Onset Latency to a First Psychogenic Nonepileptic Seizure during Inpatient EEG Monitoring: Evidence for Semiological Differences

Abstract

216 consecutive patients diagnosed with psychogenic nonepileptic seizures (PNES) admitted to the epilepsy monitoring unit at our institution over a 4.5 year period were retrospectively identified. PNESs were classified into 4 semiological subcategories: Major motor (n=123), Minor motor (n=38), Akinetic (n=32), and Subjective/experiential (n=23). The median latency to first PNES for the entire population was 7 hours (range <1 to 207 hours) confirming previous observations that the latency to first PNES upon admission is often <24 hours. The novel observation is that latency to first PNES was dependent on type. The median latency to first PNES was significantly prolonged in both the minor motor (median 21 hours) and subjective/experiential (median 22 hours) groups as compared to the major motor (median 5 hours) and akinetic (median 4 hours) groups. Thus, patients with one of these two subtypes may require longer admissions to capture the event of interest and confirm the diagnosis.

Introduction

Psychogenic nonepileptic seizures (PNES) are sudden alterations in behavior that are the result of a psychological process and not an epileptic mechanism (1, 2). The clinical manifestations of these events can range from violent, thrashing behaviors to trembling to staring and immobility to experiential complaints. As these behaviors overlap those associated with epileptic seizures, it is often necessary to admit patients with PNES to an inpatient video-EEG monitoring unit (EMU) for diagnostic evaluation. It is important to differentiate PNES from epilepsy as the treatment varies between these two medical conditions.

It has previously been demonstrated that the latency to the first spontaneous PNES following admission to an epilepsy monitoring unit (EMU) is typically less than 24 hours (3, 4). However, a recent report found that a proportion of first spontaneous PNES occur after 72 hours of monitoring (5). These prior studies grouped all patients with PNES into a single group. The effect of PNES semiology on latency was not investigated. The purpose of this study was to determine if latency to first PNES during an admission to an EMU varies by semiology.

Methods

This retrospective review was approved by the University of Virginia’s Institutional Review Board.

Patient selection

All patients, older than 18 years of age (n=645), admitted to the F.E. Dreifuss Epilepsy Monitoring Unit (EMU) at the University of Virginia over a 4.5-year period were retrospectively placed into the following categories based upon review of the daily EMU reports and discharge summaries: epileptic seizures (n=338), PNES (n=216), physiological non-epileptic events (e.g. myoclonus, tremor, TIA, restless leg syndrome, and migraine headache; n=26), and nondiagnostic admission (n=65). The epileptic seizure group includes 41 patients who had events that could not definitely be differentiated from simple partial seizures during the admission. In all cases of PNES, the EMU report specifically stated that the event occurred in the absence of an ictal or postictal EEG change.

The 216 patients who had at least one unprovoked PNES (PNES patients) and the 65 patients with nondiagnostic admissions form the basis of this report.

PNES Semiological Classification

The daily EMU reports of the 216 PNES patients were reviewed to characterize the semiological presentation. These reports were prepared after complete review of continuous simultaneous video and EEG data obtained digitally (Grass-Telefactor, AstroMed, Inc) by neurologists with additional fellowship training in epilepsy with board certification in clinical neurophysiology (American Board of Psychiatry and Neurology), electroencephalography (American Board of Electroencephalography), or both.

During the monitoring, electrodes were placed according to the 10–20 International Electrode System. Sphenoidal electrodes were not used; however, T1 and T2 electrodes were placed. Patients were observed continuously by a monitor attendant who reported all suspicious clinical and electrographic events. In addition, the patient, patient’s family/friends, and nurses were able to mark events of interest for review.

We classified the PNESs into 4 subcategories: major motor, minor motor, akinetic, and subjective/experiential. Major motor events (Gröppel cluster 1) (6) were those described as consisting of at least one of the following positive motor symptoms in the presence or absence of an altered state of consciousness: thrashing, flailing, pelvic thrusting, nonrhythmic jerking of the extremities, head nodding or head turning side-to-side, or bicycling movements. Minor motor events (Gröppel cluster 2 and 3) (6) were those described as consisting of either a fall, change in posture, trembling, or automatisms (small-amplitude movements of the face or upper extremities) in the setting of altered consciousness. Akinetic events (7) were those described as a behavioral arrest evolving to staring and unresponsiveness. The final category, subjective experiential, were those described as consisting of an alteration of cognition or awareness associated with a patient’s report of a sensory (e.g. dysesthesia, visual hallucination) or emotional experience (e.g. claustrophobia) (8,9).

Latency to the first PNES was calculated as the number of hours from the start time of the recording to the time of PNES onset. Population latencies are presented as Kaplan-Meier survival curves in terms of the probability of not having had a PNES vs. duration of monitoring.

Statistics

The specific statistical tests used are noted in the results section of the manuscript. Statistical significance was accepted at the p<0.05 level.

Results

Patient Demographics

Patient demographics are presented in table 1.

Table 1.

Patient Demographics of PNES Only, PNES/Epilepsy, and Nondiagnostic Admission Groups

		PNES (n=216)	Nondiagnostic (n=65)
Age	Median (range)	37(18–80)	40(18–78)
	Mean ± stdev	37 ± 12.5	42 ± 14
Sex	Male	56 (26%)	22 (35%)
	Female	103 (74%)	43 (65%)
Latency (hours)	Median (range)	7(<1–207)	--
	Mean ± stdev	15.8 ± 23
#Events	Median (range)	3(1–40)	--
Diagnosis of epilepsy at time of admission		23 (11%)	22 (35%)

The PNES patient cohort of 216 patients includes 23 (11%) patients who had a diagnosis of epilepsy at the time of admission. The diagnosis had been previously confirmed in 8 of these patients via video-EEG monitoring. One had a temporal lobectomy for left mesial temporal lobe epilepsy 3 years prior to admission. These patients were admitted for characterization of paroxysmal attacks that were refractory to changes in antiepileptic medications.

There were 15 patients with a clinical diagnosis of epilepsy that had not been confirmed by video-EEG monitoring prior to this admission. Nine of these patients were being treated with antiepileptic medications at the time of the admission. Six were admitted for characterization of paroxysmal attacks that differed in semiology from the habitual event upon which the clinical diagnosis of epilepsy had been made. Three were admitted for characterization of their habitual events which proved to be psychogenic and not epileptic in etiology. The remaining 6 patients had distant diagnoses of epilepsy that resolved during childhood.

The nondiagnostic admission group was included in our calculation of PNES latency as the eventual diagnosis of PNES cannot be excluded in this patient population. Compared to those who had a PNES, the nondiagnostic admission group was more likely to have a diagnosis of epilepsy at the time of admission (p<0.05, χ-square). Although there were trends to more males (p>0.05, χ-square) and older age (p>0.05, Kolmogorov-Smirnov (KS) test) within the nondiagnostic admission group, neither of these variables reached statistical significance.

Semiological Classification

For these 216 patients, the captured events were categorized as either major motor (n=123, 57%), minor motor (n=38, 17%), akinetic (n=32, 15%), and subjective/experiential (n=23, 11%). The age and gender distribution were similar across groups (p>0.05, χ-square, table 2).

Table 2.

Patient Demographics of the PNES Semiology Subcategories

		Major Motor (n=123)	Minor Motor (n=38)	Akinetic (n=32)	Subjective/Experiential (n=23)
Age	Median (range)	37(18–80)	31 (19–67)	34(18–67)	37(18–60)
	Mean ± stdev	38 ± 13	35 ± 13	36 ± 10	38 ± 12
Sex	Male	31 (25%)	12 (29%)	4 (12%)	7 (30%)
	Female	92 (75%)	27 (71%)	28 (88%)	16 (70%)
Latency (hours)	Median (range)	5(<1–65)	21(<1 – 105)*	4 (<1–67)	22 (<1–207)*
	Mean ± stdev	10 ± 13	27 ± 29	13 ± 19	28 ± 43
#Events	Median (range)	3(1–40)	3 (1–29)	3 (1–12)	2 (1–17)

Latency to first event

The median latency for the entire patient population was 7 hours (figure 1). In total, 16% (46/281) had their first event within the first hour of admission, 61% (171/281) had their first PNES within 24 hours, and 71% (199/281) had their first PNES within 48 hours. The longest latency to an unprovoked PNES was 207 hours.

Figure 1.

Kaplan-Meier survival curve of latency to first psychogenic nonepileptic seizures following start of video-EEG recording for all patients independent of semiological type.

Linear regression analysis was performed using age, gender, and PNES semiology as independent variables in a model with PNES latency as the dependent variable. Only PNES semiology (p<0.001) had a significant effect on latency.

Kaplan-Meier survival curves for PNES latency grouped by PNES semiological classification are displayed in figure 2. The survival curve for the major motor group is repeated in each graph for comparison. The median and mean latencies to first event for each group are presented in table 2. The median latency to PNES was prolonged in minor motor (21 hours) group and subjective/experiential (22 hours) group as compared to the major motor (5 hours) and akinetic (4 hours) groups (Newman-Keuls, p<0.05).

Figure 2.

Kaplan-Meier survival curves of latency to first psychogenic nonepiletpic seizure following start of video-EEG recording grouped by semiological classification. The survival curve for the major motor group is repeated in each graph for comparison. *p<0.05

Between 48 and 72 hours, there were 12 initial events captured (5% of the PNES total): 4 were major motor (3% of total major motor), 4 were minor motor (10% of total minor motor), 4 were akinetic (%12.5 of akinetic), and 0 were sensory/experiential. After 72 hours, 6 additional initial events were captured (2.5% of the PNES total): 4 were in the minor motor group (10% of the total minor motor) and 2 in the subjective/experiential group (9% of the subjective/experiential). None of the events after 72 hours were major motor or akinetic.

DISCUSSION

In this study, we have confirmed the previous finding (3,4,5) that the latency to first unprovoked PNES is relatively short with the majority of events occurring within 48 hours. The novel finding was that the latency to first event varied with PNES semiology. In this study, the latency to events classified as either minor motor or subjective/experiential was significantly longer than either major motor or akinetic events.

As Parra and colleagues (3) observed 96.2% of PNESs by 48 hours of monitoring and no events after 58 hours, they recommended that induction protocols be withheld until after 48 hours of recording. In our cohort, 5% of the total the initial PNESs occurred after the 48 hour time point and 2.5% occurred after 72 hours. In a recent study (5), 25% of PNESs occurred on the third day or after and 8% of patients did not have their first event until day 7 or after.

Semiological classification was not provided in either of these prior studies. Although our findings support the recommendation for induction protocols at 48 hours in the case of major motor and akinetic PNESs at those centers that choose to use these procedures (10,11), our findings suggest that an additional day or more of monitoring should be considered if the event of interest is consistent with a minor motor or subjective/experiential based on the operational definitions used in this report.

Recent fMRI studies have begun to elucidate those brain regions that are activated during conversion disorders (12,13). However, the psychopathogenesis of PNES has not yet been elucidated and the biological significance of our findings of semiological-dependent latency to first PNES is uncertain at this time. Prior studies investigating semiological variation have demonstrated semiological differences in psychiatric predisposition (14–16) and outcome (17,18). The presence of semiological-dependence in these variables suggests the possibility that different mechanisms underlie the semiological differences in PNES.

Future studies investigating the mechanisms of PNES would benefit from an accurate classification scheme. However, there is, of yet, no official classification schema for PNES and the classification of PNES has varied by report (6, 7, 16, 17, 19–21). At a minimum, these schema have differentiated PNES dominated by major motor activities such a thrashing, wild head movements, and pelvic thrusting from other PNESs. Gröppel et al. (6) used a cluster analysis technique based on only 28 patients to devise a classification scheme that divided PNES into 3 groups. However, their system is limited by the absence of akinetic events and subjective/experiential events, which represent approximately 25% of the events observed in this study and have been frequently observed in other studies of PNES (e.g. (7, 21)).

There were several limitations of this study. The format was a retrospective chart review and semiological classification was dependent on the description available in the EMU reports and discharge summaries. Additional analysis to determine if there were differences in psychiatric comorbidities or previous trauma could not be pursued based on the limited data available in the medical record. Furthermore, there was limited follow-up data available for either the PNES group or the nondiagnostic group. We suspect that among the nondiagnostic group are a large number of patients with epilepsy who did not happen to have an epileptic seizure during their admission as well as people who may be potentially malingering. Additional prospective studies are warranted to better characterize the group of patients who have nondiagnostic admissions to EMUs.

In summary, this study demonstrated that a large proportion of patients with PNES will have their first event within 48 hours of inpatient video EEG monitoring with many occurring within the first 24 hours. In general, although two days of monitoring are required for most patients, those patients with minor motor or subjective/experiential PNES may require longer admissions to capture the event of interest and confirm the diagnosis. What is not known is whether these variable clinical presentations are related to differences in the underlying mechanism. Further studies are warranted to determine not only the correlation of neurologic and psychiatric comorbidites and outcome but also to begin to elucidate the neuropathogenesis of PNES.