SEARCH

SEARCH BY CITATION

Keywords:

  • Nonepileptic seizures;
  • Pseudoseizures;
  • Semiology;
  • Video-EEG use in epilepsy

Summary

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

Purpose:  To systematically study the semiology of psychogenic nonepileptic seizures (PNES) captured by video–electroencephalography (EEG) monitoring (VEM) and categorize the typical patterns observed.

Methods:  VEM records of patients who underwent evaluation from January 2002 to June 2007 were reviewed to identify those who had PNES with or without a background of epilepsy. The semiology of each event was visually analyzed and entered into a statistical database. Type of movement, anatomic distribution, synchrony, symmetry, onset, offset, course, duration, vocalization, hyperventilation, eye movements, and responsiveness were evaluated. PNES were classified into distinct groups according to the predominant motor manifestation.

Results:  A total of 330 PNES from 61 patients were studied. Based on semiology, six different types of PNES were observed as follows: (1) rhythmic motor PNES characterized by rhythmic tremor or rigor-like movements (46.7%); (2) hypermotor PNES characterized by violent movements (3.3%); (3) complex motor PNES characterized by complex movements such as flexion, extension, abduction, adduction, rotation, with or without clonic-like and myoclonic-like components of varying combinations and anatomic distribution (10%); (4) dialeptic PNES characterized by unresponsiveness without motor manifestations (11.2%); (5) nonepileptic auras characterized by subjective sensations without any external manifestations, marked in the VEM records as “seizure button presses” (23.6%); and (6) mixed PNES where combinations of above seizure types were seen (5.2%). In a given patient, all the seizures belonged to a single type of PNES in 82% of cases.

Discussion:  PNES can be classified into six stereotypic categories. Contrary to common belief, PNES demonstrates stereotypy both within and across patients.

Psychogenic nonepileptic seizures (PNES) are a common problem encountered by epileptologists (Ghougassian et al., 2004). The prevalence of PNES is estimated to be 2–33 per 100,000 (Benbadis & Hauser, 2000) compared to the prevalence of 4–6 per 1,000 in epilepsy (Hauser & Kurland, 1975). The average delay between the onset of seizures and the diagnosis is 7 years (Reuber et al., 2002). The average medical cost per patient with PNES for a period of 6 months is estimated to be around 8,000 USD (Martin et al., 1998), which means that every PNES patient will utilize healthcare resources worth 112,000 USD before the diagnosis of PNES is established. Therefore, this condition is a significant neurologic disorder with considerable social and economic implications. Making a firm diagnosis is vital, and is shown to reduce the subsequent healthcare utilization costs significantly (Martin et al., 1998).

The diagnosis of PNES is challenging. Seizure semiology, psychiatric history, seizure provocation techniques, postictal prolactin assay, and psychological testing are various diagnostic methods with varying strengths and drawbacks (Kuyk et al., 1997; Crager et al., 2002). However, there is general agreement that video–electroencephalography (EEG) monitoring (VEM) is the gold standard test (Reuber & Elger, 2003).

Pattern recognition of events forms the cornerstone of interpreting video-EEG findings. Several studies have described various semiologic features of PNES (Gulick et al., 1982; Luther et al., 1982; Gates et al., 1985; Meierkord et al., 1991; Leis et al., 1992; Saygi et al., 1992; Lancman et al., 1993; Scheepers et al., 1994; DeToledo & Ramsay, 1996; Geyer et al., 2000; Vossler et al., 2004; Chung et al., 2006). However, the lack of a classification system of PNES is a major drawback in its recognition and management. This affects homogeneity of data across different studies as well as understanding of the long-term prognosis of different subgroups.

The current study was planned against this backdrop. We sought to systematically study the semiology of PNES based on inpatient VEM. By performing detailed semiologic analysis of all witnessed PNES, we sought to categorize the typical patterns observed.

Methods

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

We retrospectively reviewed medical records and video-EEG records of all adult patients who underwent monitoring at two tertiary care epilepsy centers from January 2002 to June 2007. Those who were diagnosed with PNES with or without a background of epilepsy were selected for this study. The diagnosis of PNES was made on the basis of the consensus of at least two epilepsy specialists who based their opinion on the clinical details and video-EEG monitoring, independent of the current study. The following criteria were used to diagnose PNES: at least a single typical clinical event was captured on EEG, the EEG did not show any electrographic ictal rhythm during the event, no postictal slowing was seen on the EEG immediately after the event, and there was no evidence of an alternative neurologic diagnosis such as a movement disorder for this event. When epilepsy was coexistent, it was diagnosed on the basis of supportive history with neuroimaging, interictal epileptiform discharges on video-EEG or previous EEG, and captured epileptic seizures on video-EEG.

All patients had continuous in-patient VEM for a period typically ranging from 1–8 days. Scalp electrodes were placed in accordance with the 10–20 international electrode system. Antiepileptic drugs (AEDs) were usually discontinued or reduced during the recording period. EEG and audiovisual signals were acquired and analyzed using Compumedics (Compumedics Ltd, Melbourne, Australia) and Vanguard (Lamont Systems, Cleveland, OH, U.S.A.) digital video-EEG systems.

Video-EEG studies of all selected patients were reviewed by an epilepsy specialist (U.S.). Video segments of all clinical events and the corresponding EEG studies were reviewed several times separately as well as in synchrony, until the investigator was satisfied that all the details were visualized and tabulated.

The semiology of each clinical event was visually analyzed in detail and entered into an SPSS statistical database (SPSS Inc, Chicago, IL, U.S.A.). Type of movement, anatomic distribution of the movement (extremities, head, trunk, pelvis), synchrony, symmetry, eye movement, responsiveness, vocalization, hyperventilation, onset (abrupt or gradual), offset (abrupt or gradual), course, and duration of the event were tabulated.

We used the predominant motor movement to classify the PNES into distinct groups. Other characteristics were subsequently analyzed in each group. The pattern of stereotypy of the PNES type within and across patients was studied.

Because this was a descriptive study, no statistical inferences could be drawn. Summary statistics included mean, standard deviation, minimum, maximum, and median for continuous variables and numbers and percentages for categorical variables. The statistical analysis was carried out using the SPSS software program.

The study was approved by the human research ethics committees of Alfred Hospital and Monash Medical Centre, Melbourne, Australia, where it was conducted.

Results

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

A total of 330 PNES from 61 patients were studied. The mean number of PNES recorded per patient was five. There were 45 female and 16 male patients with ages ranging from 16–83 years (mean 38 years). The mean age of onset of PNES was 29 years. The age range of onset was 9–80 years, and late onset of the condition was seen in a minority (four patients at or above the age 55). The mean duration of VEM was 3 days (range 1–8 days).

Detailed visual analysis of all PNES revealed three types of motor manifestations, based on which six different PNES seizure types were delineated as follows (Table 1):

Table 1.   Comparison of different semiologic characteristics between PNES groups
 Rhythmic Motor PNESHypermotor PNESComplex Motor PNESDialeptic PNESNonepileptic AurasMixed PNES
  1. PNES, psychogenic nonepileptic seizures; UPL, upper limb; LL, lower limb; UL, unilateral; BL, bilateral; HV, hyperventilation; head shaking, side to side movement “no-no” gesture; head nodding, anterior posterior movement “yes-yes” gesture.

Type of movementRhythmic tremor or rigor likeViolent, thrashing, kickingComplexNilNilAny of the above
UPL involvement62.3% (BL 54.5%; UL 7.8%)72.7% (all are BL)78.8% (all are BL)NilNil88.3% (BL 82.4%; UL 5.9%)
UPL symmetry and synchrony91.7% of BL are symmetric and synchronousAll asymmetric and asynchronousAll asymmetric and asynchronousNilNilAll BL are asymmetric and asynchronous
LL involvement55.8% (BL 52.6%; UL 3.2%)100% (all are BL)100% (all are BL)NilNil88.3% (all are BL)
LL symmetry and synchrony95.1% of BL are symmetric and synchronousAll asymmetric and asynchronousAll asymmetric and asynchronousNilNilAll BL are asymmetric and asynchronous
Eyes (at the event onset)Closed 94.2%, open 5.2%, not visible 0.6%Closed 54.5%, not visible 45.5%Closed 63.6%, open 6.1%, not visible 30.3%Closed 73%, rapid blinking 16.2%, not seen 10.8%Open 100%Closed 82.4%, rapid blinking 17.6%, open 0%
HeadNodding 5.8%, shaking 16.2%Shaking 18.2%, no movement 81.8%Shaking 3%, nodding 3%No movementNo movementNodding 11.8%, no movement 88.2%
TrunkRhythmic (rigor-like) 65.6%Back arching 27.3%, no movement 72.7%Back arching 51.5%, no movement 48.5%No movementNo movementBack arching 64.7%, rhythmic 17.6%
PelvisSwinging 65.6%, thrusting 0.6%Thrusting 45.5%, no movement 54.5%Thrusting 69.7%, swinging 12.1%No movementNo movementThrusting 17.6%, swinging 17.6%
ResponsivenessUnresponsive 83.8%Unresponsive 100%Unresponsive 100%Unresponsive 100%Responsive 100%Unresponsive 100%
Vocalization9.7%45.5%57.6%5.4%Nil64.7%
HV13.6%36.4%15.2%16.2%Nil88.2%
OnsetAbrupt 68.8%, gradual 31.2%Abrupt 81.8%, gradual 18.2%Abrupt 30.3%, gradual 69.7%Abrupt 100%Abrupt 100%Abrupt 94.1%, gradual 5.9%
OffsetAbrupt 66.9%, gradual 33.1%Abrupt 81.8%, gradual 18.2%Abrupt 15.2%, gradual 84.8Abrupt 100%Abrupt 100%Abrupt 29.4%, gradual 70.6%
CourseSteady 65.6%, waxing/waning 34.4%Steady 45.5%, waxing/waning 54.5%Steady 72.7%, waxing/waning 27.3%Steady 97.3%, waxing/waning 2.7%Steady 100%Steady 58.8%, waxing/waning 41.2%
Mean duration (min )2.60.939.50.194.1
  • 1
     Rhythmic tremor, trembling or rigor like movements constituted 46.7% of all PNES. We labeled this group as Rhythmic Motor PNES. These seizures were characterized by rhythmic movements involving the extremities and trunk typically in a symmetric and synchronized fashion. Upper limb involvement was more common than the lower limb involvement. Patients with rhythmic motor PNES were unresponsive during the seizures, which means they did not respond to external stimuli (i.e., they appeared unconscious), and approximately 10% demonstrated vocalizations and hyperventilation. About two-thirds had abrupt onset and offset.
  • 2
     Hyperkinetic or hypermotor movements. This group was labeled as Hypermotor PNES and constituted 3.3% of all nonepileptic seizures. These patients typically demonstrated violent thrashing, punching, or kicking type of movements involving the extremities and trunk. In the extremities, the movements were always bilateral, asymmetric, and asynchronous.
  • 3
     Complex motor movements. This group was labeled as Complex Motor PNES and represented 10% of all PNES in the cohort. These seizures were characterized by complex and multifocal movements of both proximal and distal extremities consisting of flexion, extension, abduction, and adduction with or without clonic-like and myoclonic-like components of varying combinations. The movements were always asymmetric, asynchronous, and migratory in anatomic distribution. Back-arching and pelvic thrusting was commonly seen in this group. Compared to hypermotor PNES (type 2), these movements were more subtle and less aggressive, sometimes with an apparent nonphysiologic migratory quality.
  • 4
     Prolonged, motionless, unresponsive patients with no motor manifestations. We classified these patients as Dialeptic PNES, which constituted 11.2% of all PNES. These patients presented in a coma-like state. There were no movements, and they remained unresponsive to all external stimuli. A minority demonstrated hyperventilation. These seizures tend to last longer than other PNES (ranging from 2–57 min, mean 9.5 min) and the vital signs and EEG were unchanged during the seizure.
  • 5
     Nonepileptic auras (23.6%). These events were characterized by various subjective sensations without any external manifestations, marked in the VEM records as “seizure button presses.” The EEG and electrocardiography (ECG) remained normal during these episodes, with no alternative systemic or biochemical explanation. The usual terms used by the patients to describe these sensations were, “I am going through a trance,”“I feel weird,” and “zoning out.”
  • 6
     Mixed PNES (5.2%). These PNES would present as a combination of any of the preceding types.

Ictal eye closure at the beginning of the seizure was witnessed in most patients across all PNES types. Rapid eye blinking was also seen in some patients with dialeptic and mixed PNES.

EEG demonstrated characteristic patterns in different PNES types due to movement and muscle artifacts. In dialeptic PNES and nonepileptic auras, EEG remained unchanged (Fig. 1). Rhythmic motor PNES produced a typical EEG pattern consisting of 6–9 Hz rhythmic “slow waves” due to movement artifact admixed with normal background EEG (Fig. 2). This was a very regular rhythm with no evolution in amplitude or distribution. In hypermotor PNES the entire EEG was obscured by muscle artifacts. At the offset the EEG returned to normal background rhythm immediately (Fig. 3). The ictal EEG in complex motor NES was somewhat similar to that of hypermotor PNES, but the muscle artifact was less dense, and some “slow waves” were visible underneath due to movement and electrode artifacts (Fig. 4).

image

Figure 1.   Electroencephalography (EEG) study of dialeptic psychogenic nonepileptic seizures (PNES). This patient remained unresponsive during the episode. Note rapid eye blinking and preserved background alpha rhythm.

Download figure to PowerPoint

image

Figure 2.   Electroencephalography (EEG) of rhythmic motor psychogenic nonepileptic seizures (PNES). (A) This patient had rhythmic tremor–like movements of the trunk and extremities for 4 s. Note rhythmic “slow waves” due to movement artifact with an abrupt onset and offset. The normal background rhythm resumes instantly with the offset of the event. (B) Another patient who had rhythmic rigor-like movements of the trunk with side-to-side head shaking. Rhythmic “slow waves” have slower frequency compared to A. Neither EEG studies demonstrate any evolution of the ictal rhythm.

Download figure to PowerPoint

image

Figure 3.   Electroencephalography (EEG) of hypermotor psychogenic nonepileptic seizures (PNES). Abrupt onset of violent thrashing-type movements of the extremities was seen in this patient. Note the dense muscle and movement artifact.

Download figure to PowerPoint

imageimage

Figure 4.   Electroencephalography (EEG) of complex motor psychogenic nonepileptic seizures (PNES). (A) This patient demonstrated complex, irregular, asynchronous, flexion, extension, abduction, and adduction movements of extremities. Note the abrupt onset of artifact. (B) EEG showing the ongoing seizure of the same patient. Note the artifact is less “dense” compared to Figure 3. Some “slow waves” are visible underneath. (C) The offset is gradual. The normal background rhythm returns immediately with no postictal slowing.

Eight patients (13.1%) had coexistent epilepsy consisting of temporal lobe epilepsy in four, primary generalized epilepsy in three, and frontal lobe epilepsy in one. Their PNES were semiologically different from the epileptic seizures.

We also studied the degree of stereotypy within patients. In a given patient, all PNES belonged to the same semiologic type in 82% of cases.

Discussion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

We provide a detailed analysis of semiologic features of PNES in a systematic manner. Based on that, we have delineated six stereotypic PNES types. This has potential implications for diagnosis and prognosis, with a number of studies indicating that outcome of PNES may vary among different types (Selwa et al., 2000; Reuber et al., 2003).

A number of studies have described different semiologic features of PNES (Gulick et al., 1982; Luther et al., 1982; Gates et al., 1985; Meierkord et al., 1991; Leis et al., 1992; Saygi et al., 1992; Lancman et al., 1993; Scheepers et al., 1994; DeToledo & Ramsay, 1996; Geyer et al., 2000; Vossler et al., 2004; Chung et al., 2006). However, patients with PNES usually present as a combination of multiple features. Therefore, it would seem logical to group these patients depending on such combinations. A few authors have made attempts to identify distinct semiologic groups of PNES. One such study based on video-EEG of 110 patients classified nonepileptic seizures into two groups: attacks with collapse and limpness and attacks with prominent motor activity (Meierkord et al., 1991). In a literature survey of 62 patients, four groups were identified as tonic–clonic, complex partial type attacks, combination of attack types, and special types (swoons, tantrums, arc-de-cercle or opisthotonus, and conscious simulation) (Van Merode et al., 1997). Another study on a cohort of patients in a neuropsychiatry ward classified “pseudo-seizures” into swoons, tantrums, and “abreactive attacks” (Betts & Boden, 1992). Cluster analysis was used by another group to classify PNES. They studied 27 patients and identified three groups labeled as psychogenic motor seizures, psychogenic minor motor or trembling seizures, and psychogenic atonic seizures (Groppel et al., 2000).

Consequently, it appears that there is no strict uniformity in nomenclature and classification between the different PNES types described in different studies. However, the semiologic PNES categories delineated in our study can be identified among those. “Swoons” described by Van Merode et al. (1997) and Betts and Boden (1992) are similar to dialeptic PNES. These patients have also been identified as “psychogenic pseudosyncope” (Benbadis & Chichkova, 2006), and “psychogenic unresponsiveness” (Abubakr et al., 2003). Although detailed clinical features of “psychogenic atonic seizures” in the reference Groppel et al. (2000) are not available, we suspect this group also represents dialeptic PNES. Typical “swoon” attacks are characterized by sudden collapses followed by apparent coma-like state with no response to external stimuli and without any convulsive movements (Betts & Boden, 1992). In our series the collapse component was not witnessed in most patients, as they were confined to the bed for monitoring. We believe the key semiologic feature in this group is prolonged unresponsive state.

Rhythmic motor PNES are identified as “trembling seizures” in the study conducted by Groppel et al. (2000), whereas the group described as “psychogenic motor seizures” most likely represents hypermotor PNES and complex motor PNES. The phenomenon of nonepileptic auras has been recognized by some other authors as well (Lempert & Schmidt, 1990; Lancman et al., 1993).

Type 3 (complex motor PNES) represents a group of subjects with highly heterogeneous movements including clonic-like and myoclonic-like movements. In this group, psychogenic myoclonic-like body jerks in particular pose a diagnostic challenge of differentiating from epileptic generalized myoclonus and nonepileptic nonpsychogenic myoclonus. EEG interpretation is made difficult by the predominant muscle artifact, which tends to mask the discharges in cortical myoclonus.

All VEM patients were reviewed by a multidisciplinary team including epilepsy specialists and movement disorder neurologists. Challenging cases where cortical, subcortical, or spinal myoclonus (Shibasaki, 2000) were considered after consensus opinion were subsequently evaluated with more detailed imaging of the neuroaxis and electrophysiologic testing before being excluded.

Electrophysiologic studies can be very useful in differentiating true myoclonus from psychogenic myoclonic-like jerks as seen in PNES (Shibasaki, 2000; Brown & Thompson, 2001). Cortical myoclonus is often stimulus-sensitive, irregular, and usually accompanied by EEG correlates of polyspikes and spike-and-wave discharges (Shibasaki, 2000).

When surface electromyographic (EMG) recordings are used, psychogenic jerks are characterized by long EMG bursts (>70 ms) with a triphasic pattern of activation of agonist and antagonist muscles (Brown & Thompson, 2001). Simultaneous recording of EEG and EMG to demonstrate their temporal relationship (EEG–EMG polygraphy) should be considered in patients with myoclonus. Jerk-locked back averaging (JLBA) is a technique employing this principle. In cortical myoclonus JLBA elicits a positive–negative biphasic spike over the corresponding cortical region 20–30 ms preceding the EMG activity of myoclonus (Shibasaki, 2000). In psychogenic myoclonus the same technique can be used to demonstrate movement-related cortical potentials (MRCPs), which are not elicited in true myoclonus. MRCPs are characterized by slow-rising negativity (Bereitschaftspotential) over the central cortical region starting 0.7 to 2.1 s before the onset of jerks (Terada et al., 1995).

Characteristic EEG patterns in PNES are an interesting finding. Movement, muscle, and rhythmic artifacts as “ictal” electrographic features of PNES have been described (Vinton et al., 2004; Benbadis, 2006). We have demonstrated that different PNES types are associated with distinctive EEG patterns due to characteristic muscle, movement, and electrode artifacts.

There are some limitations of this study, however. All patients were recruited from two tertiary centers, thereby introducing a potential selection bias. It is possible that more severe forms of PNES constituted the cohort and milder forms with different semiology were not represented. Although a community-based sample more representative of PNES outside a tertiary setting would rectify this potential bias, case ascertainment is likely to be challenging.

Type 5 (nonepileptic auras) without outward visual manifestations poses perhaps the greatest diagnostic challenge to our proposed observational classification system. Although we acknowledge that ultimately we are unable to definitively exclude their epileptic nature, we still feel this group is a distinct category of PNES for several reasons. First, this category is well recognized by previous authors, suggesting that our findings are consistent with the literature (Lempert & Schmidt, 1990; Lancman et al., 1993). Secondly, none of the patients with nonepileptic auras had coexistent epilepsy, making these phenomena less likely to represent merely a less severe form of partial seizure. Thirdly, patients with nonepileptic auras in our study cohort had no other corroborative clinical, EEG, or radiologic features to suggest an underlying consistent structural or functional abnormality. Finally, the verbatim language descriptors used by the patients were qualitatively and quantitatively (i.e., multiple symptoms) different from epileptic auras described previously. Patients used terms such as “I feel weird” and “I am going through a trance” often with extended duration.

Placebo induction has been proposed as a reliable tool for identifying PNES (Bazil et al., 1994), and could have been useful in this challenging group. However, there are some reservations with use of this technique. First, this method has been shown to induce atypical nonepileptic events and true epileptic seizures in some patients, thereby leading to incorrect diagnosis (Walczak et al., 1994). Second, the ethical validity of the test has been questioned by some authors given the “deceptive” nature of the technique, as the doctor is unable to give a complete explanation to the patient (Kuyk et al.,1997).

All video-EEG studies were visually analyzed and classified by a single observer. Therefore, our findings need to be validated for interobserver reproducibility.

Despite these limitations, our study highlights the importance of semiology in the diagnosis and classification of PNES. Recognition of semiology is of prime importance in the diagnosis of epileptic seizures. Proposal for a semiologic classification of epileptic seizures underscores this fact (Luders et al., 1998). This is of even greater relevance in PNES, as there are no defining electrophysiologic characteristics for this condition.

A more semiologically focused classification system of this commonly encountered disorder may aid in improved recognition and diagnosis. It is hoped that this may lead to improved standardization across different studies and ultimately better etiologic understanding and management of this important public health issue. The current study demonstrates that PNES are semiologically stereotypic. Therefore, a classification based on semiology seems logical and practical.

PNES are generally considered to present with multiple seizure types within and between patients. Contrary to popular belief, we have demonstrated from a large series of patients from two tertiary care centers that PNES are highly stereotypic both within and across individual patients.

Acknowledgments

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

We would like to thank Belinda Briggs, EEG Scientist, Department of Neuroscience, Alfred Hospital, Melbourne, for providing technical support and helping with data collection.

Disclosure

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

Dr. Udaya Seneviratne reports no disclosures. Professor David Reutens reports no disclosures. Dr. Wendyl D’Souza is funded by a National Health Medical Research Council of Australia Post-Doctoral Health Professional Fellowship. He is on the Zonergan Australian Scientific Advisory Board and has received an investigator-initiated study grant from UCB-Pharma Australia. We confirm that we have read the journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

References

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References