Psychopathology and Outcome in Psychogenic Nonepileptic Seizures

Elizabeth S. Bowman

Andres M. Kanner

Psychogenic nonepileptic seizures (PNESs) is a psychiatric disorder that presents with symptoms that resemble epileptic seizures but are not due to physiological dysfunction of the nervous system. Because they resemble epileptic seizures, they are initially diagnosed by neurologists and subsequently treated by psychiatrists or other mental health professionals. This chapter explores the psychiatric illnesses found in patients with PNES and discusses the prognosis of this troublesome illness.

Of what use is a discussion of the psychiatric illnesses associated with PNES? Neurologists are unlikely to wish to treat PNES directly, but they need to know the comorbid psychiatric illnesses when explaining the nature of these paroxysmal events to patients, screening for psychiatric emergencies in these patients, and making referrals for mental health care.

Understanding the more common pseudoseizure-associated psychiatric illnesses is essential for mental health clinicians who often find themselves faced with the daunting task of sorting out multiple coexisting psychiatric conditions in patients with pseudoseizure. For psychiatrists, determination of comorbid diagnoses may play a critical role in treatment outcome, because it is often the emotional pain created by these illnesses that is expressed through PNES. Reduction of such distress through pharmacological and psychotherapeutic treatment of depression, panic disorder, or post-traumatic stress disorder (PTSD) can bring about an immediate reduction of PNES. Some “PNESs” appear to be misdiagnosed symptoms of other psychiatric illnesses (e.g., panic attacks or dissociative trances) that require diagnosis and treatment.

How Common are Psychiatric Illnesses in Patients with Psychogenic Nonepileptic Seizure?

Do persons with PNES have more psychiatric illnesses than the general population? To clinicians, the answer may appear to be “obviously yes.” However, no direct comparative studies have addressed this question. Studies of patients with PNES repeatedly find rates of current psychiatric disorders that greatly exceed those of the general population in the National Comorbidity Survey
(NCS) of the United States (1). Table 26.1 shows reported rates of any current psychiatric illness in 16 studies of patients with PNES (2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17). These studies find some kind of psychiatric illness in 43% (2) to 100% (8,11,16,17) of PNES subjects.

TABLE 26.1 Studies Reporting Prevalence of Psychiatric Diagnoses, Past Treatment, and Past Hospitalization in Patients with Pseudoseizure

Study Reference	Any Current Diagnosis (%)	Any Lifetime Diagnosis (%)	Prior Psychiatric Treatment (%)	Prior Psychiatric Hospitalization (%)
2	43	71	—	—
3	46	—	—	25
4	47^a	—	—	—
5	60	—	50	30
6	67	—	≥51	19
7^b	67	—	—	—
8	70^c	—	—	—
9	71	—	51^d	—
10	76	—	—	—
11	80^e	—	—	30
12	84	—	—	—
13	91	—	—	—
14	95	—	—	—
15	96	—	—	—
16	100	—	—	—
17	100	—	—	—
11	100^f	—	70	—
8	100^g	—	—	—
19	—	—	38	—
20	—	—	—	37
21	—	—	—	32
29	—	—	58
^aDoes not include conversion diagnoses. ^bSystematic Structured Clinical Interview for DSM-IV diagnoses used. ^cDoes not include mental retardation diagnoses. ^dPrior psychotherapy only. ^eExcludes somatoform diagnoses. ^fIncludes somatoform diagnoses. ^gIncludes mental retardation diagnoses.

The methodological difficulties of comparing the results of PNES studies are formidable because the studies used diverse patient populations, differ greatly in the psychiatric illnesses for which they screened, and differ in the methodological rigor and
systematic nature of their screening. For instance, only three studies listed in Table 26.1 clearly reported data for current psychiatric illness and for a past (lifetime) diagnosis of psychiatric illness (2,6,16). In the remainder of the studies, readers are left to presume that only current diagnoses are being reported. Even a choice to include mental retardation as a psychiatric illness greatly changes results. In Pakalnis et al. (8), inclusion of mental retardation raised the psychiatric diagnosis rate from 70% to 100%. However, even a study that excluded mentally retarded persons found a 100% rate of psychiatric illness by doing a comprehensive diagnostic assessment that included a wide array of illnesses (16).

Most PNESs are classified in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) (18) as conversion disorder with seizures, as misdiagnosed panic disorder, or as part of another somatoform or dissociative disorder. Therefore, it would seem obvious that all persons with PNES should carry some psychiatric diagnosis, even if it is solely that of conversion disorder. So how can we explain that the median rate of psychiatric diagnoses in 16 studies is 73.5%, rather than 100%? Study aims and methodologies play a huge role. Many studies overlook the fact that the patient has a somatoform (i.e., psychosomatic) disorder and report only findings of illness in other categories such as anxiety and affective or personality disorders. As shown by the study of Pakalnis et al. (8), when the distinction between nonsomatoform and somatoform disorders is taken into account, the rate of overall psychiatric illness rose from 80% to 100%. This suggests that results from studies in which somatoform illness was not taken into account show a skew in the direction of unrealistically low overall rates of psychiatric illness in patients with PNES. Yet another methodological bias is failure to study Axis II diagnoses (personality and developmental disorders).

The NCS study of the prevalence of psychiatric illnesses in the general US population found that 50% of adults report at least one lifetime mental disorder and that close to 30% report having at least one disorder in the past 12 months (1). NCS studies assessed a wider range of psychiatric illnesses than most of the PNES studies, so NCS rates of having a current or lifetime mental illness generally would be expected to exceed those found in PNES studies. Therefore, it is striking that the rates of having any mental illness in the NCS study are lower than the lowest rates of psychiatric illness found in the 16 PNES studies listed in Table 26.1. These data strongly suggest that patients with PNES have far more comorbid psychiatric illness than do people in the general population.

Prior Psychiatric Treatment and Hospitalizations

Another way to estimate the level of psychiatric morbidity in patients with PNES is to look at rates of past psychiatric treatment. Only 20% to 40% of people with mental illnesses seek treatment, so rates of treatment seeking seriously underestimate rates of illness (1). Despite this, six studies of patients with PNES listed in Table 26.1 found that an astounding 38% to 70% of patients had previously sought mental health treatment (5,6,9,11,260,262). Four of these studies found treatment rates of 50% or higher. These findings may reflect the unusually severe illness of the tertiary care subject populations in these studies, but they also point to the patients with PNES having multiple and serious psychiatric illnesses. High psychiatric comorbidity in patients with PNES was demonstrated in one study that found a mean of 4.4 current and 6.0 lifetime Axis I psychiatric diagnoses (16). Unfortunately, few studies of patients with PNES from primary care settings exist for comparison.

A third way to estimate the frequency and seriousness of psychiatric illnesses in patients with pseudoseizure is to assess their rates of past psychiatric hospitalization as an indication of symptoms severe enough to disable or endanger them. Seven studies of past psychiatric hospitalizations (Table 26.1) show rates of 25% to 37% among patients
with PNES (3,5,6,11,20,21,22). No less than one fourth of them have been psychiatrically ill enough to warrant hospitalization. Whether we examine studies of rates of psychiatric diagnosis, of past treatment, or of hospitalizations, patients with PNES appear to be psychiatrically ill considerably more often than their general population counterparts. It is uncertain but likely that patients with PNES have higher rates of psychiatric illness than general medical or neurological outpatients.

What Types of Psychiatric Illnesses Occur in Patients with Pseudoseizure?

What kinds of comorbid psychiatric illnesses should clinicians screen for in a newly diagnosed patient with PNES? Studies indicate that more than one type of pathology is likely to be seen. The results of studies of psychiatric diagnoses in patients with PNES need to be interpreted cautiously because of diverse research methodologies and diverse patient populations.

The patient populations studied are diverse in gender proportions, seizure chronicity, proportions with comorbid epilepsy, inpatient versus outpatient status, methods of PNES diagnosis, age, and specialization of the treatment setting. Most studies reflect the relatively symptom-chronic population biases of tertiary care treatment settings. The methods used to diagnose psychiatric illnesses vary greatly in rigor and reliability. The diagnosing clinicians are not always mental health clinicians, so sophistication in psychiatric diagnosis varies. Further diagnostic confusion has occurred because different diagnostic systems in North America and Europe result in different diagnostic labels for the same symptoms.

Another methodological problem is researcher bias. For instance, most studies assess depression and panic attacks because these are common symptoms that are familiar to researchers. Therefore, they are assessed and found in more studies than are diagnoses that are less common. It is less common for patients with PNES to be assessed for substance abuse, PTSD, eating disorders, or dissociative disorders because even mental health clinicians overlook these illnesses or are inadequately trained in diagnosing them. This bias in investigation likely skews the literature toward overrepresenting depression, panic disorder, and personality disorders as the most common psychopathologies associated with PNES because they are the illnesses for which researchers have looked. Few studies have systematically screened for a wide variety of illnesses or used validated and reliable diagnostic instruments, such as the Structured Clinical Interview for DSM-IV (SCID) (23,24). The few studies that took this approach found very high rates of psychiatric illnesses (2,16) and a wider array of diagnoses. Some illnesses may be reported infrequently (e.g., eating, substance, or dissociative disorders) because they are recognized only when patients present with severe and obvious symptoms.

In light of these confounding methodological problems, studies have yielded a wide range of rates and types of psychiatric illnesses. To answer the question of what specific psychiatric diagnoses are seen with PNES, we present in the following text summations of literature findings of psychiatric diagnoses in patients with pseudoseizure, arranged in clusters of illness types as found in DSM-IV (18). In reviewing these data, keep in mind the distinction between current (met diagnostic criteria within the past month) and lifetime diagnoses (met criteria in the past, with or without current illness). Lifetime rates are always equal to or higher than current illness rates.

Mood (Affective) Disorders

Depression is reported in more pseudoseizure studies (19 studies) (Table 26.2) than any other type of psychiatric illness (2,4,5,6,7,9,10,11,12,14,15,16,17,19,20,21,25,26,27). Unfortunately, the definition of “depression” varies too greatly among studies for valid comparisons of their results. Studies that specify DSM-IV major

depression or major affective disorder, or use a validated diagnostic instrument, report current rates of depression of 45% to 56% (5,9,16). The range of reported current depression (0% to 100%) illustrates the methodological problems outlined earlier. The median rate of current depression in these studies is 31%. Therefore, clinically significant depression is present in one third to one half of persons with newly diagnosed PNES. Lifetime rates of depression are much higher (36% to 80%), reflecting the recurrent nature of depression.

TABLE 26.2 Selected^a Affective and Anxiety Disorders in Pseudoseizure Subjects

Study Reference	Depression/Major Depression (%)	Dysthymia (Current) (%)	Bipolar Disorders (%)	Any Anxiety Disorder (%)	Panic Attacks or Disorder (%)	PTSD (%)	Any Phobia (%)
25	0 Current	—	—	—	—	—	—
12	12 Current	—	—	22 Current	—	—	—
7	15^b Current	—	—	11	—	—	—
20	16 Current	—	—	—	—	—	—
2	21 Current	0	7 Current and lifetime	—	14 Current	14 Current	7 Current
	36 Lifetime	—	—	—	21 Lifetime	36 Lifetime	—
10	24^cCurrent	—	—	12 Current	—	—	—
4	25 Current	—	—	—	—	—	2 Current
26	33	—	—	11^dCurrent	—	—	—
19	35	—	4 Current	—	—	—	—
27	44^e	—	—	—	—	—	—
5	45^fCurrent	5	—	10^gCurrent 80^hCurrent	70 Current	—	—
16	47^fCurrent	13	2 Current and lifetime	47ⁱCurrent	20 Current	49 Current	33 Current
	80^fLifetime	—	—	51ⁱLifetime	27 Lifetime	58 Lifetime	42 Lifetime
11	50^j	—	—	—	80 Current	—	—
9	56^k Current	—	—	19 Current	—	—	—
15	60^f Current	7	2 Current	—	2 Current	—	—
6	65^e Lifetime	—	—	39ⁱ Life	26 Current	—	16 Current
17	100^l Lifetime	—	—	—	—	—	—
21	58 Lifetime	—	—	—	—	—	—
14	74^e,m Lifetime	—	—	—	—	—	—
28	—	—	4 BP NOS	28ⁱ Current 36^h Current	24 Current	8 Current	—
22	—	—	12 Current	18 Current	—	—	—
30	—	—	—	43^h Lifetime	—	10 Current	—
31	—	—	—	—	>90 Lifetime	—	—
33	—	—	—	—	∼33 Current	—	—
BP NOS, bipolar disorder not otherwise specified.
^aFor generalized anxiety disorder, obsessive-compulsive disorder, or anxiety disorder not otherwise specified, see discussion in text. ^bDenotes “major depression or dysthymia.” ^cDenotes “depressive syndrome.” ^dDefined as “anxiety.” ^eDenotes “depression.” ^fUtilized valid Structure Clinical Interview for DSM-III-R interviews. ^gExcludes panic subjects. ^hIncludes post-traumatic stress disorder (PTSD) and panic. ⁱExcludes PTSD. ^jDenotes “current or lifetime mood disorder.” ^kDenotes “major affective disorder.” ^lDenotes “clinical affective syndrome.” ^mIncludes bipolar disorder depressions.

Other findings in Table 26.2 show that PNES populations have mood disorders other than major depression. Dysthymia (a chronic low-grade depression) occurs in up to 13% of subjects (2,5,15,16). In many studies, it is likely lumped together with other depressive syndromes, so its true incidence is unclear. Bipolar disorder also accounts for some affective disorder diagnoses, but its occurrence in patients with PNES is uncommon (median 4%) and is near its 1% baseline rate in the general population (1,2,15,16,19,22,28). This is what would be expected for a mostly genetic illness that likely has a chance association with PNES.

The 1-year and lifetime prevalences of major depression in the general population are approximately 10% and 17%, respectively, and the prevalences for any affective disorder are 11% and 19%, respectively (1). Therefore, the high prevalence of depression in the PNES population is far more than a chance association and suggests that clinicians should always evaluate patients with PNES for depression. PNESs generally occur in persons who have a life of multiple or severe stresses (or traumas) or who have a pattern of being unable to express emotions adequately (29). PNESs are like an emotional “pop-off valve” for relief of pent-up emotional distress. The same losses and stresses that produce PNESs are likely to cause depression. The painful affects associated with depression, in turn, add to the patient’s emotional load and may contribute to having more PNESs. When depression is present, it should be treated with antidepressant medications and often with concomitant psychotherapy. Adequate treatment of depression may help decrease PNES because it decreases the pressure of painful feelings.

Anxiety Disorders

Selected anxiety disorder diagnoses in PNES subjects are listed in Table 26.2 (2,5,6,7,9,10,11,12,15,16,22,23,26,28,30,31). Anxiety has been noted often in these patients. This is not surprising, because PNESs are basically a somatic outlet for unmanageably intense feelings such as anxiety, sadness, and anger (32). The range of occurrence of anxiety disorders in patients with PNES is very broad: 0% to 80%. A more useful statistic is the median of the studies’ findings: an 18.5% prevalence of current anxiety disorders.

Data on anxiety disorders in patients with PNES are difficult to interpret because studies in which diagnoses were included are inconsistent. Some studies focus solely on panic disorder and generalized anxiety disorder (GAD) (31), whereas others separately included all DSM-IV anxiety diagnoses (16). A particular problem is that PTSD is not included in many diagnostic instruments and was omitted from many studies of other anxiety disorders. Most studies found an anxiety disorder other than PTSD in one tenth to one fifth of subjects. When PTSD is included, the rates of any anxiety disorder rise to approximately 33% to 47%. PTSD is found in 8% to 49% of PNES subjects (2,16,28,30,33), and an increasing number of investigators are commenting on the association of trauma with PNES (2,16,34,35,36). Some behaviors diagnosed as PNESs are actually dissociative flashbacks of trauma (one of the symptoms of PTSD). Therefore, clinicians should screen for PTSD in patients with PNES who have a history of trauma, especially in patients whose seizures resemble abreactions (33,35).

The anxiety disorder most commonly mentioned in PNES studies is panic disorder, which is found in 14% to 90% of subjects (2,5,6,11,15,16,28,31). The three studies that found high rates of panic disorder focused specifically on this disorder (5,11,31). Studies that assessed panic disorder
(recurrent panic attacks resulting in dysfunction) rather than isolated panic attacks generally find a prevalence of panic disorder in less than one fourth of subjects. Panic attacks cause symptoms (e.g., trembling, depersonalization, and fear) that are often confused with partial complex seizures, so persons with panic disorder may be mistakenly diagnosed with PNES (5). Patients with PNES should be questioned about symptoms of panic, keeping in mind the differences in their presentations: severely altered consciousness (i.e., amnesia) is not present during panic attacks, panic usually has a slower onset than PNES, and diaphoresis is a symptom of panic but not of PNES.

Some symptoms of generalized anxiety are often observed clinically in patients with PNES, but they are often lumped with symptoms of phobia, panic, and PTSD, and described globally as an anxiety disorder. Two studies (15,16) each found the DSM-IV (18) diagnosis of GAD, involving pervasive worry with somatic anxiety symptoms, in 9% of patients with PNES as a separate disorder from other anxiety disorders. This is likely the case because DSM criteria dictate that GAD is not diagnosed if anxiety is related to panic, PTSD, obsessive-compulsive disorder, or phobias. Although many patients with PNES are “worriers,” their anxiety may be a symptom of PTSD or panic rather than represent GAD. Clinicians should inquire carefully to assess if anxiety is related to panic or PTSD, because these disorders are treated differently than GAD. Obsessive-compulsive disorder, a largely biological illness, is found in patients with PNES at a rate of 4% (16), similar to its prevalence in the general population. Anxiety disorder not otherwise specified has been found in 2% (16) to 5% (15) of PNES subjects.

Patients with PNES suffering from panic or PTSD should receive a selective serotonin uptake inhibitor, such as sertraline or escitalopram, or serotonin norepinephrine reuptake inhibitors such as venlafaxine and they should be referred for psychotherapy. These medications are efficacious for depression disorder, panic disorder, GAD, and PTSD. Use of benzodiazepines in GAD, panic disorder, or in patients with PTSD should be a short-term strategy and should be used in conjunction with seeking psychotherapy. We do not recommend extended use of benzodiazepines without psychotherapy, because this results in medication dependence without resolution of the cause of the anxiety. The treatment of PNES and comorbid psychiatric disorders is discussed in great detail in Chapter 27.

Somatoform Disorders

Rates of somatoform disorders in PNES subjects are listed in Table 26.3 (4,5,7,8,10,11,12,15,16,19,21,22,25,31,37,38,39,40). PNES is, by definition, a somatic expression of psychological distress that usually falls under the DSM-IV diagnosis of conversion disorder. So how can we explain the findings of Table 26.3 that only one of ten studies found that 100% of patients with PNES carry a somatoform disorder diagnosis, and that the median finding of these studies is that one third have a somatoform diagnosis?

Research subject recruitment explains a great deal. These patients are diagnosed with PNES because they have symptoms that result in referral to a neurology clinic. However, they include persons whose symptoms are undiagnosed panic attacks, unrecognized dissociative amnesia, flashbacks of trauma, malingering, factitious symptoms, or difficulties attributed by researchers to personality disorders. Some subjects have misdiagnosed physical conditions (such as cardiogenic syncope) rather than PNES (40). None of these subjects would have conversion seizures, and not all would have a somatoform disorder. In addition, somatoform diagnoses can be missed unless researchers systematically ask about other conversion or somatic symptoms.

Conversion seizures account for 33% to 84% of patients with PNES (8,16,25,39). Patients with PNES have strikingly high lifetime rates (42% to 93%) of other types of conversion (10,16,21,31). Conversion seizures and other conversion symptoms exist concurrently in up to one fifth of subjects.

In subjects with concurrent conversion, conversion pseudo-Todd’s paralysis, numbness, deafness, and blindness may occur as part of their ictal or peri-ictal presentation. Vein et al. (31) noted a mean of 5.9 conversion symptoms in PNES subjects. These data suggest that patients with conversion seizure are probably a population that recurrently expresses distress through conversion, and who happen to have been studied when they presented with seizure symptoms. Therefore, over the patient’s lifetime, neurologists might expect to see untreated patients with PNES repeatedly with different forms of conversion as life stresses wax and wane. During evaluations, clinicians should ask patients with PNES about other past conversion symptoms and other somatic symptoms.

TABLE 26.3 Selected Somatoform^a and Dissociative^b Disorder Diagnoses in Pseudoseizure Subjects

Study Reference	Any Current Somatoform Disorder (%)	Nonseizure Conversion (%)	Conversion Seizures (%)	Somatoform Pain Disorder (%)	Any Current Dissociative Disorder (%)	Dissociative Disorder NOS (%)	Current Dissociative Amnesia (%)
10	2	69 Lifetime	—	—	—	—	—
12	6	42 Lifetime	—	—	—	—	—
7	17	—	—	—	11	—	—
15	22	—	—	—	22	—	22
19	23^c	—	—	—	—	—	—
37	33^d ^e	19 Current	—	14^d ^f	—	—	—
38	36	—	—	—	36	27^g	—
25	59	—	33 Lifetime	—	—	—	—
22	79	—	—	71	—	—	—
16	89	4 Current	78 Lifetime	16^f	91	62^h	13 Diagnosis
—	—	82 Lifetime	—	—	—	—	98 Symptomⁱ
11	100	—	—	—	—	—	—
31	—	93 Lifetime	—	—	—	—	53^g^,^j
4	—	21 Current	—	—	—	—	—
8	—	—	55 Lifetime	—	—	—	—
39	—	—	84 Lifetime	—	—	—	—
56	—	—	—	—	100	—	—
41	—	—	—	77	—	—	—
—	—	—	—	61^f	—	—
^aFor reports of undifferentiated somatoform disorder and somatoform pain disorder, see discussion in text.
^bFor reports of depersonalization disorder, see discussion in text. ^cDefined as Briquet’s disorder. ^dPediatric subjects. ^eDefined as any nonseizure psychosomatic diagnosis. ^fDefined as chronic headaches. Bowman and Markand (16) found 71% prevalence of headaches but diagnosed somatoform pain disorders in only 16%. ^gMethods or systematic efforts at diagnosing dissociative disorders are unclear. ^hSystematic inquiry using the Structured Clinical Interview for DSM-IV dissociative disorders. ⁱSee text for discussion on differentiating the symptom and the DSM-IV diagnosis of dissociative amnesia. ^jDenotes amnesia during pseudoseizures.