Clinical Assessment of Axis I Psychiatric Morbidity in Chronic Epilepsy: A Multicenter Investigation

Patients with a diagnosis of epilepsy, regardless of seizure type, were recruited through tertiary outpatient epilepsy centers from five universities (Medical College of Georgia, Rush Presbyterian St. Luke’s Medical Center, Stanford University, Washington University, University of Wisconsin-Madison). Inclusion criteria included chronological age 18 years and older, treated only with antiepilepsy medications (excluding patients who underwent surgery or were implanted with vagal nerve stimulator), on stable antiepilepsy medications for the previous 30 days, and intact reading ability with Wide Range Achievement Test-3 (WRAT-3) reading scores >69.

One hundred seventy-four individuals met all selection criteria, proceeded through informed consent procedures, and fully completed all interviews and questionnaires. Table 1 summarizes the demographic and clinical epilepsy characteristics of the study sample. Overall, patients in this investigation suffered from chronic epilepsy (approximately 18 years of epilepsy), almost one-half were on polytherapy (46%), and the majority (55%) had complex partial seizures.

Trained investigators at each site interviewed study participants with the Mini International Neuropsychiatric Interview (MINI) in order to identify current DSM–IV Axis I disorders.¹⁹ Additionally, each subject also underwent an interview using the Mood Disorders module of the Structured Clinical Interview for DSM–IV Axis I Disorders-Research Version (SCID-I)¹² in order to identify lifetime-to-date and current DSM–IV Axis I mood disorders. The SCID-I represented the research standard to which MINI diagnoses of depression were compared. Study interviewers at each site were experienced in standardized psychiatric assessment. However, for the purposes of this investigation all interviewers underwent centralized training in the use of the MINI and mood disorders module of the SCID-I by experienced psychiatric research clinical interviewers. Finally, the Beck Depression Inventory-II²¹ and Center for Epidemiological Studies-Depression Scale (CES-D)²² were completed by each participant in order to capture symptoms of current depressive symptoms. Further details regarding these measures are provided below.

The MINI was developed as a short and efficient diagnostic interview to be used in clinical as well as research settings.¹⁹ It follows DSM–IV and the ICD-10 criteria for psychiatric disorders, screening for 17 Axis I disorders, with brief suicidality and antisocial personality modules. The Axis I disorders included in the MINI were selected based upon the 12-month prevalence reported in the Epidemiologic Catchment Area Study²³ and the National Comorbidity Survey.²⁴ The MINI has been validated in the U.S. and Europe and is available in 20 languages.²⁵ Administration time ranges from approximately 15–20 minutes for individuals with few positively endorsed symptoms to 20–30 minutes for individuals who meet criteria for current diagnoses.

The Structured Clinical Interview for DSM–IV Axis I Disorders-Research Version is a well known semistructured interview that provides a framework upon which to make DSM–IV Axis I diagnoses.¹² Only the Mood Disorders module of the SCID-I was administered in this study in order to determine the validity of MINI mood diagnoses. The interview required approximately 15–25 minutes in order to complete the module. The SCID-I is arguably the standard semistructured psychiatric interview procedure most used for research purposes. The MINI and SCID-I were administered in a single session, and each subject was interviewed privately and independently.

The Beck Depression Inventory II is a well-known self-report instrument designed to assess and detect the severity of current (past 2 weeks) depressive symptoms in clinical, medical and community settings.²¹ The BDI-II contains 21 descriptive statements regarding depressive symptoms frequently reported by individuals diagnosed with depression. Each of the items contains a four-point severity rating scale, with administration time ranging from 5 to 10 minutes. Scores above 19 indicated moderate severity of depressive symptoms.²¹

Self-rated depressive symptoms were also assessed using the CES-D. The CES-D is a 20-item scale developed to screen for depression in primary care settings.²² It has been shown to be a valid and reliable instrument ^22,26,27 and scores of >16 are most frequently used to indicate depression.

A summary of current MINI Axis I DSM–IV disorders among the epilepsy patients is provided in Table 2. Nearly one-half of the sample had no current axis I diagnosis (N=89 or 51.1%), while the remaining half (N=85 or 48.9%) exhibited current diagnoses. Comorbidity across diagnoses was frequent. The results reported in Table 2 reflect this comorbidity in that the percentages across the Axis I disorders exceed 100% indicating that a large number of participants have more than one disorder (N=59).

Among the mood disorders, and in fact among all individual diagnoses, major depression was the most common disorder (17.2%), with a considerably smaller proportion meeting criteria for current dysthymia (4.0%) or bipolar disorders (2.8%).

Current anxiety disorders were especially prevalent with 91 of diagnoses falling in this category. Agoraphobia (15.5%), generalized anxiety disorder (13.2%), and social phobia (10.9%) were the most common diagnoses among the anxiety disorders.

Within the anxiety disorders a substantial proportion, 25 of 52 persons or 48.1%, met criteria for two or more anxiety disorders. There was also considerable comorbidity between mood and anxiety disorders. Of those patients with current depressive disorders, the majority (27 of 37 or 72.9%) were also diagnosed with a current anxiety disorder. Schizophrenia and other psychotic disorders were uncommon (N=1 or 0.6%).

Using the mood disorders module of the SCID, 45.4% of the sample was diagnosed with lifetime-to-date mood disorders of any type, 27.0% with a past major depressive episode (39.7% lifetime-to-date [includes current]).

Mini International Neuropsychiatric Interview and SCID-I diagnoses were compared using the mood disorder modules of each instrument, including current and past diagnoses where applicable. As shown in Table 3, current major depressive episodes were identified with similar frequency by the MINI (17.2%) and SCID-I (16.7%). Table 4 demonstrates significant concordance (Kappa) between the MINI and SCID-I for all DSM–IV mood diagnoses captured by both instruments. The highest concordance was evident for major depressive episodes-current (0.86) and manic episodes-past (0.79) (Table 4). While significant, considerably lower concordance between MINI and SCID-I was seen for other diagnostic categories due, at least in part, to low frequencies of diagnosis in this population.

The degree of concordance for major depressive episode was also examined for individual subjects. Twenty-three of 30 (76.7%) individuals with a major depressive episode were diagnosed by both instruments, with 7 subjects diagnosed by one or the other interview procedure (4 on MINI only, 3 on SCID only).

Less than one-half of the individuals identified as meeting criteria for a current major depressive episode were being treated with antidepressant medications. Only 13 or 43.3% of subjects diagnosed with a current major depressive episode by the MINI were taking an antidepressant medication, and only 14 or 48.3% of subjects identified by the SCID-I as experiencing a current major depressive episode were taking antidepressants. These results underscore previous reports that current depression is undertreated in this population. Interestingly, 20 (11.5%) individuals who did not meet criteria for a current major depressive episode were taking antidepressant medications. Of this group, 60% had a past major depressive episode based on the SCID-I.

Subjects diagnosed with a current major depressive episode had significantly (p<0.01) higher scores on the BDI and the CES-D as would be expected (Figure 1). For individuals identified with a current major depressive episode by the MINI and SCID-I, the mean CES-D, scores were 30.96 and 29.38, respectively. Similar results were noted on the BDI-II, with essentially equal mean scores for individuals with MINI and SCID-I diagnoses of current major depressive episode (28.27 and 27.68, respectively).

Rates of current Axis I DSM–IV disorders using the MINI were found to be common, affecting essentially one-half of the sample. This rate of current Axis I disorders is elevated compared to findings derived from population-based studies such as the National Comorbidity Survey in which 48% of patients had a lifetime history of Axis I disorders.²⁴

Most but not all investigations using standardized DSM-based structured interviews in the epilepsy literature have involved patients who were candidates for epilepsy surgery, most often with intractable temporal lobe epilepsy (e.g., references 13–18). Comparatively less research has been devoted to characterizing the distribution of Axis I disorders in the general population of patients with epilepsy attending specialty clinics. Nonetheless, the rate reported here (49%) is within the range (19%–62%) of current Axis I disorders in the few other available studies of patients with epilepsy that have reported this statistic.^17,28,29 Overall, these findings warrant a heightened index of concern for psychiatric morbidity in comparable specialty clinic settings.

Of the epilepsy patients with identified current disorders, 21.2% met criteria for depressive disorders, including current major depression (17.2%). Interestingly, anxiety disorders were unexpectedly common current disorders, with agoraphobia (15.5%), generalized anxiety disorders (13.2%), and social phobia (10.9%) being the most common. As noted previously, most studies examining patients with temporal lobe epilepsy, have reported lifetime-to-date mood disorders to be more common than anxiety disorders. This pattern has been reported by several investigators including Victoroff³⁰ (63% mood versus 32% anxiety), Altshuler et al.¹³ (44% versus 5%), Silberman et al.¹⁸ (62% versus 15%), and Hermann et al.⁴ (46% versus 13%). The rate of current major depression in these epilepsy subjects (17.2%) is considerably higher than that reported in the general population (10%).²⁴

The high rate of anxiety disorders found here was unexpected. This sample included individuals with a range of seizure types. Individuals studied here did not undergo ictal monitoring as part of the research protocol, so the relationship between mood versus anxiety disorders and seizure type could not be directly addressed in this investigation. However, there were no significant differences between individuals with complex partial seizures with and without secondary generalized seizures in terms of the frequency of mood (p=0.74) and anxiety (p=0.20) disorders in general or more specifically major depression (p=0.57).

In summary, in this sample of individuals with chronic epilepsy, major depression (17%) was the most common current Axis I DSM–IV diagnosis, followed by agoraphobia (15%), generalized anxiety disorder (13%) and social phobia (11%). The broader trend, however, revealed overrepresentation of anxiety (52.1%) compared to mood disorders (24.0%).

In the epilepsy literature there has been very little examination of the degree of comorbidity across psychiatric diagnoses. Interestingly, the rate of comorbid MINI current mood and anxiety disorders was high in this sample, with 27 of 37 (73%) patients with depressive disorders also diagnosed with a comorbid current anxiety disorder. This combination of comorbid major depression and anxiety was more common than depressive disorders alone (10 of 37 or 27.0%). In addition, within the anxiety disorders there were a large number of comorbid diagnoses. A substantial proportion, 25 of 52 persons or 48.1%, met criteria for two or more current anxiety disorders. Patterns of comorbidity are of considerable interest in the general psychiatric literature,²⁴ but remain to be fully characterized among patients with epilepsy.

Consistent with previous reports,^2,6,7–9 we found that mood disorders in epilepsy, including current major depression, often went untreated. Fewer than one-half (43%) of the patients with current major depression in this investigation were treated with antidepressant medications. The reasons for the undertreatment of depression in the general population have been reviewed previously,³¹ but remain to be characterized in epilepsy. While high rates of untreated depression typically raise the index of suspicion that psychiatric comorbidity is inadequately screened for by attending physicians,³¹ one interesting finding suggests that epilepsy specialists were concerned about and looking for comorbid mood disorders. A subset of 20 (11.5%) individuals was identified who were prescribed antidepressant medications, but they did not meet criteria for current major depression or a current mood disorder more generally. The nature of this subset of individuals remains uncertain and could reflect persons with subsyndromal depressive disorders, mood disorders that are adequately treated and no longer clinically manifest, or perhaps receiving antidepressant treatment that is not clinically warranted. Review of SCID-Iata for these 20 individuals, however, revealed that 60% suffered a past (but not current) major depressive episode, suggesting that treatment may have been appropriately targeted in the majority of cases. This subset of patients also raises the possibility that mood disorders may be even more common than our figures indicate.

Another interesting consideration is related to the individuals (43%) with current major depression who were treated with antidepressant medications. Despite receiving treatment, these individuals continued to meet criteria for major depression, suggesting that their depression had not responded to adequate treatment, the antidepressant treatment was suboptimal, patients suffered from drug resistant depression, or perhaps patients were noncompliant with their antidepressant treatment.

Overall, these results suggest that treating physicians are attempting to identify and treat comorbid mood disorders in particular, but they overlook one-half of the existing cases with current significant mood disorders, the vast majority of whom also suffer from comorbid anxiety disorders, and apparently are having difficulty controlling depressive symptoms in treated patients.

Several investigative groups in the epilepsy literature have pointed to the need for improved recognition of depression, unquestionably an important goal. However, the primary care literature has demonstrated that attempts to improve recognition of depression do not invariably translate to improved treatment of depression.³¹ In the epilepsy literature, intervention papers typically focus on the standard of care for the treatment of depression in epilepsy,^2,32 with the assumption that this information will be incorporated into the usual and typical care of patients with epilepsy by their primary physicians, subsequently improving psychiatric outcomes. However, randomized treatment trials in other (non-neurological) health care settings have challenged this assumption. These studies have demonstrated the cost-effective and treatment-effective superiority of programs of so-called collaborative care (e.g., combined psychiatry and primary medical) over improved usual care for the management of major depression.^33–36 Such issues have yet to be examined seriously in epilepsy, as do issues related to the additional health care costs and associated functional disabilities^37,38 that have been demonstrated to be associated with major depression in other patient populations. Finally, anxiety disorders appear especially common in this population, and similar data collection regarding the recognition and adequacy of treatment of these disorders remains an important task for the future.

Several limitations of this investigation, as well as opportunities for future research, should be recognized. First, the sample investigated here consisted of individuals with chronic epilepsy recruited from tertiary care centers. Findings regarding rates of Axis I disorders therefore cannot be assumed to generalize to the general population of persons with better-controlled epilepsy in the community. Second, this investigation did not focus on predictors of current mood disorders (e.g., localization and lateralization of seizure onset), but the reliability of MINI diagnoses may make such efforts feasible in clinical settings. Third, irritability and anger related to interictal dysphoric disorder⁴⁰ were not evaluated in this study and are not specifically addressed in the MINI. Finally, the high rate of anxiety disorders, either alone or in combination with mood disorders, suggests that directed study of these disorders remains an important task for the future.