Quality of Life in Psychogenic Nonepileptic Seizures
Address correspondence and reprint requests to Dr. J. P. Szaflarski at University of Cincinnati Medical Center, Department of Neurology, MSB Rm. 4506, ML 525, 231 Albert B. Sabin Way, Cincinnati, OH 45267-0525, U.S.A. E-mail: Jerzy.Szaflarski@uc.edu
Summary: Purpose: Psychogenic nonepileptic seizures (PNESs) are events that alter or seem to alter the neurologic function and, in their appearance, resemble epileptic seizures (ESs). In patients with ESs the psychological and medical aspects of epilepsy greatly influence the health-related quality of life (HRQOL). The relation between these factors and PNESs is not well established. In this study, we compared HRQOL in patients with PNESs with that of patients with ESs.
Methods: We evaluated 105 patients admitted to the Epilepsy Monitoring Unit of University Hospital between January 20, 2001, and January 20, 2002. Only patients with the definite diagnosis of ESs or PNESs were analyzed (n = 85). Patients completed an epilepsy-specific quality-of-life instrument (QOLIE-89), the Profile of Mood States (POMS), and Adverse Events Profile (AEP). We used t tests and regression analyses to contrast HRQOL in PNESs and ESs and to elucidate the main factors associated with HRQOL in patients with PNESs.
Results: In our sample, 45 patients had PNESs, and 40 had ESs. The overall HRQOL and scores on 13 of 19 QOLIE-89 subscales were significantly lower (i.e., worse) in PNES than in ES patients. AEP and scores on five of six POMS subscales also were worse in PNES patients than in ES patients. PNES versus ES diagnosis, POMS depression/dejection, and AEP were significant predictors of HRQOL, jointly explaining 65% variation in HRQOL. The lower HRQOL in PNESs versus ESs was in part explained by depression and AEP.
Conclusions: Patients with PNESs have a lower HRQOL and worse mood problems than do patients with ESs. This disadvantage is primarily due to depression and medication side effects, although these factors influence QOL in much the same way in PNES and ES patients. These baseline HRQOL data on patients with PNESs can be used to evaluate the effects of treatment in this patient population.
The concept of health-related quality of life (HRQOL) is very difficult to describe in everyday life. In the clinical setting, physicians caring for patients with epilepsy monitor seizure frequency and adverse effects of medications, but they may not specifically or systematically assess HRQOL. Therefore several measures were developed to complement the traditional outcome evaluations and to assess other dimensions of HRQOL (1–3), which include the ability to drive, freedom from bothersome medication side effects, participation in social activities, independence, successful employment, seizure unpredictability, or social embarrassment (4,5). In patients with epilepsy, HRQOL measures improve with attaining seizure freedom, independent of the treatment type used to achieve the seizure freedom (6–12). The data regarding HRQOL in psychogenic nonepileptic seizures (PNESs) are limited (13,14).
PNESs are defined as events that alter or seem to alter the neurologic function and, in their appearance, resemble ESs. They are difficult to diagnose solely on the basis of clinical criteria and a normal interictal EEG (15–18). Because clinically PNESs are very similar to ESs, it is reasonable to assume that the effects of PNESs on HRQOL are similar to the effects ESs have on patients with pharmacoresistant epilepsy. The prevalence of PNESs is ∼2–33/100,000 of the general population and between 10 and 58% of the patients with intractable epilepsy seen in specialized epilepsy centers (15,19). The incidence of definite PNESs is 1.4–3.03/100,000 per year (20,21); thus the number of patients with PNESs could be as high as 300,000 to 400,000 (22). Because PNESs are resistant to the standard medical therapies used for the treatment of ESs, the financial cost of their treatment may be very high (23). The direct lifetime cost of undiagnosed PNESs may equal that of intractable epilepsy, which for 1995 was estimated to be as high as $231,432 per patient (23). As documented by several studies, the appropriate diagnosis and management of PNESs can lead to resolution of PNESs in 19–52%, to improvement in 75–95% of patients (13,24–26), and decrease of health care utilization by 69–97%(22).
The majority of PNES research has focused on the clinical characteristics of PNESs or on predicting the diagnosis based on MMPI and clinical history data (16,27–29). Other studies evaluated the association between PNESs and psychological or sexual trauma or other psychiatric comorbities (30,31), outcome after the diagnosis of PNESs (13,24–26), or the incidence of PNESs (20,21). Only two studies evaluated HRQOL in patients with PNESs (14,32).
Because PNESs are common (19,20), effective therapies should be identified. In the meantime, more knowledge is needed about the demographic, social, and psychological factors leading to and associated with this illness. The objective of this study was to compare HRQOL in epilepsy and PNESs. We asked whether patients with PNESs have the same, lower, or higher HRQOL than do patients with intractable epilepsy. We also asked whether the medical and psychological aspects of epilepsy that are known to influence HRQOL in patients with intractable epilepsy (e.g., depression or medication side effects) explain the difference in HRQOL between the two groups of patients. If so, how much explanation do they provide?
We evaluated all patients admitted to the Epilepsy Monitoring Unit (EMU) at the University Hospital in Cincinnati between January 20, 2001, and January 20, 2002, who were eligible and who agreed to participate in the study. The data are based on medical charts and self-administered questionnaires filled out during the stay in the EMU. Patients were eligible to participate if they were aged 18 years or older, had no significant mental handicap, and were able to and completed the questionnaires on their own. The study was approved by the Institutional Review Board of the University of Cincinnati.
Study-related testing was usually performed on the day of admission, before the patient was made aware of the final diagnosis (ESs, PNESs, PNESs+ESs, other). Clinicians had no access to the patient's responses to the questionnaires when making the final diagnosis. Although the clinical features and neuropsychological profile of the patient may by suggestive of the diagnosis of PNESs, the final diagnosis can be confirmed only by simultaneous video recording of the clinical behavior and with EEG (15,16,18). Therefore the diagnosis of PNESs versus ESs in all patients included in this study is based on the results of the prolonged video-EEG monitoring (PVEM), the clinical characteristics of the events recorded, and other supporting evidence (e.g., neuropsychological testing/MMPI). On completion of PVEM, we divided patients into four groups: definite PNESs, possible PNESs, possible epilepsy, and definite epilepsy. As previously, we used the following criteria: definite PNESs is defined as habitual spells recorded on video-EEG without an ictal EEG discharge and normal background alpha rhythm during the events; possible PNESs is defined as no events recorded and continuously normal interictal EEGs; possible epilepsy is defined as no events recorded with interictal epileptiform discharges; definite epilepsy is defined as recorded stereotypic events with an ictal EEG discharge (17,18,20,29). Only the data on patients with definite PNESs and definite epilepsy were used in this study.
We measured HRQOL with QOLIE-89, which is the most widely used and most comprehensive instrument specifically developed for evaluation of HRQOL in patients with epilepsy (2). Because ESs and PNESs may be clinically very similar, QOLIE-89 also appears to be an excellent tool for evaluation of patients with PNESs (32). Higher scores indicate better QOL. Our measure of psychological status is POMS, also a widely used instrument that contains 65 items in six dimensions (depression/dejection, tension/anxiety, fatigue/inertia, confusion/bewilderment, anger/hostility, and vigor/activity). Higher scores point to worse mood problems, except for the vigor/activity scale, in which the relation is reversed (1). AEP is a 19-item inventory that assesses medication side effects (3). Higher scores indicate more severe medication adverse effects.
Demographic characteristics and medical history were collected through review of the medical chart and self-reports. Frequency of seizures was measured as the self-reported number of seizures per week. The number of the previously used antiepileptic drugs (AEDs) was obtained from the initial clinical interview and previous records. Magnetic resonance imaging (MRI) data were obtained from patient records (official report) and through a direct review of the imaging study. Psychiatric history—depression diagnosed by a primary care physician, neurologist, or psychiatrist, or other psychiatric illnesses [e.g., anxiety disorder, panic attacks, or posttraumatic stress disorder (PTSD)] diagnosed by a psychiatrist—was determined based on previous records (medical chart) and a self-report during the EMU admission; the measure is coded 0 for “no psychiatric history” and 1 for “psychiatric history.” Length of stay is the number of full days spent in EMU. Comorbidities are defined as chronic health problems (including psychiatric problems) other than epilepsy, psychogenic seizures, and depression. In the descriptive statistics, we report the mean number of comorbidities; in the multivariate analysis, we use a binary variable coded as 0 for “none” and 1 for “one or more.” Because comorbidities overlap with psychiatric history, only comorbidities were used as a predictor in regression analyses.
We analyzed a combined epilepsy–PNES sample to preserve the full sample size and to be able to test directly for differences between the two groups. Except for two missing MRI values, we replaced other missing data according to commonly used statistical procedures. One missing value on psychiatric history was replaced based on the physician's assessment of the patient's mental health on admission to EMU. Two values missing on QOLIE-89 overall score because of nonresponses on some items within QOLIE-89 instrument were replaced by using a diagnosis-specific mean (mean for the ES group or mean for the PNES group).
As the first step in our analysis, we compared the ES and PNES groups in terms of demographic characteristics, health/medical conditions, and HRQOL and its correlates (POMS, AEP). To examine the differences between the two groups, we computed χ2 tests for categoric variables and t tests for continuous variables; we made no correction for multiple comparisons because of the exploratory nature of our study. As the second step, we applied hierarchic multiple regression to predict and explain QOL in our sample. This analysis is based on a simple assumption that HRQOL is affected by a person's mental or physical disability, and that factors associated with the disease also may influence HRQOL. Axiomatic assumptions of this kind generally underlie research on HRQOL (33,34). Our outcome variable was the QOLIE-89 overall score. We first ran a simple equation with definite diagnosis (ES/PNES) as the only predictor to test a hypothesis about the difference in the HRQOL score between the ES and PNES groups, adjusted for presence of comorbidities (as they also may affect HRQOL). Next we tested an equation that also included POMS depression/dejection scale and AEP to estimate the difference in HRQOL between the ES and PNES patients. This model was then adjusted for gender, age, and the other medical history variables (abnormal MRI, duration, seizure frequency, AEDs). To test for differential effects of psychological status on HRQOL for the PNES and ES patients, we finally added to the model (one at a time) interactions between definite diagnosis and POMS depression/dejection and between definite diagnosis and AEP. Each consecutive regression model was compared with the previous one and retained based on (a) significant effects of the added variables, and (b) a significant increment-to-R2 test (35); if these criteria were not met, the simpler model was retained. The 0.05 alpha level was used in this study.
The practical question before us was to predict HRQOL in patients with certain characteristics. The key advantage of regression analysis is the ability to improve estimates of an outcome variable in a population by using the regression line instead of merely reporting the sample mean of the variable. That is, prediction of an outcome is improved by taking into account the variable's correlation with other factors (the resulting estimates will be closer to the true values of the outcome). Furthermore, by introducing predictor variables gradually into the model, it is possible to compare their relative contribution to the variation in the outcome variable (based on the R2 values and the increment-to-R2 test) (35).
We must caution the reader that although we attempted to identify factors predictive of HRQOL and we describe our results in terms of regression, or effect, coefficients, it is best to interpret the coefficients as measures of association between the outcome variable (HRQOL) and the predictor variables and not as measures of causation. We assume certain causal paths in our theoretical model, but building a true causal model and testing for possible two-way effects between the outcome variable and the predictor variables lies beyond the scope of this study.
Of the 141 patients admitted to the EMU, 105 were willing and able to participate in the study. Of the enrolled patients, 45 were diagnosed with intractable epilepsy, and 40, with PNESs. In the remaining 20 patients, the diagnosis was either unclear (no events recorded in 12 patients) or they received a mixed diagnosis (eight patients). Thirty-six patients did not participate in the study, of whom 20 had a mental handicap (16 had the diagnosis of mental retardation, and four had decreased mental status during the admission to the EMU related to the medication side effects or concurrent medical illness); eight refused to participate; three were younger than 18 years; three did not complete the questionnaires before the diagnosis was established; and two were excluded because of nonresponses to the questionnaires. One patient with presumptive diagnosis of frontal lobe complex partial seizures was included in the category of possible seizures because of lack of any EEG changes during the stereotypical events. The clinical characteristics of the events were suggestive of frontal lobe complex partial seizures, which are sometimes difficult to distinguish from the PNESs (16,18).
We present the characteristics of the patients with the definite diagnosis of epilepsy and the definite diagnosis of PNESs in Table 1. Patients with epilepsy and with PNESs had a similar mean age (36 and 38 years). PNES patients had a later onset of the disease, used fewer AEDs before evaluation in the EMU, had a higher frequency of events per week, and were less likely to have abnormal MRI. The PNESs group had more women and a higher frequency of psychiatric problems (e.g., depression). Little difference between the ES and PNES groups was observed in terms of the average number or presence of other comorbidities (e.g., cardiovascular diseases, diabetes, or other chronic conditions).
Table 1. Demographic and medical characteristics of patients with definite epilepsy (ES) and definite psychogenic, nonepileptic seizures (PNES) monitored in EMU (n = 85)
|Females||60.0|| ||85.0|| || ||6.532||0.011|
|Abnormal MRIb||52.3|| ||15.4|| || ||12.376||<0.001|
|Psychiatric history||35.6|| ||60.0|| || ||5.079||0.024|
|Age (yr)|| ||36.0 (35.9)|| ||37.6 (10.6)||−0.732|| ||0.466|
|Age at onset (yr)|| ||19.6 (13.7)|| ||29.4 (13.3)||−3.340|| ||0.001|
|Frequency of seizures/week|| ||7.5 (15.0)|| ||16.8 (23.8)||−2.181|| ||0.032|
|Number of AEDs|| ||4.6 (2.86)|| ||3.4 (1.9)||2.301|| ||0.024|
|Length of stay (days)|| ||4.4 (1.8)|| ||3.0 (1.9)||3.764|| ||<0.001|
|Number of comorbidities|| ||1.8 (1.7)|| ||2.2 (1.5)||−1.227|| ||0.223|
Table 2 shows the differences in HRQOL, mood states, and adverse events between the ES and PNES groups. The overall quality of life was 16.6 points lower in the PNES group than in intractable epilepsy patients (mean, 41.0; SD, 15.6 vs. mean, 57.6; SD, 18.2; p < 0.001). Patients with PNESs scored lower on all QOLIE-89 subscales compared with patients with intractable epilepsy, with a difference of ≥15 points on eight of 19 subscales (p < 0.05). PNES patients also had worse mood problems than did ES patients. The POMS scores were higher in patients with the diagnosis of PNES for all subscales except for the vigor/activity subscale, indicating more mood problems in this group. Scores for the vigor/activity scale point toward higher vigor, ebullience, and energy in patients with epilepsy versus those with PNESs. Although patients with ESs were treated on average with a larger number of AEDs than were PNES patients (Table 1), AEP score was higher for the PNES patients (Table 2; p = 0.002), indicating that these patients have more side effects than do epilepsy patients.
Table 2. Comparison of average HRQOL scores (QOLIE-89 overall score, QOLIE-89 subscales, POMS subscales, and Adverse Events Profile) in the epilepsy and PNES patients (n = 85)
|QOLIE-89 (overall score)||57.6||(18.2)||41.0||(15.6)||4.451||<0.001|
|Role Limitations: Physical||50.7||(41.3)||19.8b||(27.6)||4.008||<0.001|
|Role Limitations: Emotional||63.8||(41.1)||49.0||(44.4)||1.602||0.113|
|Work, Driving, Social Function||52.0||(24.7)||27.8||(17.6)||5.127||<0.001|
|Change in Health||37.6||(25.2)||32.5||(26.7)||0.898||0.436|
|Adverse events profile (AEP)||43.1||(10.4)||51.1||(12.2)||−3.242||0.002|
The results of the regression analysis are presented in Table 3. We present the unadjusted results because they do not significantly differ from the results adjusted for gender, age, and other medical history. The negative regression coefficient in model 1 (b = –15.49; p < 0.001) confirms the bivariate results (Table 2): the HRQOL score for patients with PNESs is, on average, 15.5 points lower than that for patients with intractable epilepsy, adjusting for comorbidities. This model explains 25% variation in HRQOL (adjusted R2 = 0.25). Model 2 adds POMS depression/dejection and AEP as predictors; as shown, these variables are significantly associated with HRQOL (b = 0.574 and b = 0.569; p < 0.001). Model 2 explains a significantly larger amount of variance in HRQOL of 65% (adjusted R2 = 0.645; increment-to-R2 test = 35.979; p < 0.001). A reduced coefficient for definite diagnosis in model 2 (b = –7.162; p < 0.01) indicates that POMS depression/dejection and AEP explain almost a half of the PNES disadvantage in HRQOL versus epilepsy.
Table 3. Effects of ES/PNES diagnosis, POMS depression/dejection, and Adverse Events Profile on health-related quality of life (QOLIE-89 overall score), adjusted for comorbidities (n = 85)
|Definite diagnosis (ES, 0; PNES, 1)||−15.485a||(3.609)||−7.162b||(2.795)|
|Comorbidities (“none,” 0; “≥1,” 1)||−14.474b||(5.174)||−12.355b||(3.557)|
|POMS: Depression/Dejection (scale)|| || ||−0.574a||(0.111)|
|AEP (scale)|| || ||−0.569a||(0.131)|
|Increment-to-R2 test|| || ||35.979a (df = 2)|
Finally, we tested for the presence of interactions between definite diagnosis and POMS depression/dejection and definite diagnosis and AEP, to see whether different levels of depression/dejection and AEP in PNES and ES patients result in different QOL. We found no significant interaction effects (power > 0.99). Thus our present finding is that the effect of depression (POMS depression/dejection) on HRQOL is not different for patients with PNESs and patients with refractory epilepsy.
In this study we report the results of HRQOL evaluation in patients with PNESs and contrast the results to HRQOL in patients with intractable epilepsy.
The overall QOLIE-89 score for the intractable epilepsy patient group in our study (57.6) was somewhat lower than the data presented by in the original QOLIE-89 publication for the most severely affected patients, but the difference is likely not significant (65.5) (2). It may be related to the inclusion of the patients with the most severe epilepsy in our study (patients admitted to the EMU for presurgical evaluation or differential diagnosis), whereas patients enrolled in the original study were recruited on an outpatient basis and excluded from participation if they had any active and significant psychiatric comorbidities (2). A recent randomized epilepsy surgery trial enrolled patients very similar to our ESs population (12). The presurgical QOLIE-89 scores in that population were between 52.9 and 60.5; thus the score in our study falls in the middle of the data from that trial (12). The age at seizure onset and age at the evaluation were similar in both studies, suggesting that our intractable epilepsy group is similar to the patients with intractable epilepsies evaluated at other centers and is well suited for the analysis intended in this study (12).
Whereas the overall HRQOL in the intractable epilepsy cohort was similar to that in other studies, HRQOL in the cohort of patients with PNESs in our study was much worse (57.6 vs. 41.0; p < 0.001; a 17-point difference). According to a recent study of patients with epilepsy evaluated with the QOLIE-89 instrument, the difference of ≥15 points on QOLIE-89 overall score is of clinical significance (with a p < 0.05) and is not related to chance or measure error (36). The low overall HRQOL in the PNES patients in our sample is confirmed by the data from all QOLIE-89 subscales where the patients with PNESs scored significantly lower (and by >15 points) than the patients with ESs on eight of 19 scales (Table 2). Our findings are in agreement with the results of a study comparing the quality of life between ESs and PNESs by using QOLIE-31 instrument (14). The results also indicate that patients with PNESs have lower HRQOL than do epilepsy patients who are usually severely impaired by intractable seizures. This is despite less evidence for organic brain disorders in patients with PNESs, as suggested by lower incidence of MRI abnormalities in this group (Table 1; p < 0.001). Thus our findings are not only statistically significant and consistent with earlier reports, but they also are clinically important.
In addition to overall HRQOL, we compared patients with PNESs and ESs on specific subscales of QOLIE-89 that tackle social and psychological aspects of QOL. A prior study showed that multiple psychological concerns including social embarrassment, mood/stress problems, stigma of epilepsy, and social/professional discrimination affect HRQOL in patients with intractable epilepsy (11). In our study, we tested whether these factors also play a role in HRQOL of patients with PNESs. We found that PNES patients have significantly more severe psychological problems than do ES patients. Besides more severe and frequent depression (POMS depression/dejection), PNES patients have more anxiety and bewilderment (POMS tension/anxiety and confusion/bewilderment subscales) and less energy/ebullience (POMS vigor/activity subscale), which can represent a self-report of cognitive inefficiency, possibly a by-product of anxiety or related states (1). Although some authors reported that seizure severity is associated with low self-esteem and anxiety in patients with intractable epilepsy (37), our analysis did not confirm the association between the duration and number of seizures per week and lower HRQOL. However, HRQOL in PNESs can still be affected by seizures, for example, if there is a threshold effect of a small number of seizures that was missed because all were refractory. Overall, our findings confirm the effect of a chronic and intractable disease on HRQOL in PNESs to an even larger extent than in intractable epilepsy.
Even though a majority of epilepsy patients lead a normal emotional and cognitive life, neurobehavioral abnormalities are found in a many of them, especially in patients with intractable epilepsies (4,38,39). These psychiatric conditions may be related to the localization of the epileptic focus, seizure spread, neuropathologic damage related to or preceding the epilepsy onset, or low educational and socioeconomic status (38). Multiple psychiatric conditions are known to occur in epilepsy patients, including depression, anxiety disorders, psychoses, Geschwind syndrome, and dissociative disorders (38,40). AEDs also can cause psychiatric adverse events such as psychoses, affective syndromes, encephalopathies, or depression (41). As mentioned previously, HRQOL in epilepsy may be related in part to seizure frequency and severity, medication side effects, and social problems associated with unemployment and lack of independence. HRQOL also is related to the presence or absence of psychiatric and other comorbidities. All of these conditions coexist in our cohort of PNES patients. When controlled for these conditions, the QOL in patients with PNESs is still significantly lower in that group, suggesting that, although the cognitive and mood problems associated with PNES and AEDs significantly reduce the QOL in PNESs, they are not the sole contributors. Other medical and psychosocial factors (e.g., independence or history of abuse) not included in this study also may shape the HRQOL of patients with PNESs (11,31).
Although depression is present in ∼25% of the epilepsy patients in community studies, 67% of the outpatient intractable epilepsy population is estimated to have depression (42). Only 36% of the ES group was diagnosed with a psychiatric comorbidity (Table 1). The POMS depression/dejection score in our ES cohort (13.9) is only slightly higher than the normative data for the healthy adult population aged between 18 and 65 years (men, 8.3; women, 10.2), but this could be related to underreporting because the demographic data in our study were obtained by an interview only (43). The history of psychiatric comorbidities is higher in the PNES group (p = 0.014), with the POMS depression/dejection score being higher (13.9 vs. 21.2; p = 0.013), indicating higher incidence and severity of mood problems in the PNES group. Unfortunately, we cannot infer from our study whether the severe mood problems are related to the concomitant diagnosis of PNESs or whether the preexisting psychiatric problem leads to the development of PNESs.
Although prior data on HRQOL in PNESs are limited, we believe that our ES–PNES comparison and its results are valid. First, life experiences of PNES and ES patients overlap in terms of physical symptoms and social and psychological consequences of disease. Therefore QOLIE-89, a valid and reliable tool to measure HRQOL in epilepsy patients, also can be applied to measure HRQOL in patients with PNES. We assume that our data are representative of both cohorts of patients: ESs and PNESs. The mean QOLIE-89 overall and subscale scores for epilepsy patients in our sample are similar to the data reported in the original QOLIE-89 study (2). We assume our whole sample, including the PNESs subsample, to be representative of the population. These theoretic and methodologic considerations strengthen our findings.
Our study shows that the diagnosis of PNESs is associated with worse HRQOL than the diagnosis of epilepsy. The study also shows that medication side effects significantly reduce HRQOL in patients with PNESs. Therefore we recommend that patients with suspected PNESs should be referred for further evaluation as soon as the suspicion of PNESs is raised. We also recommend weaning the AEDs whenever the diagnosis of PNESs without concomitant ESs is made. We expect that elimination of the AEDs after the accurate diagnosis would improve HRQOL in patients with PNESs because many of their concerns revolve around the medication side effects, fear of seizures, and limitations related to uncontrolled spells.
Presented in part at the Annual Meeting of the American Epilepsy Society, Philadelphia, PA, November 30–December 5, 2001.