- Top of page
- Patients and Methods
- Supporting Information
Purpose: Psychogenic nonepileptic seizures (PNES) continue to represent a serious diagnostic challenge for neurologists. Video–electroencephalography (EEG) studies have provided detailed knowledge of the spectrum of visible PNES manifestations. However, little is known about how patients or seizure witnesses experience PNES, although many diagnoses in seizure clinics are made on the basis of self-reported information rather than video-EEG observations. This study describes the range of PNES manifestations as they are reported by patients or seizure witnesses.
Methods: Three hundred eight candidates for this study were consecutively diagnosed with PNES on the basis of video-EEG recordings of habitual seizures involving impairment of consciousness without epileptic ictal EEG activity at the Royal Hallamshire Hospital in Sheffield and the National Hospital for Neurology in London, United Kingdom. One hundred patients responded to a postal questionnaire and participated in this study. Eighty-four of the questionnaires completed by patients were accompanied by questionnaires completed by seizure witness. The patient questionnaire contained 12 demographic and clinical questions and the 86-item Paroxysmal Event Profile (PEP), asking patients to rate statements about their attacks on a five-point Likert scale (“always,”“frequently,”“sometimes,”“rarely,”“never”). The Paroxysmal Event Observer (PEO) questionnaire uses 34-items with the same Likert scale. The PEP questionnaire includes inquiries about symptoms of panic or dissociation as well as symptoms previously found to distinguish between generalized tonic–clonic seizures and syncope or thought to differentiate between epilepsy and PNES.
Key findings: The item-by-item analysis revealed the inter- and intraindividual variability of PNES experiences. The majority of patients with PNES reported some phenomena, which have traditionally been attributed to epilepsy (such as seizures from sleep, experiencing a rising sensation in their body, postictal myalgia). Although most PNES were experienced as striking without warning and reported to cause loss or impairment of consciousness, most patients also reported seizure warnings in at least some of the seizures. Despite the clinical heterogeneity apparent from these findings, a correlation matrix showed that symptoms were not randomly distributed. Significant correlations were seen between duration of seizures and seizures from reported sleep (r = −0.28, p = 0.006), seizure-related motor activity and seizures from reported sleep (p = −0.48, p < 0.001), flashbacks and anxiety (p = 0.44, p < 0.001) or dissociation (p = 0.66, p < 0.001), and anxiety and dissociation (r = 0.53, p < 0.001). The comparison of similarly worded items on the PEP and PEO questionnaires showed that witnesses were more often aware of seizure triggers and a relationship between PNES and emotional stress than were patients (p = 0.001/p < 0.001).
Significance: These findings based on the self-report of patients with well-characterized PNES and witnesses of their seizures demonstrate why it can be difficult to distinguish descriptions of PNES from those of epilepsy on the basis of factual items. The differences between patient and witness reports suggest that clinicians have to take note of the source of information they use in their diagnostic considerations. The intra- and interindividual variability of reported PNES manifestations demonstrates the clinical heterogeneity of PNES disorders. The positive correlation of symptoms of dissociation and anxiety in these patients may reflect psychopathologic differences between subgroups of PNES patients.
Psychogenic nonepileptic seizures (PNES) are episodes that superficially resemble epileptic seizures but are not caused by epileptic discharges in the brain. They are paroxysmal events characterized by an impairment of self-control associated with a range of motor, sensory, and mental manifestations, which represent an experiential or behavioral response to distress. The overwhelming majority of PNES are considered as beyond patients’ voluntary control (Gates et al., 1998; LaFrance & Devinsky, 2002), and can be categorized as manifestations of conversion or dissociative disorder (Reuber, 2009).
PNES are one of the most common differential diagnoses of epileptic seizures: Between 12% and 18% newly presenting to neurologists with blackouts, and about one fourth of the patients referred to a typical U.S. epilepsy monitoring unit have PNES rather than epilepsy (Kotsopoulos et al., 2003; Benbadis et al., 2004; Hitiris et al., 2005).
The visible manifestations of PNES have been studied in great detail since the introduction of video–electroencephalography (EEG) monitoring 30 years ago (Reuber, 2008). Although there is no single feature that distinguishes reliably between PNES and epileptic seizures, a number of visible observations make a diagnosis of PNES likely (Avbersek & Sisodiya, 2010).
Our understanding of the subjective symptoms of PNES is more limited. It has been suggested that many patients experience physical symptoms of panic or hyperventilation during their seizures—perhaps without mental symptoms of anxiety (Snyder et al., 1994; Vein et al., 1994; Witgert et al., 2005; Goldstein & Mellers, 2006). Even in the absence of panic symptoms, most patients experience their seizures as confusing and beyond their control (Green et al., 2004; Thompson et al., 2009). A study comparing the use of metaphors found that patients with epilepsy were more likely to conceptualize their seizures as a hostile, external agent acting of its own volition, whereas patients with PNES preferred to talk of their seizures as a space or place into which they travel (Plug et al., 2009a). In keeping with this (and unlike patients with epilepsy), patients with PNES showed resistance to the doctor’s use of the label “attack” for their ictal experiences (for instance, by not using the term in an answer although the doctor had used it in his question) (Plug et al., 2009b). Several studies suggest that (if present) impairment of consciousness tends to be less profound in PNES than in complex partial epileptic seizures (Reuber & Kurthen, 2011).
There is also a relative dearth of research into what patients and witnesses report about seizure manifestations, although studies relating such reports to observations made during monitoring with video-EEG have revealed significant discrepancies (Rugg-Gunn et al., 2001; Heo et al., 2008; Syed et al., 2008), suggesting that some ictal observations that are useful diagnostically if seen on a video-recording (such as the initial eye-opening characteristic of epileptic seizures) (DeToledo & Ramsay, 1996; Chung et al., 2006) do not help clinicians without video confirmation (Syed et al., 2010). This does not mean that information derived from patient or witness reports cannot be used for the diagnosis; however, it demonstrates that what people report about seizures should be studied directly rather than predicted on the basis of video-EEG studies.
The present study is intended to improve our understanding of the phenomenology of PNES as they present to physicians talking to patients in the outpatient clinic. It describes seizure manifestations in a large cohort of securely diagnosed patients. By combining a descriptive approach, a correlation matrix, and comparisons of patient and witness statements, this study aims to improve our understanding of how PNES are experienced and aid the diagnostic process by making physicians more aware of what patients and witnesses are likely to tell them about the seizures.
- Top of page
- Patients and Methods
- Supporting Information
Of the 308 patients contacted, 100 (32%) responded to the questionnaire and took part in this study; 71% of the respondents were women. Respondents’ ages ranged from 19–81 years [mean 41.7, standard deviation (SD) 13.6]. The mean age of the nonrespondents was 37 years (SD 12.4); 71% of nonrespondents were female. There were no significant differences between respondents and nonrespondents in terms of age and gender ratio. Eighty-four (84%) of the PEP questionnaires returned were accompanied by PEO questionnaires completed by witnesses. The witnesses had known the patients for a mean of 16.9 years (SD 15.8). There were no significant demographic differences between participants from London or Sheffield. A median of 5% (range 0–12%) of items on the PEP and a median of 4% (range 0–25%) on the PEO questionnaire were left blank and filled with medium item scores.
Respondents had first developed seizures a mean of 12.9 years (SD 12.5) before the study. Ten percent of respondents had not had a seizure over the year preceding the study, 7% had 1–5 attacks, 40% had 6–50, and 43% had >50. Twenty-four percent of the participants had never been admitted to hospital acutely with a prolonged seizure, 17% had been admitted once, 31% two to five times, and 28% more than five times. Sixteen percent of respondents reported at least one admission to an intensive care unit with an attack. The same number reported attacks, blackouts, or epilepsy in immediate family members. Twenty-five percent of patients stated that they had experienced a head injury with loss of consciousness (nor necessarily before the onset of seizures), 1% a history of meningitis, and 4% a history of febrile seizures in childhood. Table 1 shows the replies to some of the items included in the PEP (1a) and PEO questionnaire (1b); Appendix S1 is a table with the responses to all questions.
Table 1. (a) Respondents replies to some of the items on the PEP and (b) PEO questionnaires
|Item||Always (%)||Frequently (%)||Sometimes (%)||Rarely (%)||Never (%)|
| I am aware of a trigger for my attacks||10||4||44||9||31|
| My attacks are associated with emotional stress||8||9||47||13||23|
| My attacks come on when I am asleep||3||8||43||10||36|
| My attacks come on out of the blue without any warning||43||24||25||2||6|
| In my attacks I feel a rising sensation in my body||14||15||28||5||38|
| My attacks cause a bad taste in my mouth||14||10||16||6||54|
| In my attacks I feel sick||5||6||24||24||41|
| In my attacks I have a sense of feeling I’ve seen something before when I have not||0||5||14||8||73|
| During my attacks memories seem to flash into my mind||2||11||14||6||64|
| During my attacks I have no idea what is happening around me||30||25||26||5||14|
| In my attacks I am conscious but I can’t react to things||28||20||28||1||23|
| During my attacks I feel very frightened||18||17||37||8||20|
| I wake from my attacks with a cut tongue||1||4||20||13||62|
| After my attacks I find that I have burned myself||4||1||7||8||80|
| After the attacks my muscles ache||36||26||22||9||7|
| There are triggers form the attacks||11||14||43||11||21|
| The attacks come on out of the blue without any warning||49||27||17||5||2|
| Attacks cause falls||38||23||26||6||7|
| The attacks involve a complete loss of consciousness or blackout||56||19||8||5||12|
| Loss of consciousness, loss of awareness or loss of speech last over 1 min||45||18||19||2||15|
| Shaking of the arms and legs goes on for more than 1 min||26||26||20||4||24|
| The attacks involve chewing, smacking, or licking movements of the mouth and lips||21||12||16||12||39|
| The attacks involve fiddling, picking, or fumbling movements of the hands||17||18||11||17||38|
| The eyes are mostly open in the attacks||21||13||32||10||24|
| In the attacks the head moves rapidly from side to side||14||16||30||12||29|
| The head turns to one side in the attacks||6||8||21||23||22|
| The mouth is tightly shut during the attacks||7||16||52||11||14|
| The attacks involve crying or weeping||2||4||24||8||62|
| The attacks involve arching of the back||17||11||14||26||32|
| The attacks involve trembling||27||31||21||1||19|
| The attacks involve violent shaking of arms and legs||24||20||17||17||23|
| The arms and legs are rigid in the attacks||11||13||33||16||24|
| The arms and legs are limp in the attacks||12||16||29||13||31|
| Attacks look a bit like normal sleep||8||7||18||8||58|
| The attacks involve frothing from the mouth||6||2||11||8||72|
| The skin and lips look blue in the attacks||8||11||24||13||44|
| I am immediately recognized after the attacks||33||23||31||6||1|
Statistically significant correlations of clinically defined dimensions are shown in Table 2. All dimensions excluding duration (n = 93) and ictal motor activity (n = 84) are based on 100 patient responses. There were no significant correlations between the seizure frequency category, hospitalization frequency, history of intensive care admission, and any of the other dimensions included in the table. The respondents’ ages at PNES onset and at the time of completion of the questionnaire correlated positively with each other. The age at onset correlated negatively, and the age at completion of the questionnaire positively with the duration of the disorder.
Table 2. Correlation matrix of clinically defined dimensions
| ||Duration||Dissociation||Flashbacks||Anxiety||Motor||From sleep|
|Duration||X||n. s.||n. s.||n. s.||n. s.||−0.28 (p = 0.006)|
|Dissociation||X||X||0.66 (p < 0.001)||0.53 (p < 0.001)||n. s.||n. s.|
|Flashbacks||X||X||X||0.44 (p < 0.001)||n. s.||n. s.|
|Anxiety||X||X||X||X|| || |
|Motor||X||X||X||X||X||−0.48 (p < 0.001)|
Table 3 shows the comparison of responses to three key questions about seizures on the PEP and PEO questionnaires, which differed only in terms of the personal pronoun used (these were the only questions that differed in this way). Witnesses thought that seizure triggers could be identified significantly more often than reported by patients and that more seizures were related to emotional stress. A question about ictal impairment of consciousness was not worded identically on the PEP and PEO questionnaires, but responses to two closely related questions suggested that witnesses reported more frequent events with complete loss of consciousness than patients did (PEP: “during my attacks I have no idea what is happening around me”/PEO: “the attacks involve complete loss of consciousness or blackout”).
Table 3. Comparison of patient and witness responses (ordinal scale: 5 = always, 4 = frequently, 3 = sometimes, 2 = rarely, 1 = never)
|Question||Mean patient response category (SD)||Mean witness response category (SD)||Significance (MWU test)|
|I am (there is) a trigger for the attacks||2.5 (1.3)||2.8 (1.3)||0.001|
|My (the) attacks are associated with emotional stress||2.7 (1.2)||2.8 (1.3)||<0.001|
|My (the) attacks come on out of the blue without any warning||4.0 (1.2)||4.2 (1.0)||n. s.|
- Top of page
- Patients and Methods
- Supporting Information
The range of patients’ responses to the PEP questionnaire reveals the experiential heterogeneity of PNES. Almost one half of the respondents always experience their attacks as striking out of the blue and without any warning, but just over 50% report being aware of triggers for, at least, some of their seizures. More than two thirds of patients regularly feel frightened in their attacks but one in five never does. Whereas nearly one third of patients report being completely unaware of their surroundings in all attacks, a similar number remains conscious of their environment but are unable to react to it. The majority of patients experience either state in different attacks. The responses from the seizure witnesses to the PEO questionnaire also paint a varied picture. Most state that seizure-related falls, shaking, or limpness of the limbs and rapid reorientation occur in some but not all attacks.
Patients’ and witnesses’ responses to several items demonstrate the phenomenologic overlap between epilepsy and PNES. This illustrates why it can be so difficult to differentiate between these two seizure disorders on the basis of factual items in patient or witness accounts (Syed et al., 2010), and suggests why the diagnosis of PNES is typically delayed by several years (Reuber et al., 2002). The majority of patients with PNES report some phenomena, which have traditionally been attributed to epilepsy (such as seizures from reported sleep, feeling sick or experiencing a rising sensation in their body, postictal myalgia). Similarly, the majority of witnesses describe open eyes, behaviors suggestive of oral and manual automatisms and cyanosis (in at least some attacks). Some features thought to be more common in PNES than epilepsy (including ictal weeping and seizures looking like normal sleep) are rarely experience or observed. Confusingly, rapid side-to-side head movements (described as indicative of PNES in video-EEG studies) (Gates et al., 1985; Groppel et al., 2000) are reported as less common by the observers of PNES than are unilateral head turns (which occur in many focal epileptic seizures). Somewhat more helpfully, déjà vu, ictal tongue biting, burns, and frothing from the mouth (pointing to a diagnosis of epilepsy) were never experienced by the majority of patients. However, even these items were reported by significant minorities of PNES patients or witnesses.
Although the item-level analysis illustrates the inter- and intraindividual heterogeneity of the phenomenology of PNES, the correlational matrix reveals that PNES symptoms are not randomly distributed. A longer duration of the seizure disorder was associated with a lower proportion of seizures from (self-reported) sleep. This observation suggests that sleep-related seizures are particularly likely to be reported in the earlier phase of PNES disorders, when the diagnosis may still be uncertain. Unlike epileptic seizures, nocturnal PNES typically occur from “pseudosleep” (characterized by the EEG appearances of wakefulness) (Benbadis et al., 1996), although apparently “sleep”-related PNES are self-reported by one half of PNES patients (Duncan et al., 2004). The witnesses’ reports revealed that the seizures of patients claiming to have had more seizures from sleep were characterized by less motor activity (rigidity and shaking).
Significant correlations were also found between scores reflecting symptoms attributed to different forms of psychopathology in the current diagnostic manuals of mental disorders. There was a strong association between symptoms of dissociation (depersonalization, derealization, and amnesia) and flashbacks. This may not be surprising given their co-occurence in conditions such as posttraumatic stress disorder or borderline personality disorder, and the links of all of these symptoms with a previous history of trauma. The positive correlation between the symptoms of anxiety and dissociation is more interesting. Whereas some patients report few anxiety and dissociative symptoms, other patients frequently experience both types of symptoms. PNES have previously been interpreted as a dissociative response to physical arousal, or as “panic attacks without panic” (paroxysmal physical symptoms of panic without cognitive symptoms of anxiety) (Goldstein & Mellers, 2006).
Our data based on patients’ experience rather than objective measures of physiologic arousal cannot refute this model. However, we note that the majority of respondents to our questionnaire “never” experienced any of the items captured on the DES-Taxon as characteristic of “pathological dissociation.” Furthermore, the correlation matrix suggests that there are markedly different forms of PNES experience. This means that it is questionable whether one explanatory approach fits all cases in this context. It interesting to speculate whether the combination of high levels of anxiety and features of dissociation (especially flashbacks) relate to a particular etiology (one previous study showed that this combination of symptoms was more likely in patients with antecedent sexual abuse) (Selkirk et al., 2008). Of course, there may be a difference between having symptoms of anxiety/dissociation and being aware of or reporting them (Reuber & Kurthen, 2011). It is also possible that the adaptation of the DES-taxon used here fails to measure the relevant form of dissociation (Lawton et al., 2008). However, neither the descriptive analysis of questionnaire items relating to phenomena typically interpreted as dissociative nor the correlation matrix lends support to the idea that patients with PNES experience dissociative symptoms instead of anxiety or that patients increasingly dissociate from unpleasant symptoms of arousal as their disorder becomes more chronic.
There were interesting differences between patients’ own perceptions of their seizures captured by the PEP and witnesses’ observations recorded on the PEO questionnaires. Patients were less often aware of seizure triggers and emotional stress than witnesses were. This finding may reflect the alexithymia that characterizes many patients with PNES (Bewley et al., 2005), or their cognitive bias against a psychological explanation for their seizures (Stone et al., 2004). The replies to the PEP and PEO questionnaires also indicate that witnesses may overestimate patients’ degree of unconsciousness.
This study has a number of limitations. There were at least two sources of possible selection bias: first, all patients described in this study had presented to specialist units and had video-EEG documented seizures. Responses in these patients may differ from the one third of patients with PNES who will fail to have seizures during even prolonged video-EEG monitoring. Second, only one-third of the patients invited to participate in this study returned the PEP questionnaire and one in six of the respondents could not provide a matching PEO. Although we did not find any demographic markers of a responder bias, it is possible that the subgroup of responders is not characteristic of all patients with PNES. Patients who are unable or unwilling to say much about the symptomatology of their seizures may have been underrepresented in the group of responders to the survey. It is important to recognize that (at this point) the findings in this patient group with chronic PNES disorders and their friends, family members, or carers cannot be generalized to patients presenting with seizures of recent onset. However, given that most diagnoses of PNES are made only after patients have been treated for epilepsy for many years (Reuber et al., 2002), the results may still be directly relevant in the diagnostic process. It is also conceivable, that the responses of patients or observers to the questionnaires (especially with regard to triggers or reported emotional symptoms) were influenced by the fact that they had received the diagnosis of PNES and that their neurologist had explained this condition to them. Furthermore, we have no other information about patients’ state of mind while completing the questionnaire other than the responses provided on the questionnaire. Most importantly, we do not know whether some of the responses (for instance the report of anxiety or dissociative symptoms) may have been influenced by comorbid depression. Lastly, we should point out that small differences in how a question is put can have big effects on how it is answered. It is possible that some replies would be different if questions were asked face-to-face in the clinic rather than included in a questionnaire.
Despite these drawbacks, a number of conclusions can be drawn: Although most PNES are experienced as striking without warning and are reported to cause loss or impairment of consciousness, there is considerable intra- and interindividual variability. Our analysis based on the reports of seizure manifestations by patients and witnesses, therefore, offers further support for the clinical heterogeneity of PNES disorders described previously on the basis of video-EEG observations (Gates et al., 1985). Future comparisons with responses from patients with epilepsy will need to test whether the intraindividual variability observed here is a diagnostic pointer to PNES rather than epilepsy. The positive correlation of symptoms of dissociation and anxiety in these patients suggests that there may reflect important psychopathologic differences between subgroups of PNES patients. Placing too much emphasis on patients’ or witnesses’ responses to individual factual items may risk incorrect diagnoses. We have previously shown that it is diagnostically relevant to consider whether patients have volunteered seizure symptoms or whether significant prompting was required in the history-taking process (Reuber et al., 2009). The discrepancies between patients’ and witnesses’ assessments of the relevance of seizure triggers and emotional stress demonstrate that it is also important to consider who has provided clinical information when considering its diagnostic significance.