SEARCH

SEARCH BY CITATION

Keywords:

  • Epilepsy;
  • Psychopathology;
  • Epidemiology;
  • Comorbidity;
  • Neuropsychiatry

Summary

  1. Top of page
  2. Summary
  3. Method
  4. Results
  5. Discussion
  6. Implications and Future Research
  7. Acknowledgments
  8. Disclosure
  9. Author Contributions
  10. References

Purpose:  In a nationally representative population-based study in England, we estimated the burden of psychiatric and neurodevelopmental comorbidities in people with epilepsy. We investigated whether any overrepresentation of comorbidities could be explained by epilepsy being a chronic medical or neurologic condition, or by the confounding effect of demographic and socioeconomic factors or other health conditions.

Methods:  The Adult Psychiatric Morbidity Survey 2007 comprised detailed interviews with 7,403 individuals living in private households in England. Doctor-diagnosed epilepsy (and asthma, diabetes, and migraine, chronic conditions for comparison) was ascertained by self-report, and extensive diagnostic and screening interviews were used to assess psychiatric and neurodevelopmental conditions.

Key Findings:  The estimated lifetime prevalence of epilepsy in the adult (≥16 years old) population of England was 1.2% (95% confidence interval [CI] 1.0–1.5). Almost one-third of the people with epilepsy had an International Classification of Diseases, Tenth Revision (ICD-10) anxiety or depressive disorder (compared with one in six people without epilepsy). Among these, social phobia and agoraphobia, generalized anxiety disorder, depression, and measures of suicidality had strong associations with epilepsy, which remained robust after accounting for potential confounders. These associations were consistently stronger than those in people with asthma or diabetes, and similar to those in people reporting migraine or chronic headaches. Epilepsy was also strongly associated with autism spectrum disorders (odds ratio [OR] 7.4, 95% CI 1.5–35.5) and possible eating disorders, and these associations were not evident in people with asthma, diabetes, or migraine.

Significance:  Psychiatric and neurodevelopmental conditions were overrepresented in people with epilepsy. These associations were stronger than with other nonneurologic chronic conditions, and not explained by confounding. Some overlap in the psychopathology observed in epilepsy and migraine cannot rule out the presence of common pathways of psychiatric comorbidity in neurologic conditions. However, associations of epilepsy with conditions such as autism spectrum disorders point to comorbidities specific to epilepsy that may not be shared by other neurologic conditions.

Epilepsy is the most common serious neurologic disorder, and it affects >50 million people worldwide (World Health Organization 2005; Ngugi et al., 2010). Psychiatric disorders are commonly encountered in people with epilepsy, and these may negatively influence the course of epilepsy, lead to inadequate response to treatment, and contribute to a poor quality of life as well as increased mortality (Christensen et al., 2007; Thapar et al., 2009). However, psychopathology is frequently unrecognized and untreated in people with epilepsy (Hermann et al., 2000), and many questions remain about both the extent and the nature of the relationships between epilepsy and specific psychiatric conditions. A good understanding of the burden of these comorbidities is essential for better recognition and treatment, and to inform the training needs of clinicians. This has the potential to translate into improved care of people with epilepsy. In addition, observed associations may illuminate understanding of common mechanisms behind epilepsy and specific forms of psychopathology.

Population-based studies in adults have often focused on comorbid depression in epilepsy (reviewed in Tellez-Zenteno et al., 2007), but potentially important relationships with other conditions such as anxiety (Beyenburg et al., 2005) and neurodevelopmental disorders (Garcia-Morales et al., 2008; Bolton et al., 2011) have not been adequately examined. The reported prevalence of comorbid psychopathology in epilepsy also varies widely across studies, reflecting methodologic limitations including possible selection (due to recruitment of epilepsy cases from unrepresentative clinical populations) (Swinkels et al., 2005; Tellez-Zenteno et al., 2007; Garcia-Morales et al., 2008) or measurement (lack of diagnostic assessment tools to ascertain psychopathology) (Gilliam et al., 2003; Gaitatzis et al., 2004b; Lacey et al., 2009; Ottman et al., 2011) biases. Furthermore, studies without adequate control populations (Gilliam et al., 2003; Gaitatzis et al., 2004b) may have led to inaccurate estimates of the risk of psychopathology in people with epilepsy compared with the general population.

Two other important issues remain unexplored. First, because socioeconomic disadvantage and health problems are common in people with epilepsy (Scambler & Hopkins, 1980; Gaitatzis et al., 2004a) as well as psychiatric disorders (Muntaner et al., 2004; Prince et al., 2007), confounding by these factors may explain the observed relationships, but this has not been previously considered. Second, depression and anxiety disorders are commonly associated with many chronic conditions (Prince et al., 2007). Depressive symptoms (assessed by a screening instrument) appeared to be more common among people with epilepsy than among people with asthma in one previous study (Ettinger et al., 2004), but this issue has not been explored in studies using diagnostic instruments, or for disorders other than depression. Specifically, no study has investigated how comorbid psychopathology in epilepsy compares with that in people with chronic neurologic presentations such as migraines or chronic headaches. Therefore, it is largely unknown whether epilepsy is associated with an increased risk of psychopathology in addition to the effect of having a chronic medical or neurologic condition.

We hypothesized that epilepsy would be associated with psychiatric and neurodevelopmental disorders, even after adjusting for potential confounders. We also hypothesized that epilepsy would have stronger associations with depressive and anxiety disorders than similar associations in people with other chronic conditions such as asthma, diabetes, and migraine. We tested these hypotheses using face-to-face interview data from a large nationally representative population-based study in England, which used validated diagnostic and screening measures to ascertain a range of psychiatric and neurodevelopmental conditions.

Method

  1. Top of page
  2. Summary
  3. Method
  4. Results
  5. Discussion
  6. Implications and Future Research
  7. Acknowledgments
  8. Disclosure
  9. Author Contributions
  10. References

The Adult Psychiatric Morbidity Survey (APMS) 2007

The sample for APMS 2007 was designed to be representative of the population 16 years and older living in private households in England. A multistage stratified probability sampling design was used. The sampling frame was Royal Mail’s small user Postcode Address File (PAF), with postcode sectors (on average comprising 2,550 households) representing the primary sampling units. Postcode sectors were first stratified by Strategic Health Authority (SHA). All of the primary sampling units within each SHA were then further stratified on the basis of the proportion of persons in nonmanual occupations and sorted by the proportion of households without a car based on 2001 Census data. Postal sectors were then sampled from each stratum with a probability proportional to size (where size was measured by the number of delivery points). In this way, 519 postal sectors were selected. Within these selected postal sectors, 28 delivery points were then randomly selected, yielding 14,532 delivery points. Interviewers visited these addresses to identify private households with at least one person aged ≥16 years. Within the potentially eligible sample of 12,694 addresses, one person from each household where contact was made was randomly chosen to take part in the survey, and 7,461 (57%) individuals agreed to be interviewed. The interviews were conducted in two phases. The first phase included structured diagnostic and screening assessments for a range of psychiatric and neurodevelopmental disorders, along with measures of general health and demographics. The second phase interviews were carried out on a subsample by clinically trained research interviewers using semistructured diagnostic instruments. The probabilities of selection to the second phase interviews were based on respondents’ responses to screening questions in phase 1. Fieldwork was completed between October 2006 and December 2007. Further details of the survey methodology are available elsewhere (Brugha et al., 2009; McManus et al., 2009). Ethical approval for the survey was obtained from the Royal Free Hospital and Medical School Research Ethics Committee.

Ascertainment of epilepsy, asthma, diabetes, and migraine

The first phase interviews included a section on general health, in which participants were asked if they had any specific physical health conditions listed on a show card. Those indicating that they had epilepsy/fits since the age of 16 were then asked if they had received a diagnosis of epilepsy by a doctor, the age of onset of the problem, whether they had fits in the last year, and if they had been on medication for the same in the last year. Similar questions have been employed and found to be valid in other population-based studies of epilepsy in Canada and the United States (Tellez-Zenteno et al., 2007; Ottman et al., 2011). People who reported that they had epilepsy since age 16 and had received a diagnosis of epilepsy by a doctor were considered cases of epilepsy in this study. A similar methodology was used to identify people reporting a lifetime doctor-diagnosed history of migraines or frequent headaches, asthma, and diabetes, respectively. Asthma and diabetes are chronic conditions known to have a higher burden of comorbid psychopathology (Prince et al., 2007), and we chose to study them as comparison groups against our findings for comorbidity in epilepsy. Furthermore, migraines and chronic headaches are conditions that may be neurologic in origin, but that may also reflect neurotic psychopathology, and we studied this group as a comparison group (Antonaci et al., 2011).

Measures of psychopathology

Measures described below are detailed further in the APMS reports (Brugha et al., 2009; McManus et al., 2009).

Depression and anxiety disorders

The revised Clinical Interview Schedule (CIS-R) (Lewis et al., 1992), a structured psychiatric interview designed to be used by trained lay interviewers was used to identify people meeting International Classification of Diseases, Tenth Edition (ICD-10) (World Health Organization 1992) diagnostic criteria for depressive episodes, social phobia, specific phobia, panic disorder, agoraphobia, generalized anxiety disorder (GAD), and obsessive compulsive disorder (OCD). In addition, when CIS-R scores indicate significant psychopathology (≥12) (Lewis et al., 1992), but diagnostic criteria of any of the above disorders is not met, a diagnosis of mixed anxiety and depression is applied, although the ICD-10 does not give specific diagnostic criteria for this category (World Health Organization 1992). To avoid confusion, we refer to these presentations as nonspecific psychiatric morbidity.

Suicidality

Respondents were asked questions about a lifetime or previous year history of suicidal thoughts, suicide attempts, and self-harm in the face-to-face interview.

Autism Spectrum Disorder (ASD)

The Autism Diagnostic Observation Schedule, Module 4 (ADOS) (Lord et al., 2000), a semistructured diagnostic instrument for ASD was used in the phase 2 subsample (Brugha et al., 2011). The subsample was selected from phase 1 responses to a validated self-reported screening tool for ASD (Brugha et al., 2011). The ADOS ratings corresponding to Diagnostic and Statistical Manual of Mental Disorders, 4th ed. (American Psychiatric Association, 2000) criteria were summed to produce an overall score, and the recommended threshold of ≥10 was used to indicate a case of adult ASD (Brugha et al., 2011).

Psychotic disorders

The Schedule for Clinical Assessment in Neuropsychiatry (SCAN) interview was administered in the phase 2 subsample for a confirmatory diagnosis of psychotic disorder in the preceding year (Wing et al., 1990). The subsample was selected based on phase 1 screening questions, which included current use of antipsychotic medication, inpatient psychiatric admissions, self-reported diagnoses, or psychotic symptoms, including auditory hallucinations (McManus et al., 2009).

Screening for eating disorders

The SCOFF (Morgan et al., 1999) questionnaire was used to screen for eating disorders. We defined people with a possible eating disorder if they scored ≥2 on the SCOFF (the recommended cutoff) and answered affirmative to an additional item—“did your feelings about food interfere with your ability to work, meet personal responsibilities and/or enjoy a social life?”—capturing the potential impact on social functioning related to eating problems.

Screening for posttraumatic stress disorder (PTSD)

A brief self-report screening questionnaire, the Trauma Screening Questionnaire (TSQ) (Brewin et al., 2002), was used to indicate presence of trauma-related symptoms in the past 2 weeks and that clinical assessment for PTSD was warranted.

Screening for adult attention-deficit/hyperactivity disorder (ADHD)

The Adult Self-Report Scale-v1.1 (ASRS) (Kessler et al., 2005) was used to screen for symptoms suggestive of ADHD. A recommended score of ≥4 was considered to be a positive screen, indicating that a clinical assessment for ADHD may be warranted.

Possible confounders

Because previous studies have not studied confounding, we identified a number of potential sociodemographic and health confounders a priori, since these are associated with both epilepsy (Scambler & Hopkins, 1980; Gaitatzis et al., 2004a) and psychopathology (Muntaner et al., 2004; Prince et al., 2007). These included age, gender, marital status, employment status (currently working, or either economically inactive or unemployed), tenure of accommodation (owner or renting property), highest educational qualification, household income, and occupational social class. A somatic comorbidity variable was made in which respondents were categorized as having 0, 1, or 2 or more current comorbid doctor-diagnosed somatic illnesses (items included cancer, diabetes, stroke, angina or myocardial infarction, hypertension, bronchitis or emphysema, asthma, peptic ulcer or other gastrointestinal problems, liver disease, bowel or colon diseases, bladder problems or incontinence, arthritis and migraine, or frequent headaches). Similar variables were also made for comorbid somatic illnesses in people with migraine or frequent headaches, asthma, and diabetes.

Analysis

All analyses were conducted using the survey commands in STATA 10.1 for Windows (StataCorp LP, College Station, TX, U.S.A.). Data were weighted to account for study design and nonresponse, so that the results were representative in terms of age, sex, region, and area characteristics of the household population 16 years or older in England (McManus et al., 2009). First, sample weights were applied to account for the different probabilities of selecting respondents in different-sized households. Second, to reduce household nonresponse bias, a household level weight was calculated from a logistic regression model using interviewer observation and area-level variables (collected from Census 2001 data) available for responding and nonresponding households. The nonresponse weight for each household was calculated as the inverse of the probability of response estimated from the model, multiplied by the household’s selection weight. Finally, weights were applied using calibration weighting based on age, sex, and region to weight the data up to represent the structure of the national population, taking into account the differential nonresponse between regions and age-by-sex groups. The population control totals used were the Office for National Statistics 2006 midyear household population estimates. For psychosis and ASDs, (conditions ascertained in phase 2 interviews), additional weights were designed to take into account the probability of selection of participants for phase 2 interviews. Complete details of the weighting procedures are available elsewhere (McManus et al., 2009).

In descriptive analysis, we first studied sociodemographic characteristics of people with epilepsy and compared them with people without epilepsy. We then estimated the prevalence of each psychiatric condition in people with epilepsy. We used weighted logistic regression models to estimate odds ratios (ORs) and 95% confidence intervals (95% CIs) of the association of epilepsy with each comorbid condition. As recommended by the literature, we used multiple measures of socioeconomic position to minimize residual confounding due to these factors (Galobardes et al., 2007), and included all of the potential confounders in adjusted models to estimate adjusted ORs and 95% CIs. There was a higher proportion of missing data for income and occupational class, but omission of these variables in our final adjusted models presented here had no substantial effect on the results. For analysis of neurodevelopmental conditions (ASD and ADHD), we used verbal IQ estimated using the National Adult Reading Test (Crawford et al., 1988) as an additional confounder. In separate comparative analyses, we used a similar methodology to estimate associations in people with asthma, diabetes, and migraine.

Results

  1. Top of page
  2. Summary
  3. Method
  4. Results
  5. Discussion
  6. Implications and Future Research
  7. Acknowledgments
  8. Disclosure
  9. Author Contributions
  10. References

Doctor-diagnosed epilepsy, present since age 16, was reported by 101 people corresponding to a weighted prevalence of 1.2% (95% CI 1.0–1.5) in the general population of England. About 3% of our original sample had missing data in covariates, and we included 94 people with epilepsy and 7,071 people without epilepsy with complete data in our analysis. People with epilepsy were more likely to be divorced or separated (as compared to being married), have lower educational qualifications, live in rented accommodation, and be economically inactive compared with people without epilepsy (Table 1).

Table 1.   Characteristics of people with epilepsy as compared to the general populationwithout epilepsy in England.
 Epilepsy (n = 94) %No epilepsy (n = 7,071)a %OR (95% CI)
  1. OR (95% CI), odds ratio with 95% confidence interval.

  2. aLower n’s for income (n = 5,744) and for occupational class (n = 6,809).

  3. bOccupational social class based on U.K. Registrar General’s classification (I/II = professionals or intermediate professions, III = nonmanual and manual skilled occupations, IV/V = partially skilled or unskilled occupations).

Age (years)
 16–3432.130.8Ref
 35–5445.135.71.2 (0.7–2.2)
 55 or more22.933.50.7 (0.3–1.3)
Gender
 Female52.451.61.0 (0.7–1.6)
Marital status
 Married/cohabiting54.763.1Ref
 Single/widowed34.129.61.3 (0.8–2.3)
 Divorced/separated11.27.31.8 (1.0–3.2)
Household income
 Tertile 1 (highest)33.036.1Ref
 Tertile 225.632.70.9 (0.5–1.6)
 Tertile 3 (lowest)41.431.21.5 (0.9–2.5)
Employment
 Economically inactive49.538.71.6 (1.0–2.4)
Occupational social classb
 I/II37.438.7Ref
 III45.840.61.2 (0.7–2.0)
 IV/V16.920.70.8 (0.4–1.7)
Highest educational qualification
 Degree level21.827.9Ref
 A/0 Level45.946.11.3 (0.7–2.4)
 No qualifications32.326.11.6 (0.9–2.9)
Tenure of accommodation
 Rented accommodation43.529.41.9 (1.2–3.0)
Comorbid physical illness
 None55.857.1Ref
 125.125.31.0 (0.6–1.6)
 2 or more19.217.61.1 (0.6–1.9)

Generalized anxiety disorder (12.5%), nonspecific psychiatric morbidity (12.2%), and depressive disorders (9.6%) were the most common psychiatric comorbidities in people with epilepsy (Table 2). When compared with the rest of the population without epilepsy—and after adjusting for potential confounders—social phobias, agoraphobia, generalized anxiety disorder, and depression were the common mental disorders with the strongest independent associations with epilepsy (Table 2). People with epilepsy also had strong relationships with all measures of lifetime and recent suicidal thoughts and behavior. These associations were largely similar with those observed in the group of people reporting migraines, or chronic headaches, bar some differences, although the CIs overlapped. Specifically, the association with social phobias appeared to be much stronger, and nonspecific psychiatric morbidity, panic disorder and specific phobias appeared to be weaker in people with epilepsy as compared to those reporting migraines or chronic headaches (Table 3). In contrast, similar associations were either absent or much weaker in people with asthma and diabetes (Table 3). The associations between epilepsy and all measures of suicidality were attenuated after adjusting for concurrent depression or anxiety disorders (Table 4).

Table 2.   Prevalence of psychiatric conditions in people with epilepsy, and results of logistic regression analysis comparing psychopathology in people with epilepsy with the general population of England without epilepsy.
Mental disorderPrevalence in people with epilepsy % (95% CI)Crude OR (95% CI)Adjusted OR (95% CI)
  1. Adjusted OR, model adjusted for age, gender, marital status, highest educational qualification, employed or economically inactive, and number of comorbid physical illnesses. Verbal IQ score further included in the final model for association between autism spectrum disorder and attention deficit disorder screen (it had no confounding effect on other conditions under study).

  2. OR, odds ratio; 95% CI, 95% confidence interval.

  3. *p < 0.05; **p < 0.01, ***p < 0.001.

Any depressive or anxiety disorder30.6 (21.5–41.4)2.3 (1.4–3.8)**1.9 (1.2–3.2)*
Depression and an anxiety disorder4.2 (2.2–7.2)4.1 (2.1–8.2)***3.1 (1.5–6.3)**
Depression9.6 (5.3–16.9)3.7 (1.9–7.1)***2.7 (1.4–5.4)**
Generalized anxiety disorder12.5 (7.6–20.1)3.3 (1.8–5.9)***2.6 (1.5–4.7)**
Social phobia6.0 (2.7–12.8)7.1 (3.0–16.7)***5.2 (2.1–13.1)***
Specific phobias1.8 (0.6–5.7)2.0 (0.6–6.7)1.2 (0.3–4.5)
Panic disorderNo observation
Agoraphobia4.9 (2.7–11.4)4.7 (1.8–12.5)**3.2 (1.2–8.7)*
Obsessive compulsive disorder3.1 (1.5–6.3)2.9 (1.3–6.4)**1.8 (0.8–4.4)
Nonspecific psychiatric morbidity12.2 (6.4–21.9)1.5 (0.8–3.1)1.3 (0.6–2.6)
Suicide and self-harm
 Suicidal thoughts lifetime26.5 (18.0–37.1)2.3 (1.4–3.8)**2.0 (1.2–3.2)**
 Suicidal thoughts in last year12.2 (6.9–20.8)3.3 (1.7–6.2)***2.5 (1.3–4.8)**
 Suicide attempts lifetime12.5 (6.7–21.2)3.0 (1.5–5.9)**2.3 (1.1–4.6)*
 Suicide attempts last year4.1 (1.5–10.5)7.3 (2.4–21.9)***4.6 (1.6–13.6)*
 Deliberate self-harm12.7 (7.2–21.5)2.9 (1.5–5.5)**2.3 (1.2–4.5)*
Neurodevelopmental/other conditions
 Autism spectrum disorder8.1 (2.2–25.9)9.3 (2.0–42.4)**7.4 (1.8–30.6)*
 Psychotic disorder1.1 (0.1–7.2)2.7 (0.4–20.8)1.7 (0.2–15.8)
 Eating disorder (SCREEN)5.0 (1.9–12.6)3.4 (1.2–9.7)*2.9 (1.1–7.7)*
 Posttraumatic stress disorder (SCREEN) 4.9 (2.0–11.4)1.8 (0.7–4.4)1.2 (0.5–2.9)
 Attention deficit disorder (SCREEN) 15.4 (9.2–24.7)2.0 (1.1–3.7)*1.6 (0.9–3.0)
Table 3.   Logistic regression analysis of associations of anxiety, depression, and suicidality in people with asthma, diabetes, and migraine or chronic headaches as compared to the general population of England (without these respective conditions).
Mental disorderAsthma (n = 886) Adjusted OR (95% CI)Diabetes (n = 406) Adjusted OR (95% CI)Migraine (n = 936) Adjusted OR (95% CI)
  1. Adjusted OR, model adjusted for age, gender, marital status, highest educational qualification, employed or economically inactive, and number of comorbid physical illnesses.

  2. OR, odds ratio; 95% CI, confidence interval.

  3. *p < 0.05; **p < 0.01; ***p < 0.001.

Any depressive or anxiety disorder1.7 (1.3–1.9)***1.0 (0.7–1.3)2.3 (1.9–2.8)***
Depression and an anxiety disorder1.6 (1.1–2.3)*0.8 (0.5–1.4)3.1 (2.1–4.7)***
Depression1.4 (1.0–2.0)0.9 (0.5–1.5)2.5 (1.7–3.8)***
Generalized anxiety disorder1.2 (0.9–1.6)1.0 (0.6–1.7)2.1 (1.6–2.9)***
Social phobia0.8 (0.4–1.6)1.9 (1.0–3.5)
Specific phobias1.5 (0.8–2.9)0.1 (0.0–1.0)2.7 (1.5–5.0)**
Panic disorder1.3 (0.7–2.4)2.2 (1.0–5.0)1.9 (1.0–3.4)*
Agoraphobia1.1 (0.6–2.2)0.4 (0.1–1.3)2.3 (1.3–3.9)**
Obsessive compulsive disorder1.7 (0.9–3.2)1.0 (0.4–2.6)1.8 (1.0–3.2)
Nonspecific psychiatric morbidity1.7 (1.3–2.1)***0.9 (0.6–1.3)1.8 (1.4–2.2)***
Suicide and self-harm
Suicidal thoughts lifetime1.5 (1.2–1.8)***1.0 (0.7–1.4)2.2 (1.8–2.6)***
Suicidal thoughts in last year1.5 (1.1–2.1)*1.1 (0.7–1.8)2.5 (1.8–3.5)***
Suicide attempts lifetime1.5 (1.1–2.1)**1.1 (0.6–1.9)1.9 (1.3–2.6)**
Suicide attempts last year3.5 (1.6–7.7)**1.3 (0.3–6.8)4.3 (1.8–10.1)**
Deliberate self-harm1.5 (1.1–2.0)*0.5 (0.2–1.1)1.5 (1.0–2.2)*
Table 4.   Evidence of mediation of the relationship between epilepsy and suicidality by anxiety and depressive disorders (model 2) and symptoms (model 3).
Measure of suicidalityModel 1 OR (95% CI)Model 2 OR (95% CI)Model 3 OR (95% CI)
  1. Model 1: adjusted estimates after adjustment for confounders (age, gender, marital status, highest educational qualification, employed or economically inactive, tenure of accommodation, and number of comorbid physical illnesses).

  2. Model 2: Model 1 adjusted for presence of any anxiety or depressive disorder to assess mediation by anxiety or depressive disorders.

  3. Model 3: Model 1 adjusted for total CIS-R score to assess mediation by anxiety or depressive symptoms (this approach would ensure subthreshold psychopathology is also captured).

  4. OR, odds ratio; 95% CI, 95% confidence interval.

  5. *p < 0.05; **p < 0.01.

Suicidal thoughts lifetime2.3 (1.1–4.6)**1.6 (0.9–2.8)1.5 (0.9–2.8)
Suicidal thoughts in last year2.5 (1.3–4.9)**1.9 (0.9–3.8)1.7 (0.8–3.5)
Suicide attempts lifetime2.2 (1.1–4.6)*1.8 (0.8–4.0)1.7 (0.8–3.8)
Suicide attempts last year4.3 (1.4–13.5)*3.2 (1.0–10.4)2.7 (0.8–9.1)
Deliberate self-harm2.2 (1.1–4.2)*1.8 (0.9–3.5)1.6 (0.8–3.3)

A strong association between epilepsy and autism spectrum disorders was observed even after accounting for sociodemographic and health covariates (OR 6.2) and increased further after adjusting for verbal IQ (OR 7.4, 95% CI 1.5–35.5). Epilepsy was also associated with a 70% increase in odds of a psychotic disorder, although CIs were wide (adjusted OR 1.7, 95% CI 0.2–14.2). The associations with screen-positive adult ADHD were attenuated when covariates were introduced in the models. Finally, epilepsy was associated with an almost threefold increase in odds of screen-positive eating disorder. None of these conditions appeared to have significant associations in people with asthma, diabetes, or migraine.

Discussion

  1. Top of page
  2. Summary
  3. Method
  4. Results
  5. Discussion
  6. Implications and Future Research
  7. Acknowledgments
  8. Disclosure
  9. Author Contributions
  10. References

In this comprehensive, nationally representative population-based study, we examined the burden and relationships of a range of comorbid psychiatric conditions in epilepsy, including several that have never been investigated previously in adult populations. Almost one third of the people with epilepsy met ICD-10 diagnostic criteria for an anxiety or depressive disorder in this study, compared with about one in six people in the general population without epilepsy. After adjustment for confounders, people with epilepsy had significantly elevated odds of having social phobias, agoraphobia, generalized anxiety disorder, and depression and all measures of suicidality. These associations were consistently stronger than similar relationships in people with asthma or diabetes, and comparable to people reporting migraines or chronic headaches. People with epilepsy had a more than sevenfold increase in odds of having an autism spectrum disorder and an almost threefold increase in the odds of being screen positive for eating disorder, compared with people without epilepsy. These associations were not observed in people with asthma, diabetes, or migraine.

To our knowledge, this nationally representative population-based study represents the most detailed examination of the psychiatric and neurodevelopmental comorbidities of epilepsy ascertained by face-to-face interviews using standardized validated measures of psychopathology. In addition, our ability to control for appropriate confounders, and compare our results with similar associations in people with asthma, diabetes, and migraines offers a significant advantage in comparison to previous literature.

Some limitations of our study should be noted. The ascertainment of epilepsy in this household survey was by self-report; therefore, the possibility of some reporting bias cannot be excluded. In addition, we had no information on the nature of the diagnostic process, seizure type, frequency, or treatment of epilepsy. However, these limitations are common to almost all previous population-based studies on this topic (Ettinger et al., 2004; Gaitatzis et al., 2004b; Tellez-Zenteno et al., 2007; Ottman et al., 2011), and yielded an internationally comparable prevalence of epilepsy. Furthermore, due to the recognized challenges in the diagnosis of epilepsy it is possible that a proportion of our sample classified as epilepsy actually have an alternative diagnosis such as non–epileptic attack disorder; however, this potential misclassification is common to all population-identified epilepsy samples. Population-based studies with more robust measures of epilepsy and detailed information on the various types of epilepsy are required to overcome this limitation (Berg et al., 2010; Thurman et al., 2011). Similarly, the comparison conditions included in this study were ascertained by self-report, although an effort was made to ascertain if these had been diagnosed by a doctor. Despite the large sample, as expected, the number of people with epilepsy was relatively low, which limited our statistical power, particularly evident through wide CIs of associations with other rare conditions. Response rates in national surveys are falling throughout the world (Groves, 2002), and the 7,461 individuals interviewed in the APMS constituted 57% of the eligible sample, comparable to other population-based studies on this topic (Ettinger et al., 2004). We used detailed weighting procedures to minimize the likelihood of nonresponse bias (McManus et al., 2009). Finally, the cross-sectional design, although important to estimate the burden of comorbidities, does not allow for directional inferences. For instance, the relationship between epilepsy and depression may be bidirectional (Kanner, 2008), and longitudinal studies are required to clarify causal pathways of our observations.

The prevalence estimates of depression and anxiety disorders in people with epilepsy in our study are consistent with the relatively sparse population-based literature on this topic (Ettinger et al., 2004; Gaitatzis et al., 2004b; Tellez-Zenteno et al., 2007; Garcia-Morales et al., 2008). In addition, our study clarifies several issues not adequately addressed previously. First, it is notable, that anxiety disorders, particularly phobias and GAD, are common comorbidities in epilepsy. This is important, since these may be as disabling as depression (Rai et al., 2010), and are potentially treatable. Second, the associations of epilepsy with depression, anxiety disorders, and suicidality were consistently stronger than similar estimates in asthma and diabetes, indicating that epilepsy has specific and stronger relationships with these comorbidities than other chronic physical conditions. Similarly, although people reporting migraines or chronic headaches may actually represent the presence of neurotic psychopathology rather than a primary neurologic condition, (Antonaci et al., 2011) comparing the rates of psychopathology in people with epilepsy with this group only reinforce our results that depression, anxiety disorders, and suicidality are significant issues in people with epilepsy. Our observation that apart from depression, anxiety, and suicidality, no other conditions we studied had significant relationships with asthma, diabetes, or migraine suggests that epilepsy has unique associations with psychopathology that are not simply explained due to it being a chronic (neurologic) condition.

The associations between suicidality and epilepsy in our study were largely mediated by comorbid depression or anxiety disorder. This contrasts with findings from a Danish study (Christensen et al., 2007), where secondary-care diagnosis (ascertained mostly from inpatient records) of psychiatric conditions did not fully explain the risk of suicide in people with epilepsy. Two important issues may explain this discrepancy. First, depressive and anxiety disorders are mostly managed in primary care, and second, psychopathology is often undiagnosed in people with epilepsy (Hermann et al., 2000). Hence our comprehensive face-to-face assessment of psychopathology was an inherent advantage as compared to the use of a registered secondary-care diagnosis.

Another notable contribution of our study is in relation to the relatively high prevalence of nonspecific psychiatric morbidity in people with epilepsy, but without a significant increase in odds ratios as compared to the general population (Table 2). These presentations, where significant anxiety or depressive symptoms do not meet diagnostic criteria for specific disorders, are reminiscent of the “atypical” presentations of depression and anxiety that epileptologists have previously termed as “dysthymic like disorders” or “dysphoric disorder of epilepsy” (Kanner, 2003; Blumer et al., 2004). These presentations were among the most common psychiatric morbidities in the general population, and were no more common in people with epilepsy than those without.

Our data also allowed us to estimate associations between epilepsy and several conditions that are relatively uncommon in the general population. Noteworthy among these was the strong association between adult autism spectrum disorders and epilepsy. This relationship is widely noted in children, although pediatric population-based studies on this topic are also limited (Tuchman & Rapin, 2002). Furthermore, a recent study suggests seizures may, in fact, arise in late adolescence or adulthood in people with ASD (Bolton et al., 2011), reinforcing our confidence in our results. In addition, in keeping with some previous literature, we found a 70% increase in odds of a psychotic disorder in people with epilepsy, albeit with wide CIs. Clinical studies have reported a higher prevalence (2–9%) of psychotic disorders (Garcia-Morales et al., 2008) in people with epilepsy (particularly temporal lobe epilepsy) than our modest 1% prevalence. However, it has been noted that previous reports may have overestimated this association by studying highly selected clinical populations (Swinkels et al., 2005).

Although questions covering adult ADHD, eating disorders, and PTSD were administered using screening instruments, our study represents the first population-based estimates of their possible relationship with epilepsy in adults. It has previously been suggested that the prevalence of ADHD in adults with epilepsy may be underestimated (Garcia-Morales et al., 2008), the rationale being that 25–30% of the pediatric population with epilepsy have comorbid ADHD, and that 50–75% of these children may continue to have ADHD in adulthood (Garcia-Morales et al., 2008). We found partial support for this assertion. More than 15% of adults with epilepsy in our sample with epilepsy screened positive for adult ADHD, representing an over twofold increase in odds when compared with the general population without epilepsy. This association, however, was largely explained after adjustment for confounders. We also found a significant association between epilepsy and screen-positive eating disorders, which has not been previously investigated. We did not find evidence of a relationship between screen-positive PTSD and epilepsy in contrast with findings from one small study (Rosenberg et al., 2000) of 35 patients with intractable seizures. Further exploration of the relationships between epilepsy and these conditions, using diagnostic instruments are obvious questions for future research.

Possible mechanisms

Epilepsy has been a stigmatized condition throughout history, and negative societal (mis)perceptions, stereotyping, and discrimination against people with epilepsy continues to be common (Scambler & Hopkins, 1980; Baxendale & O’Toole, 2007; de Boer et al., 2008; de Boer, 2010; The Lancet Neurology, 2010), and is known to propagate social isolation and low self-esteem (Jacoby, 1994; Krauss et al., 2000). People with epilepsy often live in uncertainty and in fear of having further seizures (Fisher et al., 2000). These issues could explain our results, and conditions associated with worry (GAD and other anxiety disorders), avoidance (phobias, particularly social phobia), stigma (depression, phobias), and low self-esteem (depression and suicidality, eating disorders) all had strong relationships with epilepsy.

The understanding of a common neurobiology between epilepsy and psychiatric comorbidities is also emerging (Tecott et al., 1995; Cowan & Kandel, 2001; Martin, 2002; Tuchman & Rapin, 2002; Jobe, 2003; Kanner, 2005), although most evidence is based on animal models or observed indirectly (e.g., through efficacy of anticonvulsants on mood disorders) (Grunze, 2008). Some overlap in the psychopathology observed in epilepsy and migraine cannot rule out the presence of common pathways of comorbidity in neurologic conditions. However, associations of epilepsy with conditions such as autism spectrum disorders point to comorbidities specific to epilepsy that may not be shared by other neurologic conditions. Furthermore, advances in the neurosciences hold potential for elucidation of pathways between seizures, other neurologic conditions, and psychopathology (Cowan & Kandel, 2001; Martin, 2002).

Implications and Future Research

  1. Top of page
  2. Summary
  3. Method
  4. Results
  5. Discussion
  6. Implications and Future Research
  7. Acknowledgments
  8. Disclosure
  9. Author Contributions
  10. References

The causal pathways of our findings need further elucidation through well-designed cross-disciplinary longitudinal studies using detailed measures of epilepsy (Thurman et al., 2011). The impoverishing “wall” (Cowan & Kandel, 2001; Baker et al., 2002; Martin, 2002) between neurology and psychiatry is notable through the documentation of the “sore lack” of epidemiologic data on epilepsy (The Lancet Neurology, 2010), when detailed psychiatric morbidity surveys such as the APMS are being conducted. Future psychiatric morbidity surveys hold potential as vehicles for detailed population-based studies of epilepsy. In clinical practice, greater awareness of the range of psychiatric presentations in people with epilepsy, and their effective identification and management is warranted.

Acknowledgments

  1. Top of page
  2. Summary
  3. Method
  4. Results
  5. Discussion
  6. Implications and Future Research
  7. Acknowledgments
  8. Disclosure
  9. Author Contributions
  10. References

The APMS survey was carried out by the National Centre for Social Research in collaboration with the University of Leicester, and was commissioned by The NHS Information Centre for health and social care with funds from the Department of Health. The APMS data are available from the UK Data Archive, University of Essex, Colchester. DR is funded through a clinical lecturer award funded by Severn Deanery, Bristol, U.K. We thank Professor Howard Meltzer for useful comments on the manuscript.

Disclosure

  1. Top of page
  2. Summary
  3. Method
  4. Results
  5. Discussion
  6. Implications and Future Research
  7. Acknowledgments
  8. Disclosure
  9. Author Contributions
  10. References

None of the authors has any conflict of interest to disclose.

We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines

Author Contributions

  1. Top of page
  2. Summary
  3. Method
  4. Results
  5. Discussion
  6. Implications and Future Research
  7. Acknowledgments
  8. Disclosure
  9. Author Contributions
  10. References

(1) All coauthors have been substantively involved in the study and/or the preparation of the manuscript; (2) no undisclosed groups or persons have had a primary role in the study and/or in manuscript preparation (i.e., there are no “ghost-writers”); and (3) all coauthors have seen and approved the submitted version of the paper and accept responsibility for its content. Specific contributions: DR conceived this paper, conducted the analysis, wrote the first and subsequent drafts of the manuscript, and is the guarantor. MK, VJ, SM, GL, and TSB contributed significantly toward the scientific development of this article at all stages, and edited successive versions of the manuscript. SM and TSB made important contributions to the design and conduct of the parent APMS survey. DR had full access to the data and takes responsibility for submission.

References

  1. Top of page
  2. Summary
  3. Method
  4. Results
  5. Discussion
  6. Implications and Future Research
  7. Acknowledgments
  8. Disclosure
  9. Author Contributions
  10. References