and the cause was not specified in 3. Post‐traumatic stress disorder was associated with both civilian and military traumas in 4 PNES patients. Nine of 11 women with PNES (82%) had been diagnosed with PTSD (including 4 with military sexual trauma) vs. 19 of 39 men (49%; ns). Univariate analyses (Fisher's Exact Test) revealed significant differences between the PNES and ES groups for diagnoses of PTSD
Percent of Patients
60
*** Psychogenic
50
Epileptic
40 30 20
*
347
Table 2 Logistic regression results for all measures and for the final model (below line). Significant p values are bolded. All variables
Odds ratio
95% confidence interval
Total axis I diagnoses Any axis II diagnosis PTSD Major depression Alcohol abuse Substance abuse Adjustment d/o Anxiety d/o Other depression Age at admission Sex (% female) Duration of illness
0.673 2.375 20.522 1.010 0.718 3.600 0.790 2.293 0.376 0.989 1.664 0.995
0.302–1.497 0.286–19.713 2.625–160.441 0.188–5.418 0.142–3.619 0.310–41.770 0.107–5.816 0.343–15.330 0.057–2.462 0.943–1.038 0.257–10.772 0.991–1.000
Final model PTSD Duration of illness
9.151 0.995
2.878–29.091 0.991–0.999
p > chi-square 0.332 0.423 0.004 0.991 0.688 0.306 0.817 0.392 0.307 0.652 0.593 0.041
b0.001 b0.01
(p b 0.001); any personality disorder (p = 0.01); bipolar disorder (p = 0.02); and other substance abuse (p = 0.03), all with higher frequency in the PNES group (Table 1; Fig. 1). Logistic regression analysis is presented in Table 2. Post‐traumatic stress disorder was the only significant psychiatric classification variable (at p ≤ 0.05). The large confidence interval in the model with all variables is in part due to the large number of variables and relatively small number of subjects. The final model consisted of PTSD and duration of seizures. Overall classification accuracy of the model was 71.3% (at 0.5 probability). However, classification was predominately related to the diagnosis of PTSD. Table 3 is a 2 × 2 classification for seizure type diagnosis (PNES vs. ES), based on a prior diagnosis of PTSD only. Overall classification accuracy was 70.1%. Sensitivity was 58%, specificity 87%, and positive predictive value 85%.
4. Discussion The results demonstrate that in veterans admitted for EMU evaluation, a prior diagnosis of PTSD is predictive of PNES rather than ES. Psychogenic non‐epileptic seizure patients also had a significantly greater overall burden of comorbid psychiatric illness as compared to those with ES, including a greater number of axis I diagnoses and a greater prevalence of axis II diagnoses. However, PTSD was the only diagnosis that significantly contributed to seizure type classification in the logistic regression model. This observation has direct implications for earlier identification of veterans with PNES, a known problem within the VA medical system [8]. An association between PTSD and PNES has been reported previously, and studies comparing PTSD rates in patients with PNES and ES have generally reported higher rates in patients with PNES [14,16,17]. Civilian studies using the Structured Clinical Interview for DSM III or IV have reported PTSD rates of 29–58%, and a study utilizing the PTSD checklist diagnosed PTSD in all 8 patients with
*
10 0
Fig. 1. Axis I diagnoses preceding diagnostic EMU admission for veterans with psychogenic vs. epileptic seizures. PTSD — post traumatic stress disorder. ***p b 0.001; *p b 0.05.
Table 3 2 × 2 classification table for the diagnosis of psychogenic non-epileptic seizures (PNES) vs. epileptic seizures (ES) based on a preceding diagnosis of post-traumatic stress disorder. Observed
Predicted PNES
ES
PNES ES % correct
29 5 85.3%
21 32 60.4% 70.1% (overall correct)
348
M. Salinsky et al. / Epilepsy & Behavior 25 (2012) 345–349
PNES [12–14,16]. In the only previous study of veterans, PTSD was diagnosed in 14 of 22 men with PNES as compared to 3 of 34 with ES [3]. Our results support and extend these observations. Post‐traumatic stress disorder was the most common axis I diagnosis in veterans with PNES, diagnosed in 58% prior to EMU admission. The association between PTSD and PNES contrasts to the relatively weak association with ES. In the present study, PTSD was diagnosed in 13.5% of veterans with ES, only mildly elevated as compared to the general population rate of 8% estimated by the National Co-morbidity Survey, and similar to the estimated prevalence of 15% for Vietnam theater veterans and 17% for veterans of Iraq [18–20]. The contrast between PTSD rates in veterans with PNES and those with ES resulted in a positive predictive value of >85%. Psychogenic non‐epileptic seizures were more frequently comorbid with axis I and II psychiatric diagnoses as compared to ES. This observation is not explained by availability of medical records, as these were substantially more available for the ES group. Excluding PTSD, major depression was the most common axis I diagnosis, recorded in nearly 50% of the PNES group. However, the difference in the incidence of major depression (or all types of depression combined) was not statistically significant, and depression did not contribute to the classification of ES vs. PNES. Similar findings have been reported in both civilian and veteran comparison studies [3,14,21]. Rates of major depression are increased in ES patients as compared to the general population, particularly in patients with poorly controlled seizures, the group likely to be admitted to an EMU [22]. Differences between PNES and ES groups may, therefore, be reduced. In the same way, a diagnosis of substance abuse (including alcohol abuse) did not separate veterans with PNES and ES due to the high prevalence in both groups. Personality disorders, though more common in PNES patients, similarly did not contribute to seizure type classification on multivariate analysis. The roles of physical and psychological traumas in the development of PNES and PTSD have been discussed elsewhere [12,14, 16,17,23]. The relatively high rates of both disorders in veterans reflect the burden of combat and other military traumas, added to those traumas typical in civilian populations. Half of the veterans with PNES and PTSD in the present study were diagnosed with PTSD related to military experiences. These were primarily combat experiences and also included military sexual trauma and other traumas. Others were thought to have PTSD on the basis of civilian traumas, primarily physical and sexual abuse in childhood as is commonly reported in civilian PNES series [12,16,23]. Exposure to military traumas is an issue specific to veterans and may result in unique psychological effects. As such, the strong association between PNES and PTSD found in this study may not be generalizable to civilians. We have previously shown a relatively high ratio of PNES to ES diagnoses in veterans (as compared to civilians) admitted to the EMU [8]. The percent of patients discharged from EMU evaluation with a diagnosis of PNES was identical for veterans and civilians (~ 25%), but fewer veterans were diagnosed with ES (18% of discharges vs. 40% in civilians). The reasons for this high ratio are not known but may in part be mediated through PTSD and through awareness of the relationship between TBI and the development of epileptic seizures. Post‐traumatic stress disorder is relatively common in veterans of recent US military conflicts and is associated with an increased risk of somatic symptoms, including neurological symptoms [20,24]. The current findings suggest that PTSD is also strongly associated with the development of PNES. It can be speculated that the development of seizures (as opposed to other somatic symptoms) is related to awareness of the association between TBI, particularly military TBI, and the development of epileptic seizures [25]. This awareness and the emergence of TBI as the signal injury of veterans of recent military conflicts may prime the development of seizure-like symptoms in veterans predisposed to somatization. Veterans with TBI (self-reported in 58% of our PNES patients) and PTSD may, therefore, be at particularly
high risk [8]. In general, PTSD may mediate negative health outcomes in veterans with mild TBI [26,27]. This retrospective study has several limitations. We were able to take advantage of the linked EMR systems within the VAMCs. However, we could not be certain that all psychiatric evaluations had been reviewed, and it is likely that the diagnoses abstracted are an underestimate of psychiatric diagnoses prior to EMU evaluation. We accepted only DSM III‐ or IV‐based diagnoses from psychiatrists, psychologists, or mental health nurse practitioners so as to optimize consistency. However, care providers did not use a structured inventory and may have used somewhat different criteria in reaching their diagnoses. Also, the patients reviewed in this study represent those admitted to an EMU, usually for the evaluation of poorly controlled seizures. As such, they may not represent the broader population of seizure patients served by VAMCs. Our observations are only valid for those veterans seeking care within the VA medical system, a minority of veterans overall. Many of these limitations will be addressed in ongoing prospective studies. Early diagnosis of PNES may be the most effective way of reducing long-term disability. Civilian studies have shown a reduction in seizures, AED usage, and health care utilization following diagnosis and conventional care [7,9–11,28]. Delay in diagnosis may also be associated with worse prognosis [29,30]. In a previous study, we documented a nearly 5-fold delay in the diagnosis of PNES in veterans as compared to civilians [8]. The present study identifies an easily elicited historical risk factor for PNES. A prior diagnosis of PTSD in veterans with poorly controlled seizures should alert clinicians to the possibility of PNES and lead to early referral for diagnostic monitoring. Acknowledgments The authors thank J. Cereghino MD and D. Smith MD for review of this manuscript. This material is the result of work supported with resources and the use of facilities at the Portland VA Medical Center. References [1] Martin R, Burneo JG, Prasad A, et al. Frequency of epilepsy in patients with psychogenic seizures monitored by video-EEG. Neurology 2003;61:1791-2. [2] Benbadis SR, O'Neill E, Tatum WO, Heriaud L. Outcome of prolonged video-EEG monitoring at a typical referral epilepsy center. Epilepsia 2004;45:1150-3. [3] Dworetsky B, Strahonja-Packard A, Shanahan C, Paz J, Schauble B, Bromfield E. Characteristics of male veterans with psychogenic nonepileptic seizures. Epilepsia 2005;46:1418-22. [4] Pierelli F, Chatrian G, Erdly W, Swanson P. Long-term EEG-video-audio monitoring: detection of partial epileptic seizures and psychogenic episodes by 24-hour EEG review. Epilepsia 1989;30:512-23. [5] Selwa LM, Geyer J, Nikakhtar N, Brown MB, Schuh LA, Drury I. Nonepileptic seizure outcome varies by type of spell and duration of illness. Epilepsia 2000;41:1330-4. [6] Benbadis SR, Agrawal V, Tatum WO. How many patients with psychogenic nonepileptic seizures also have epilepsy? Neurology 2001;57:915-7. [7] McKenzie P, Oto M, Russell A, Pelosi A, Duncan R. Early outcomes and predictors in 260 patients with psychogenic nonepileptic attacks. Neurology 2010;74:64-9. [8] Salinsky MC, Spencer D, Ferguson F, Boudreau E. Psychogenic seizures in U.S. veterans. Neurology 2011;77:945-50. [9] Ettinger AB, Devinsky O, Weisbrot DM, Ramakrishna RK, Goyal A. A comprehensive profile of clinical, psychiatric, and psychosocial characteristics of patients with psychogenic nonepileptic seizures. Epilepsia 1999;40:1292-8. [10] Reuber M, Pukrop R, Bauer J, Helmstaedter C, Tessendorf N, Elger CE. Outcome in psychogenic nonepileptic seizures: 1 to 10 year follow-up in 164 patients. Ann Neurol 2003;53:305-11. [11] Walczak T, Papacostas S, Williams DT, Scheuer ML, Lebowitz N, Notarfrancesco A. Outcome after diagnosis of psychogenic nonepileptic seizures. Epilepsia 1995;36: 1131-7. [12] Bowman E, Markand O. Psychodynamics and psychiatric diagnoses of pseudoseizure patients. Am J Psychiatry 1996;153:57-63. [13] D'Alessio L, Giagante B, Oddo S, et al. Psychiatric disorders in patients with psychogenic nonepileptic seizures, with and without comorbid epilepsy. Seizure 2006;15:333-9. [14] Arnold L, Privatera M. Psychopathology and trauma in epileptic and psychogenic seizure patients. Psychosomatics 1996;37:438-43. [15] Marchetti R, Kurcgant D, Neto J, Bismark M, Marchetti L, Fiore L. Psychiatric diagnoses in psychogenic non-epileptic seizures. Seizure 2010;17:247-53.
M. Salinsky et al. / Epilepsy & Behavior 25 (2012) 345–349 [16] Rosenberg HJ, Rosenberg SD, Williamson PD, Wolford GL. A comparative study of trauma and posttraumatic stress disorder in prevalence in epilepsy patients and psychogenic nonepileptic seizure patients. Epilepsia 2000;41:447-52. [17] Dikel TN, Fennell EB, Gilmore RL. Posttraumatic stress disorder, dissociation, and sexual abuse history in epileptic and nonepileptic seizure patients. Epilepsy Behav 2003;4:644-50. [18] Kessler R, Sonnega A, Bromet E, Hughes M, Nelson C. Post-traumatic stress disorder in the national comorbidity study. Arch Gen Psychiatry 1995;52:1048-59. [19] Schlenger W, Kulka R, Fairbank J, et al. The prevalence of post-traumatic stress disorder in the Vietnam generation: a multimethod, multisource assessment of psychiatric disorder. J Trauma Stress 1992;5:333-63. [20] Hoge CW, Terhakopian A, Castro CA, Messer SC, Engel C. Association of post-traumatic stress disorder with somatic symptoms, health care visits, and absenteeism among Iraq war veterans. Am J Psychiatry 2007;164:150-3. [21] Asmussen SB, Kirlin KA, Gale SD, Chung SS. Differences in self-reported depressive symptoms between patients with epileptic and psychogenic nonepileptic seizures. Seizure 2009;18:564-6. [22] La France W, Kanner A, Hermann B. Psychiatric comorbidities in epilepsy. Int Rev Neurobiol 2008;83:347-83.
349
[23] Alper K, Devinsky O, Perrine K, Vazquez B, Luciano D. Nonepileptic seizures and childhood sexual and physical abuse. Neurology 1993;43:1950-3. [24] Andreski P, Chilcoat H, Breslau N. Post-traumatic stress disorder and somatization symptoms: a prospective study. Psychiatry Res 1998;79:131-8. [25] Raymont V, Salizar AM, Lipsky R, Goldman D, Tasick G, Grafman J. Correlates of posttraumatic epilepsy 35 years after combat head injury. Neurology 2010;75:224-9. [26] Hoge CW, McGurk D, Thomas JL, Cox AL, Engel CC, Castro CA. Mild traumatic brain injury in U.S. soldiers returning from Iraq. N Engl J Med 2008;358:453-63. [27] Pietrzak RH, Johnson DC, Goldstein MB, Malley JC, Southwick SM. Posttraumatic stress disorder mediates the relationship between mild traumatic brain injury and health and psychosocial functioning in veterans of Operations Enduring Freedom and Iraqi Freedom. J Nerv Ment Dis 2009;197:748-53. [28] Martin R, Gilliam FG, Kilgore M, Faught E, Kuzniecky R. Improved health care resource utilization following video-EEG confirmed diagnosis of nonepileptic seizures. Seizure 1998;7:385-90. [29] Lempert T, Schmidt D. Natural history and outcome of psychogenic seizures: a clinical study in 50 patients. J Neurol 1990;237:35-8. [30] Buchanan N, Snars J. Pseudoseizures: clinical management and outcome in 50 patients. Seizure 1993;2:141-6.