JLE

Epileptic Disorders

MENU

The aetiology of psychogenic non-epileptic seizures: risk factors and comorbidities Volume 21, issue 6, December 2019

Figures


  • Figure 1

  • Figure 2

  • Figure 3

  • Figure 4

Psychogenic non-epileptic seizures (PNES), also known as dissociative seizures, are episodes of altered subjective experience, involuntary movements and reduced self-control that can resemble epilepsy, syncope, or other paroxysmal disorders. However, rather than being caused by neuronal hypersynchronisation (as in epilepsy) or cerebral hypoperfusion (as in convulsive syncope), PNES are the result of complex neuropsychiatric dysfunction. The diagnosis is primarily based on the analysis of seizure experiences and visible manifestations (the patient's history and observable seizure semiology), and is supplemented by the careful exclusion of alternative explanations such as epilepsy (Avbersek and Sisodiya, 2010; LaFrance et al., 2013). This narrative review addresses key learning objectives of the ILAE curriculum by describing the demographic profile, common risk factors and comorbid disorders of PNES (Blümcke et al., 2019).

While different psychophysiological mechanisms may make variable contributions to the aetiology of PNES in individual patients, a range of predisposing, precipitating and perpetuating factors have been identified across patient populations and integrated into theoretical models (Brown and Reuber, 2016a, 2016b; LaFrance and Bjønæs, 2019). Notably, many possible aetiological contributors defy a simple, dualistic categorization into “psychological” or “physical” factors (e.g. childhood abuse affects both brain maturation and social functioning). Within a biopsychosocial framework, it is somewhat arbitrary to discuss disorders which have been defined as being nosologically distinct (e.g. panic disorder or post-traumatic stress disorder [PTSD]) as “comorbidities” and not as primary disorders, which include PNES as one manifestation. With this caveat in mind, we discuss risk factors and comorbidities along the lines of the current nosologies in which most PNES are classified as a functional neurological symptom (conversion) disorder according to Diagnostic and Statistical Manual of Mental Disorders (DSM-5), or dissociative disorder in the International Classification of Diseases 11th edition (Perez et al., 2015; Erro et al., 2016).

Epidemiology

PNES is a common disorder which is recognised around the world (Kanemoto et al., 2017). Major differences in access to diagnostic facilities and expertise mean that there is no reliable epidemiological evidence allowing prevalence comparisons between different countries (Hingray et al., 2018). PNES have an annual incidence of at least 1.5-6.17/100,000 (Sigurdardottir and Olafsson, 1998; Szaflarski et al., 2000; Duncan et al., 2011). This range is likely to be an underestimate of the general population incidence because it only accounts for video-EEG documented diagnoses. In general neurology outpatient clinics, PNES account for 2% of new referrals (Stone et al., 2010). In specialised “first seizure” clinics, they comprise 8-12% of presentations (Angus-Leppan, 2008; Duncan et al., 2011). Among patients presenting with convulsive seizures to emergency services, PNES are recognised in 11% of cases (Dickson et al., 2017). The proportion of those with PNES among patients with apparently refractory epilepsy referred to tertiary epilepsy units is around 30% (Asadi-Pooya and Sperling, 2015). The prevalence of PNES in the general population has been estimated at 2-50/100,000 (Benbadis and Allen Hauser, 2000; Kanemoto et al., 2017). Patients with the same disorder also account for about 5% of referrals to specialist syncope services (although their condition is most commonly labelled as “psychogenic syncope” or “pseudosyncope”) (Tannemaat et al., 2013; Popkirov et al., 2014).

PNES cohort studies have identified a number of socioeconomic and demographic risk factors for PNES (McKenzie et al., 2010; Duncan et al., 2011; Goldstein et al., 2019). Women (and girls) make up 60-80% of all patients, although the gender disparity is smaller in older adults and those with intellectual disability (Duncan et al., 2006; Duncan and Oto, 2008; Goldstein et al., 2019). The mean and median age at onset is around 28 years (although the modal age is 19 years). Young women are at particularly high risk (Asadi-Pooya and Sperling, 2015; Goldstein et al., 2019). Like epilepsy, PNES is more commonly diagnosed in individuals from lower socioeconomic groups (Duncan et al., 2012), with one large study from the United Kingdom reporting that over 50% of all patients live in areas categorised within the highest deprivation quintile and over two-thirds of patients being unemployed at the point of seeking treatment for PNES (Goldstein et al., 2019).

Explanatory frameworks

A number of theoretical models have been proposed and summarized for PNES (LaFrance and Bjønæs, 2019). A recently proposed model, the “integrative cognitive model” (ICM), accommodates current research on experiential, psychological and biological risk factors for the development of PNES, but in view of the considerable heterogeneity of presentations and medical context, it is not certain that a universal model can capture the full range of PNES manifestations (figure 1) (Brown and Reuber, 2016b; Reuber and Brown, 2017). In the ICM, PNES are conceptualised as experiential and behavioural paroxysms that result from the activation of a learnt mental representation (akin to a computer programme, i.e. the idea of a seizure, or “seizure scaffold”), often combined with concurrent physiological arousal. The seizure scaffold can contain elements of instinctive automatisms (e.g. freezing or thrashing movements), personal illness experiences (e.g. syncope or epilepsy) or illness beliefs (e.g. derived from witnessed seizures). The activation of the seizure scaffold is triggered by perceived threat and/or conditioned cues, and facilitated by a failure of inhibitory systems. This disinhibition, along with reduced behavioural awareness (or sense of agency), can occur in dissociative states, but also in the context of chronic stress or rumination. The activation of the seizure scaffold (associated with a change in awareness of the self and the environment) ultimately leads to the resolution of a state of distress and parasympathetic activation. This has a physiological component (a calming effect of dissociation through detachment and emotional numbing) and a psychosocial component (an escape from the trigger, e.g. through the interruption of interpersonal confrontation, catastrophising thoughts or intrusive memories). The “effective” resolution of the distressed state means that each PNES reinforces the system and makes the next PNES more likely to occur.

The ICM can accommodate much of the aetiological and clinical heterogeneity which characterises PNES. In the following sections, we discuss common aetiological factors and how they could contribute to the predisposition, precipitation and perpetuation of PNES according to the ICM.

Experiential risk factors

Traumatic abuse and stressful life events

The epidemiological association of PNES with previous life adversity and psychological trauma, which has been recognised since the nineteenth century (Breuer and Freud, 1895), continues to be a central element in the current understanding of the condition (Bowman, 2018). Studies have demonstrated increased rates of both childhood maltreatment (including sexual, physical and psychological abuse and neglect) and stressful life events, such as bereavement or illness (Bowman, 2018). A recent systematic review calculated that among 903 patients with PNES across 24 highly heterogeneous studies (compared to 1,023 controls), the odds ratio of retrospective reports of stressors in childhood and adulthood was 3.1 (confidence interval 1.7-5.6) (Ludwig et al., 2018). Of note, different studies identified some patients (14-70%) who reported no severely stressful life events or childhood maltreatment (Ludwig et al., 2018). This means that, although trauma and neglect may occasionally occur too early in life to be explicitly remembered, and the findings may have been affected by poor sensitivity of some of the methods used to capture trauma (Baldwin et al., 2019), an attitude of “there must have been something traumatic” is not justified. Among patients with PNES, trauma is more prevalent in those with more severe psychiatric comorbidity and stronger dissociative tendencies (Hingray et al., 2011). Patients with PNES reporting sexual trauma also have more severe, and more commonly convulsive seizures, often characterised by emotional triggers and prodromal symptoms (Selkirk et al., 2008) although this was not replicated in one recent study (Asadi-Pooya and Bahrami, 2019). While the identification of these features justifies a more extensive psychiatric exploration, the absence of reported trauma or stress does not contradict a diagnosis of PNES.

Trauma can contribute to the occurrence of PNES in various ways. Firstly, childhood maltreatment and traumatization have a clearly observable impact on brain development and maturation (Perry and Pollard, 1998). Studies have shown reproducible structural and functional changes in the brains of adults who have suffered childhood trauma (Herringa, 2017). Reduced grey matter in limbic and prefrontal areas is consistently found and related to abnormal stress and emotion regulation (Paquola et al., 2016). These effects on frontolimbic structures may be aggravated by reduced functional connectivity between limbic and prefrontal areas as well as increased amygdala reactivity to negative stimuli (Herringa, 2017). Although imaging studies in patients with PNES have so far failed to converge on pathognomonic structural abnormalities across unselected PNES patient populations (Asadi-Pooya, 2015; McSweeney et al., 2017), altered structural frontolimbic connectivity has been demonstrated in some cases (Hernando et al., 2015; Lee et al., 2015; Perez et al., 2017), possibly related to disruption of age-dependent maturation processes (Popkirov et al., 2018a).

The long-term developmental effects of childhood trauma also include changes in various developmentally sensitive allostatic systems relevant to PNES and related disorders (Keynejad et al., 2018). Childhood maltreatment affects inflammatory states in adulthood (Danese et al., 2007) with studies consistently showing chronic alterations independent of psychiatric morbidity (Coelho et al., 2014). Closely connected to functional changes in inflammatory systems are those in the stress response system, and within it the hypothalamic-pituitary-adrenal axis (Roelofs and Pasman, 2016; Keynejad et al., 2018). Limited empirical evidence suggests that basal diurnal cortisol is increased in patients with PNES, especially those with a history of sexual trauma (Bakvis et al., 2010), and correlated with heightened threat perception (Bakvis et al., 2009).

Lastly, traumatic events can also contribute to the manifestation of PNES through their subsequent mental representation, i. e. memories (Bowman, 1993). Seizures can be precipitated by trauma cues and ictal experience can include reliving of trauma memories (flashbacks) (Reuber et al., 2011, 2016). The suppression of intolerable memories has long been presumed to be a primary function of dissociation according to Pierre Janet, or, in Freudian terms, the “primary gain” of “conversion” (Bowman, 2018). Studies using functional MRI have shown that the neurophysiology behind wilful memory suppression in general involves activation of the lateral prefrontal cortex which suppresses hippocampal activity (Anderson et al., 2004). Similar network responses to traumatic memories have been observed in patients with dissociative movement disorders (“motor conversion disorder”) (Aybek et al., 2014). Furthermore, compared to healthy controls, patients also had accompanying activation of areas involved in motor planning and body schema (supplementary motor area and temporoparietal junction). This kind of motor system activation has been associated with the initiation of previously learnt motor behaviour independent of top-down voluntary control (Voon et al., 2011; Aybek et al., 2014).

Acute and persistent stress

Stress does not have to be severe or traumatic to contribute to the occurrence of PNES. A stress-diathesis model can help explain the variability in trauma and stress history in patients (Keynejad et al., 2018). In individuals with high biological susceptibility to stress-related pathology, even mildly or moderately stressful experiences may precipitate PNES. Conversely, even in the absence of particularly strong innate predispositions, chronic stress or major psychological trauma may lead to PNES or other functional neurological symptoms. Acute stress is frequently cited by patients and (more so) caregivers as a trigger of PNES, and bodily signs of distress are often reported as premonitory or ictal phenomena (Reuber and Rawlings, 2016). What is more, PNES are associated with acute physiological arousal which may be followed by parasympathetic activation as a PNES subsides (Ponnusamy et al., 2012; van der Kruijs et al., 2016). This supports the idea that dissociation in general and PNES in particular can serve as an allostatic mechanism that relieves distress, especially when other coping mechanisms are unavailable or have been overwhelmed (Goldstein and Mellers, 2006; Stone and Carson, 2013; Brown and Reuber, 2016b).

Dysfunctional relationships and attachment

In addition to leading to long-term problems with stress and emotion regulation, childhood trauma is known to be associated with the development of dysfunctional attachment styles in adulthood. In keeping with this, problems in interpersonal functioning are common among patients with PNES (Brown and Reuber, 2016a). Relationships within the family, with therapists, and in a wider social environment are often characterized by insecure attachment, social anxiety and avoidance. This can be a major source of emotional distress for patients (Green et al., 2017; Wardrope et al., 2019). In line with the ICM, PNES can remove patients from situations characterised by interpersonal challenges, thus reinforcing an unintentional “escape”-mechanism (Brown and Reuber, 2016b). This is evident in cases when PNES occur during psychotherapy, where stressors such as relationship conflicts and traumatic memories are often addressed (Kemp et al., 2018). Aside from the immediate stress relief afforded by a PNES, secondary reinforcement may be provided by the “sick role” conferred upon patients within their social environment.

Illness beliefs, social contagion and iatrogenesis

The power of suggestion in the precipitation of PNES is a well-known phenomenon, which is sometimes used as a diagnostic technique in epilepsy centres (Popkirov et al., 2015a). Purely “ideogenic” cases, whereby the conviction that one has a seizure disorder is sufficient to cause seizures, are very rare (Roach and Langley, 2004). Nonetheless, certain illness beliefs can be potent precipitating and perpetuating factors. In military veterans, for example, among 81 patients with PNES, 47% cited (usually mild) traumatic brain injury as a cause of their seizures (possibly influenced by the well-known association of head injury and epilepsy) (Salinsky et al., 2018). Misattribution effects like this can be reasonably extrapolated to head injuries in a civilian context (see below), especially with the recent media attention around mild traumatic brain injury (Popkirov et al., 2018a).

Another common perpetuator of physician-related illness misattribution is medication. Taking an anti-epileptic drug (AED) may contribute to the delay of correcting an epilepsy misdiagnosis (Bahrami et al., 2019). While it is conceivable that AEDs could predispose to PNES through (unwanted) physiological effects (Niedermeyer et al., 1970; Jabeen et al., 2018), the prescription of AEDs is likely to shape illness expectations, and there have even been cases when AED prescribed for a different reason (e.g. neuropathic pain) have contributed to the subsequent misinterpretation of seizure-like events as epilepsy (Oto et al., 2003). For this reason, once a diagnosis of PNES (and absence of epilepsy) has been established, AEDs given for suspected epilepsy should be tapered immediately, not least to support the change in illness perceptions (Thompson et al., 2009; Oto et al., 2010). Explaining that AEDs do not treat PNES and removing them can have positive effects on quality of life (Rawlings et al., 2017a) and long-term outcome (Duncan et al., 2016; Chen et al., 2018).

Psychiatric comorbidity

Dissociative and functional disorders

PNES are commonly associated with other dissociative and functional neurological (conversion) symptoms. Other dissociative disorders have been identified in a median of 33% of patients with PNES (Bowman, 2018). Additional functional, or “medically unexplained”, symptoms are found in 60-80% of patients with PNES (Bowman and Markand, 1996; Duncan et al., 2011; McKenzie et al., 2011). Conversely, a study of the first 100 consecutive outpatients evaluated at a speciality clinic for motor functional neurological disorders (FND) reported that 17 patients had PNES combined with a functional movement disorder and/or a functional limb weakness (Matin et al., 2017). In another study on 73 patients with motor FND, 17 (23%) also had PNES (Crimlisk et al., 1998). In a large series of 65 cases of functional movement disorders, three had coexisting PNES (5%), while 8/157 (5%) patients with PNES from the same study had a coexisting functional movement disorder (Driver-Dunckley et al., 2011). In a series of 107 cases of functional limb weakness, eight patients (7%) reported a “non-epileptic attack” at symptom onset (Stone et al., 2012). Somatization tendencies, identified by high levels of self-reported physical symptoms correlate positively with greater severity and poor outcome (Reuber et al., 2003a, 2003b; Brown and Reuber, 2016a).

Post-traumatic stress disorder (PTSD)

Considering the strong association between trauma and PNES, it is not surprising that PTSD is a common comorbidity. In nine studies on adults with PNES that included PTSD assessment (reviewed by Fiszman et al. [2004]), 79/207 (38%) fulfilled diagnostic criteria for current PTSD. Comparable rates are reported in recent studies (reviewed and meta-analysed by Diprose et al. [2016]), with a higher rate of 58-64% among military veterans (Salinsky et al., 2012; Salinsky et al., 2018). It has been suggested that patients with PNES and repeated trauma experience in childhood show symptom overlap across psychological domains with patients with so-called “complex” PTSD (Hingray et al., 2017). Patients with both PNES and epilepsy tend to have lower rates of PTSD than those without comorbid epilepsy (D’Alessio et al., 2006; Labudda et al., 2018), suggesting alternative aetiological routes (see below). PTSD comorbidity has potential therapeutic implications, as PNES patients with PTSD have higher levels of alexithymia and tend to apply more emotion-focused coping strategies compared to patients without trauma or with trauma but no PTSD (Zeng et al., 2018). Prolonged exposure therapy, a PTSD-specific therapy that relies on guided confrontation with activities, places and situations associated with trauma, has been shown to lead to high rates of seizure remission in patients with PNES and comorbid PTSD (Myers et al., 2017). Eye movement desensitization and processing (EMDR) -another psychotherapeutic technique used for PTSD- has also been suggested for the treatment of traumatized patients with PNES (Chemali and Meadows, 2004).

Personality disorder

A high prevalence of personality disorders among patients with PNES has been noted, with borderline personality disorder (BPD; also known as emotionally unstable personality disorder), or at least borderline personality traits, being a particularly common phenotype (although some studies report cluster C personality disorders [avoidant, dependent, or obsessive-compulsive] at similar rates) (Lacey et al., 2007; Bermeo-Ovalle and Kanner, 2018). Childhood trauma is common in BPD and is associated with higher rates of dissociative symptoms (Popkirov et al., 2018b). Conversely, emotion dysregulation, a hallmark of BPD, is frequently seen in patients with PNES and is likely to be an important factor in seizure pathophysiology (Reuber et al., 2004; Williams et al., 2018; Jungilligens et al., 2019). BPD is further characterised by a marked instability of interpersonal relationships, which is a common and clinically relevant finding in patients with PNES (Holman et al., 2008; Green et al., 2017; Wardrope et al., 2019). Dialectical-behaviour therapy, an intervention that specifically targets emotion dysregulation and interpersonal problems in BPD, can be helpful for patients with PNES (Bullock et al., 2015).

Depression and anxiety

Exact rates vary, but a recent meta-analysis has calculated overall rates of clinical depression of around 40% (Walsh et al., 2018). Comparative studies have revealed higher rates of depressive and anxiety disorders in patients with PNES compared to the general population (Bermeo-Ovalle and Kanner, 2018). Most studies of depression and anxiety disorders also found higher levels of these pathologies in patients with PNES than those with epilepsy (Diprose et al., 2016; Walsh et al., 2018). Compared to patients with epilepsy, depression in patients with PNES tends to manifest more through physical (rather than affective or cognitive) symptoms, is associated more closely with relationship problems (Green et al., 2017), and has a stronger impact on quality of life (Walsh et al., 2018). The frequency of anxiety disorders ranged between 9 and 71% in different studies (Bermeo-Ovalle and Kanner, 2018). In a recent meta-analysis of nine studies, 20% of patients with PNES had a comorbid panic disorder (Indranada et al., 2018). Like PTSD, panic disorder might play a particularly relevant role in the pathophysiology of some PNES (Indranada et al., 2018). Dissociative symptoms are common during panic attacks, and, conversely, anxiety symptoms are sometimes reported during PNES (Rawlings et al., 2017b; Indranada et al., 2018). The causal system underlying panic attacks (perceived threat -arousal- escape) can be found in some patients with PNES (Goldstein and Mellers, 2006; Brown and Reuber, 2016b).

Medical and neurological comorbidity

Epilepsy and epilepsy surgery

The combination of epilepsy and PNES is not uncommon and presents significant diagnostic and therapeutic challenges. When present, epilepsy almost invariably precedes the onset of PNES (Reuber et al., 2003c). In a recent meta-analysis of over 118 studies with a pooled sample size of 17,478 patients, comorbid epilepsy was reported in 22% of cases on average (95% confidence interval: 19-25). The pooled estimate of the prevalence of comorbid PNES in cohorts of patients with epilepsy was 12% (Kutlubaev et al., 2018). However, most studies originated from specialised epilepsy centres, so it is possible that the rate of patients with comorbid epilepsy would be lower in primary care, psychiatric or general neurology settings. In two population-based studies, the pooled rate of comorbid epilepsy was 14% (Kutlubaev et al., 2018). The rate was also lower (below 10%) in studies requiring video-EEG proof of additional epileptic seizures (Lesser et al., 1983; Benbadis et al., 2001). Conversely, the rate of comorbid epilepsy may be higher in particularly selected patient groups, for instance, those with learning disabilities (Duncan and Oto, 2008) or children (Vincentiis et al., 2006). The aetiological role of coexisting epilepsy is manifold, with illness-related chronic stress, psychiatric comorbidity, biological predisposition and symptom modelling all likely to play a role.

New onset of PNES can occur after epilepsy surgery in 2.4-8.8% of patients, with lower rates reported in larger studies (Ney et al., 1998; Glosser et al., 1999; Markoula et al., 2013; Asadi-Pooya et al., 2016). Brain surgery, other than epilepsy surgery, can precipitate PNES at similar rates (Reuber et al., 2002). Peri-operative stress or the change in psychosocial dynamics associated with becoming seizure-free have been considered as potential aetiological factors (Ney et al., 1998; Asadi-Pooya et al., 2016), as well as lesional changes to brain network connectivity (Popkirov et al., 2018a).

Traumatic brain injury (TBI)

A history of head injury has been reported in patients with PNES at rates of 16-83%, with a pooled frequency of 42% among 1,039 adults across 17 studies (Popkirov et al., 2018a). This is higher than the rate of 12% found among the general population (Frost et al., 2013). The reported injuries are classified as mild TBI in the majority of cases with the distribution of severity approximating that found in the general population (Vincentiis et al., 2006). Mild TBI can have various neuropsychiatric consequences that could contribute to the occurrence of PNES (Popkirov et al., 2018a). Personal illness beliefs can be structured around a head injury (see above) even in the absence of a biologically plausible pathophysiology (Salinsky et al., 2018), and an injury that occurs in association with significant stress (combat, accident, emergency treatment) can contribute to maladaptive learning processes that incorporate acute concussive symptoms into conditioned stress responses (Brown and Reuber, 2016b). Lastly, diffuse axonal injury associated with temporary metacognitive impairments could be relevant in some cases (Popkirov et al., 2018a).

Cognitive complaints

Subjective cognitive complaints are reported by a majority of PNES patients (60%), but standard neuropsychology rarely confirms corresponding objective deficits (Driver-Dunckley et al., 2011). Lack of effort or task engagement is only very rarely to blame - instead, cognitive problems such as forgetfulness and difficulties with concentration are thought to arise from attentional dysfunction and slower information processing, both strongly correlated with comorbid mood disorders, possibly compounded by medication (see Teodoro et al. [2018] for a review of cognitive symptoms and deficit in PNES). This pattern of cognitive problems is not unique to PNES, but can be found in patients with functional movement disorders, fibromyalgia and chronic fatigue syndrome (Teodoro et al., 2018) or in isolation as a functional cognitive disorder (Stone et al., 2015). Cognitive complaints correlate significantly with quality of life in PNES patients (Myers et al., 2012) and should receive special attention. Possible medication effects need to be addressed, and patients need to be educated about the nature of the problems, potential interactions with disordered sleep and mood, and coping strategies. Considering the role of stress and cognitive control in the pathophysiology of PNES, it is easy to see how symptomatic attentional dysfunction in everyday life might contribute to precipitation and perpetuation of PNES in some cases (Brown and Reuber, 2016a).

Sleep disturbance

Patients with PNES more frequently report problems with sleep (Latreille et al., 2018) and sleep apnoea symptoms than those with epilepsy (Karakis et al., 2014). A history of sleep disorder is noted in a third of patients (Elliott and Charyton, 2014; Latreille et al., 2018). Four small polysomnographic studies on a total of 68 patients have been conducted (Bazil et al., 2003; Phillips et al., 2013; Latreille et al., 2019; Popkirov et al., 2019). Clinically significant sleep-disordered breathing is found in 14-29% of patients (Phillips et al., 2013; Popkirov et al., 2019). A probable periodic limb movement disorder was found in 27% of the patients examined (Popkirov et al., 2019). Overall sleep quality is compromised, with patients showing a mean sleep onset latency of around 45 minutes, and spending an average of 1-2 hours awake per night after first falling asleep (Latreille et al., 2018; Popkirov et al., 2019). This might explain why a large proportion of patients take sleep medication (Latreille et al., 2019). Sleep disturbances are common symptoms of depressive and anxiety disorders, but the link with PNES may be more than simple comorbidity. Both experimental and clinical studies have suggested that sleep deprivation promotes dissociative tendencies (van der Kloet et al., 2012a), and one study on a mixed psychiatric inpatient sample has shown a decrease in dissociative symptoms after sleep normalization (van der Kloet et al., 2012b), suggesting that improvement of sleep might be of therapeutic use in patients with PNES.

Chronic pain

Individuals with PNES are much more likely to have chronic pain conditions than epilepsy controls (Benbadis, 2005; Kerr et al., 2017), with pain reported by up to 86% of patients (Dworetzky et al., 2005; Driver-Dunckley et al., 2011; Gazzola et al., 2012). Chronic pain is more common in women (Thomas et al., 2013), and is often part of a somatic symptom disorder or fibromyalgia (Ettinger et al., 1999; Mokleby et al., 2002; Benbadis, 2005). A substantial proportion of those with PNES (24-47%) uses regular pain medication, with opioids being used by 14-32% of patients in different case series (Hantke et al., 2007), similar to the levels of analgesic usage among patients with functional motor disorders (O’Connell et al., 2019). Common risk factors may contribute to the emergence of chronic pain and PNES (Ettinger et al., 1999; Mokleby et al., 2002). Alternatively, persistent pain (with or without associated insomnia) and disabilities associated with this symptom can be a source of acute and chronic stress predisposing, precipitating or perpetuating maladaptive coping responses including dissociation. Use (or abuse) of opioid medication could also play a role in the pathophysiology of PNES. Chronic inescapable trauma leads to chronic downregulation of the endogenous opioid system and increased stress-related opioid release, with implications for self-destructive behaviours, social attachment processes, and pain (Lanius et al., 2018). The opioid system has been implicated in defensive “shut-down” responses under conditions of inescapable stress and is thought to mediate alterations of consciousness during dissociative responses (Lanius et al., 2018). Pharmacological antagonism of opioids with naltrexone has been shown to reduce dissociative symptoms in BPD (Schmahl et al., 2012) and has been reported as beneficial in patients with daily refractory PNES undergoing inpatient psychotherapy (Straub and Bohlmann, 2009).

Migraine

About 50-60% of patients with PNES have comorbid migraine (Elliott and Charyton, 2014; Shepard et al., 2016), far exceeding the rate of 15% found in the general population (Steiner et al., 2013). One large study (n = 1,000) from India reported a very high rate of swoon-type reactions accompanying severe migraine attacks in women (13% of girls and 23% of women), but not in men (0% of boys; 1% of men), probably reflecting a culture-associated phenomenon (Chakravarty et al., 2010). Interestingly, patients with PNES report more frequent and longer lasting migraine attacks than those with epilepsy (Shepard et al., 2016). Severe pain, combined with the anxiety of unpredictable attack durations, can serve as a potent trigger (as suggested by the observations by Chakravarty et al. [2010]), while sensorimotor and cognitive aura symptoms (common in dissociative seizure patients [Shepard et al., 2016]) could contribute to the seizure scaffold. Neurophysiologically, individuals with migraine consistently show reduced habituation to sensory and noxious stimuli of various modalities, including defensive reflexes such as the blink reflex (Coppola et al., 2013). One speculative interpretation of the high rate of comorbid migraine in patients with PNES is thus that an inability to habituate to (i.e. tolerate) distressing interoceptive stimuli or thoughts is compensated through alternative system-wide adaptations such as dissociation (Goldstein and Mellers, 2006; Stone and Carson, 2013). Put simply, as one major adaptive dynamic (habituation) is impaired, other strategies are utilised to deal with persistent/intolerable distress.

Asthma

Asthma has been reported in a third of patients with PNES (de Wet et al., 2003; Elliott and Charyton, 2014). As with other common comorbidities, the recurrent distress of asthma attacks and the anxiety associated with an unpredictable and potentially life-threatening disorder can contribute to the emergence of PNES. The high prevalence of reported asthma in PNES patients could, however, have a different explanation. Vocal cord dysfunction (VCD), a functional respiratory disorder, is often misdiagnosed as asthma, although it typically presents with inspiratory dyspnoea and stridor (rather than the expiratory dyspnoea observed in asthma) accompanied by panic and agitation (Gimenez and Zafra, 2011). It shares several risk factors with PNES (anxiety, depression, history of abuse, female gender), and involves a vicious cycle of anxiety, hyperventilation, dysfunctional breathing and enhanced laryngeal reflexes (Bardin et al., 2017). The diagnosis of asthma / VCD in patients with PNES is complicated by the fact that all of these conditions are often associated with a tendency to hyperventilate (de Wet et al., 2003). A meta-analysis of seven studies examining ictal phenomena in patients with PNES has shown that 68% (CI: 55.4-79.5%) of patients had symptoms which could be related to hyperventilation (feeling dizzy or light-headed) (Indranada et al., 2018). Furthermore, intentional and unintentional hyperventilation can precipitate or accompany PNES and deliberate hyperventilation is an effective provocation procedure for PNES during video-EEG recordings (Popkirov et al., 2015b; Indranada et al., 2018).

Critical appraisal of the literature

The nature and direction of the relationship between PNES and most of the comorbidities reviewed here cannot be derived from correlational or epidemiological studies. Comorbidities may be coincidental without any pathophysiological inter-relation; they can directly or indirectly contribute to PNES (e.g. concurrent epilepsy); they can be caused by PNES (e.g. secondary social phobia); they can have a bidirectional causal association (e.g. depression); or be associated with each other through shared risk factors (e.g. other functional neurological disorders). Furthermore, in some cases, PNES cannot be clearly separated from a “comorbid” condition - either pathophysiologically or clinically (e.g. when PNES develop out of a classic panic attack or represent an embodied flashback in PTSD). In addition to these conceptual considerations, the methodological limitations of the available literature have to be considered. The rates of occurrence of most risk factors and comorbidities reviewed above have been reproduced in reasonably-sized populations. However, different definitions and classifications have been used, especially in relation to psychiatric comorbidities. Many studies rely on subjective symptom reporting rather than clinical diagnoses. The association with potential risk factors has typically relied on retrospective recall of such factors rather than prospective longitudinal research. Findings are usually not adjusted for socioeconomic status, although there is evidence that PNES are more common among poorer individuals who may have different rates of trauma and comorbidities relative to control populations with a different socioeconomic background (Sigurdardottir and Olafsson, 1998; Szaflarski et al., 2000; Duncan et al., 2011; Goldstein et al., 2019). Many published patient series were recruited in specialist settings where patients with more complex disorders may be over-represented, for instance those with a higher rate of comorbid epilepsy (Kutlubaev et al., 2018). Finally, (under-reported) cultural differences may play an important role, such as the high incidence of PNES accompanying migraine attacks in Indian women (Chakravarty et al., 2010). The cultural aspects of PNES aetiology tend to be neglected in academic research, but they can be of substantial explanatory, diagnostic and therapeutic value in clinical practice. Seizure semiology could be shaped by culturally established gestures and other idioms of distress (e. g. shielding of the face), and social situations and family dynamics need to be seen through the lens of culture, religion and politics to understand their psychosocial impact (Kanemoto et al., 2017; Martínez-Taboas et al., 2019).

Risk factors and comorbidities in clinical practice

Diagnostic value of risk factors and comorbidities

The mainstay of diagnosis of PNES is the semiological seizure analysis (clinical features can have a cumulative predictive value approximating 100% [Avbersek and Sisodiya, 2010]) supported, when necessary and possible, by ictal video-EEG (LaFrance et al., 2013; Popkirov et al., 2017). While at group level PNES are associated with a comorbidity profile, which differs from that associated with epilepsy, comorbidities and risk factors alone are not sufficient to make the diagnosis of PNES and may in fact be misleading. However, the diagnostic process does not end with the label “PNES” (and much less with that of a “non-epileptic event”): the ultimate diagnostic goal is to understand the biopsychosocial aetiology of the condition in an individual patient. Of course, neurologists do not face this task alone, but they need to be aware of how major risk factors and comorbidities can interact with seizures in those with PNES. Neither nosology nor nomenclature currently provide an appropriate reflection of the heterogeneity of PNES, so an individualised understanding of the patient's condition requires careful consideration of how risk factors and comorbidities interact in each particular case.

Risk factors and comorbidities for choice of treatment

Manualized treatments exist for PNES (Reiter et al., 2015), addressing not only the seizures, but also the risk factors and comorbidities, illustrating that individual aetiological factors need to be considered when planning treatment for PNES (LaFrance and Bjønæs, 2019). There is an increasing body of evidence supporting the therapeutic value of a range of psychotherapeutic interventions for PNES. A recent meta-analysis demonstrated that a pooled mean of 47% of patients treated became free of PNES after intervention and 82% experienced a greater than 50% improvement in seizure frequency (Carlson and Nicholson Perry, 2017). Irrespective of modality, indispensable components of treatment are: communicating the diagnosis with sufficient regard for the patient's understanding and acceptance; involving carers and devising strategies for acute seizures; involving therapists early; and treating medical comorbidities in appropriate and complementary ways (Reuber, 2019).

Competencies and learning points

Learning objective: “Describe the epidemiology, psychiatric and experiential risk factors of PNES”:

  • Describe the incidence and estimated prevalence of PNES and its importance as one of the most common causes of transient loss of consciousness.
  • Describe the prevalence and importance of common experiential risk factors and how they can interact with PNES pathophysiology.
  • Describe common medical and psychiatric comorbidities, and how they can relate to PNES.
  • Describe how risk factors and comorbidities influence the diagnostic workup and individual treatment planning.

Key points

  • PNES are episodes of impaired awareness and behavioural control that can resemble epilepsy or other paroxysmal disorders, although they have neuropsychiatric underpinnings.
  • PNES are commonly misdiagnosed as epilepsy or syncope (the opposite diagnostic error occurs as well, but is less frequent).
  • PNES are the result of complex and heterogeneous neurocognitive dysfunction often involving elements of abnormal stress responses and emotion processing combined with attentional and metacognitive problems.
  • Psychiatric comorbidities are common, particularly dissociative, depressive, anxiety and posttraumatic stress disorders, and personality disorders (especially borderline pattern).
  • A history of trauma and stressful life events are frequently but not invariably reported, and interpersonal relationships are often characterized by insecure attachment.
  • Comorbid epilepsy (almost invariably preceding the development of PNES) occurs in roughly 10% of adults and up to 30% of children with PNES, and may present an important diagnostic and therapeutic challenge.
  • Various comorbidities such as chronic pain, sleep problems, migraine, asthma, and history of head injury are found at higher rates in patients with PNES than in the general population.
  • In an individual patient, it is difficult to establish whether a co-occurring condition is a true comorbidity, a predisposition, or an underlying cause of PNES.
  • The identification of risk factors and comorbidities of PNES alone does not allow a reliable distinction from other causes of transient impairment of awareness and self-control, but their characterisation and consideration are crucial in the planning of individual treatment.
  • There are evidence-based treatments for PNES which address underlying risk factors and comorbidities.

Case studies

Case 1

A 23-year old man presents to the epilepsy-monitoring unit (EMU) with suspected pharmacoresistant post-traumatic epilepsy. Eight months ago, he was involved in a serious road traffic accident (RTA). In the RTA he suffered a concussion (loss of consciousness for about 15 minutes and peritraumatic amnesia), but sustained no further injuries. CT and MRI revealed no abnormalities at the time. Since then, he has experienced high levels of anxiety, low mood, fatigue, nightmares, disturbed sleep, intrusive memories and irritability. He has stopped driving his car and avoided traffic in general. Four months ago, he started having seizures which were usually preceded by palpitations, a sensation of a lump in the throat, and feeling “spaced out”. He would then become unresponsive and develop slight symmetrical shaking of his limbs. Attacks would last between two and 20 minutes and occur between one and six times a month. When seizures lasted for more than a few minutes, the patient's family called the emergency services. His past medical history includes asthma in childhood and migraine without aura (currently four attacks/month). After the onset of the seizures, he was started on levetiracetam for suspected epilepsy and zopiclone for his sleeping problems.

On Day 2 of the EMU monitoring, the patient becomes unresponsive shortly after hyperventilation has been performed and a typical seizure is observed. The patient later reports having experienced his usual prodromal warning signs. Based on the characteristic semiology (and the absence of epileptic activity on the EEG), PNES are diagnosed. Since only one type of seizure is reported, the epilepsy diagnosis is reversed. A psychiatric evaluation confirms the diagnosis of post-traumatic stress disorder.

The patient and his family are informed about the diagnoses and potential treatments. Levetiracetam is discontinued and the irritability is markedly reduced. He is started on venlafaxine (as a treatment for PTSD and migraine). The neurologist discusses emergency strategies regarding future seizures, explains a few simple “grounding” techniques, and the patient is referred for cognitive-behavioural therapy (figure 2).

Case 2

A 36-year-old teacher with long-standing temporal lobe epilepsy undergoes right selective amygdalohippocampectomy. After surgery, she remains seizure-free (outcome Engel Class IA), but complains of memory impairments that make teaching very stressful. Six months after the operation, she insists on reducing her topiramate, hoping to improve her memory performance, and is explicitly warned that seizures could recur. Shortly after reducing the dose, she suffers a seizure. She reports having first felt slight dizziness and a dry mouth, and was then “simply not there” for a few minutes. Her sister reports that the patient had tears in her eyes immediately following the seizure. This kind of seizure started to occur about once a month despite the reversal of the topiramate dose reduction. During a five-day EMU stay, no seizure can be recorded until one is elicited using verbal suggestion on the sixth day. The ictal video is shown to the patient and her sister, and both confirm that this is the “new kind” of seizure. Ictal EEG shows no epileptic activity.

The patient is diagnosed with new-onset PNES. After a long discussion with the neuropsychologist, she realizes that her attacks are usually preceded by worries about “the epilepsy coming back” and anxiety related to work. She declines cognitive-behavioural therapy, explaining she’ll try to “figure it out” herself first. On follow-up, six months later, she reports no further seizures since the diagnosis (figure 3).

Case 3

A 19-year-old woman is brought to the emergency department (ED) with a convulsive seizure. According to the emergency personnel, the seizure began after an altercation with her boyfriend. At the time of arrival in the ED, the seizure has lasted approximately 30 minutes and is ongoing. The patient's eyes are tightly shut, there are continuous thrashing movements of all limbs of variable amplitude and stable frequency, and her back is arched intermittently. Two doses of intravenous lorazepam (2x4 mg) administered by the receiving ED doctor, on the basis of a misdiagnosis of status epilepticus, aggravate the seizure manifestations. A neurologist and anaesthetist are called. Before the patient can be sedated and ventilated, the neurologist arrives and makes the correct diagnosis of a prolonged PNES. The neurologist asks the patients’ relatives to step out and the ED treatment team to step back from the patient. He addresses the patient by her name, explains that she is in a safe place and that she is experiencing a non-epileptic seizure. He reassures her that this seizure will stop shortly and will not damage her brain. The neurologist tells the patient that he would like to explain more about non-epileptic seizures once the seizure is over. The seizure manifestations subside over a further five minutes.

When she recovers sufficiently, the patient reports that she has been having these kinds of seizures with increasing frequency over the last year, and that they are currently occurring twice a week. Sometimes she is partially aware of what is happening in the seizures. They seem to occur at any time and she does not recognise specific triggers although they have been more frequent at times when she has felt stressed. A doctor has suggested antiepileptic drug treatment but she was not convinced that she had epilepsy. Her medical records reveal a history of severe childhood abuse, previous self-harm, recurrent syncope, and asthma.

The diagnosis of PNES is explained to the patient and her boyfriend. The verbal explanation is supported by a leaflet about PNES and information about a PNES website. The patient is referred to a psychiatrist for further evaluation (figure 4).

Recommended references

Asadi-Pooya AA, Sperling MR. Epidemiology of psychogenic nonepileptic seizures. Epilepsy Behav 2015; 46: 60-5.

Avbersek A, Sisodiya S. Does the primary literature provide support for clinical signs used to distinguish psychogenic nonepileptic seizures from epileptic seizures? J Neurol Neurosurg Psychiatry 2010; 81(7): 719-25.

Popkirov S, Carson AJ, Stone J. Scared or scarred: could ’dissociogenic’ lesions predispose to nonepileptic seizures after head trauma? Seizure 2018; 58: 127-32.

Reuber M, Brown RJ. Understanding psychogenic nonepileptic seizures-Phenomenology, semiology and the Integrative Cognitive Model. Seizure 2017; 44: 199-205.

Schachter SC, LaFrance JWC. Gates and Rowan's nonepileptic seizures. 4th Ed. Cambridge: Cambridge University Press, 2019.

Supplementary data

Summary didactic slides are available on the www.epilepticdisorders.com website.

Acknowledgements and disclosures

SP has received a speaking fee from Novartis. AA-P has received honoraria from Cobel Darou and RaymandRad; Royalty: Oxford University Press (Book publication). RD has received an honorarium from UCB PHARMA. DG reports no disclosures. CH has no disclosures relative to PNES. AMK received honoraria from Eisai laboratories for participating on the Scientific advisory Board, from Neuropace for participation in symposia, and from the Epilepsy Foundation of America, for being co-Editor-In-Chief of Epilepsy.com. None have a conflict with the topic of this manuscript. WCL has served on the editorial boards of Epilepsia, Epilepsy & Behavior, Journal of Neurology, Neurosurgery and Psychiatry, and Journal of Neuropsychiatry and Clinical Neurosciences; receives editor's royalties from the publication of Gates and Rowan's Non-epileptic Seizures, 3rd ed. (Cambridge University Press, 2010) and 4th ed. (2017); author's royalties for Taking Control of Your Seizures: Workbook and Therapist Guide (Oxford University Press, 2015); has received research support from the NIH (NINDS 5K23NS45902 [PI]), Department of Defense (W81XWH-17-0169 [PI]), Rhode Island Hospital, the American Epilepsy Society (AES), the Epilepsy Foundation (EF), Brown University and the Siravo Foundation; serves on the Epilepsy Foundation Professional Advisory Board; has received honoraria for the American Academy of Neurology Annual Meeting Annual Course; has served as a clinic development consultant at University of Colorado Denver, Cleveland Clinic, Spectrum Health and Emory University; and has provided medico legal expert testimony. CP reports no disclosures. MR has benefited from an educational grant from UCB Pharma and has accepted speaker's fees from Eisai, UCB Pharma and LivaNova. He receives royalties for “In our words: Personal Accounts of Living with Non-epileptic Seizures”, “Epilepsy Explained” and “Borderlands of Epilepsy Revisited” from Oxford University Press and is paid for his work as Editor-in-Chief of “Seizure - European Journal of Epilepsy” by Elsevier.

Disclaimer.

This report was written by experts selected by the International League Against Epilepsy (ILAE) and was approved for publication by the ILAE. Opinions expressed by the authors, however, do not necessarily represent the policy or position of the ILAE.