- Top of page
Purpose: Whether a specific lesion such as mesial temporal sclerosis (MTS) increases the risk for a mood disorder in epilepsy remains subject to debate. Despite evidence of limbic system involvement in the genesis of emotional symptoms, recent studies fail to support an association between depression and MTS. We aimed to clarify this controversial issue by overcoming prior methodologic limitations, hypothesizing that rates of major depressive disorder (MDD) would be higher in patients with MTS.
Methods: Three hundred eight patients with focal epilepsy (International League Against Epilepsy [ILAE] criteria), were classified into three groups on the basis of neuroimaging findings: MTS, a lesion different from MTS, or absence of lesion. Patients were assessed using the Structured Interview for DSM-IV axis I psychiatric disorders (SCID-I), by a psychiatrist blinded to epilepsy subtype. The Spanish version of the Hospital Anxiety and Depression Scale (HADS) was also administered. A complete logistic regression analysis was performed to investigate the association between MTS and MDD.
Key Findings: MTS increased the likelihood of a lifetime MDD by nearly 2.5. No other current or “postseizure onset” lifetime Axis I DSM-IV psychiatric disorder was associated with MTS. Female gender, primary education, comorbid anxiety disorders, and antidepressant treatment were also associated with an increased risk of MDD. Marriage was found to be a protective factor for MDD.
Significance: Our results support a specific association between MTS and lifetime “postseizure onset,” MDD. The lack of association with current depression is in line with the hypothesis that the link between MTS and depression is more of a chronic than a state-dependent condition.
High rates of psychopathologic disturbances have been reported in patients with epilepsy (Jones et al., 2010), in comparison not only with the general population, but also with patients who have other neurological disorders and chronic nonneurological diseases (Gaitatzis et al., 2004).
Furthermore, between 20% and 30% of patients with epilepsy have persistent seizures despite antiepileptic drug (AED) therapy, and 25–50% of these refractory patients have mesial temporal lobe epilepsy (Engel, 2001). Despite methodologic differences between studies, it is now well established that among patients with epilepsy, those with intractable epilepsy have the highest rates of psychiatric comorbidity (Engel, 2001). Rates of DSM-IV axis I disorders in refractory epilepsy patients vary between 44% and 70% (Gaitatzis et al., 2004), of which mood, followed by anxiety disorders, is the most common diagnosis (Swinkels et al., 2005). With regard to mood disorders: lifetime comorbidity with major depressive disorder (MDD) across studies ranges up to 30–50%, being the most frequent psychiatric disorder in epilepsy (Gaitatzis et al., 2004; Jones et al., 2010). This strong association suggests that the identification of specific, seizure-related risk factors could potentially contribute toward the early detection of patients with epilepsy at high risk of depression.
Nevertheless, studies aiming to determine whether patients with temporal lobe epilepsy (TLE) are at higher risk of depression compared to those with extratemporal epilepsy (extra-TLE), have reported inconsistent results (Swinkels et al., 2006) A recent study by our group (Sanchez-Gistau et al., 2010) aimed to clarify this issue, and concluded that MDD was more prevalent in patients with TLE. However, whether specific brain lesions, regardless of their temporal or extratemporal origin, confer an increased risk of depressive disorders in patients with epilepsy is a topic that remains to be resolved.
It is widely accepted that the limbic system is directly implicated in the regulation of emotions, and that it is tightly connected to the prefrontal cortex. Given their roles in cognition, emotional processing, and hypothalamic-pituitary–adrenal (HPA) axis regulation, the hippocampus, amygdala and prefrontal regions have been studied extensively in primary MDD (Lorenzetti et al., 2009). Specifically, volumetric reductions of the hippocampus have been consistently reported in MDD (Bremner et al., 2000; Malykhin et al., 2010). Of note, sclerosis of mesial temporal lobe structures (mesial temporal sclerosis, MTS), such as hippocampal sclerosis, are the most common lesions in patients with refractory epilepsy who undergo surgical treatment (Thom et al., 2010).
Given the reported relationship between structural abnormalities and depressive symptoms, it is reasonable to assume that lesions in regions involving mood regulation such as MTS may increase the risk of depression. Evidence from neuroimaging studies of stroke (Mayberg et al., 1988; Vataja et al., 2004; Terroni et al., 2011) and traumatic brain injury (Jorge et al., 2004, 2007) grant further support to this hypothesis, and underscore the relevance of lesions in specific neural circuits to mood regulation and postinjury onset of depression.
Despite the shared neural structures underlying the pathogenesis of both MTS epilepsy and mood disorders, studies investigating whether MTS epilepsy increases the risk of depression relative to other subtypes of epilepsy have yielded mixed results. In the first study (Quiske et al., 2000), sixty patients with TLE were subdivided according to mesial versus neocortical temporal sclerosis, and depressive symptoms were assessed using the self-reported Beck Depression Inventory (BDI). The authors found that depression occurred more frequently in the presence of MTS and was not influenced by lateralization. Following this report, the majority of studies have provided divergent findings, and a later study by the same group (Helmstaedter et al., 2004) was unable to replicate their previous results. Furthermore, increased rates of depression in patients without MTS have been reported by other authors (Heuser et al., 2009). Of note, a recent study by de Oliveira et al. (2010) concluded that patients with left MTS had a higher frequency of mood and anxiety disorders in comparison with those with right or bilateral MTS or absence thereof.
Taking the existing body of research into account, it is clear that the relationship between MTS and depression is still subject to debate.
One study to date (Adams et al., 2008) has attempted to shift the previous focus of research, by including patients with extratemporal epilepsy in their sample. The authors were interested in studying whether a diagnosis of depression was best predicted by the presence of a defined lesion (MTS, a lesion different from MTS or absence of lesion), independent of its origin. They concluded that depression was associated with nonlesional epilepsy. To our knowledge, no other study including both patients with temporal and extratemporal epilepsy has been performed in order to clarify this issue. Taking into account that our group has previously documented that TLE patients are at greater risk of MDD and that brain lesions have been reported to increase the risk of depression in both stroke and brain injury studies, we hypothesize that patients with MTS will have a higher prevalence of MDD, as compared to those with nonlesional and non-MTS lesional epilepsy. We will also investigate whether MTS patients are at greater risk for anxiety. For this purpose, we have evaluated a large sample of patients with temporal and extratemporal focal epilepsy, using a standardized instrument for the assessment of Axis I DSM-IV psychiatric disorders.
- Top of page
Sociodemographic data and seizure-related variables did not differ significantly between groups (Table 1).
Table 1. Sociodemographic information and seizure-related features
|Variable||Total (N = 308) N (%)||Lesional MTS (N = 115) N (%)||Lesional non- MTS (N = 106) N (%)||Nonlesional (N = 87) N (%)||p-Value|
|Average age (years)a|| 35.07 (12.08)|| 35.36 (11.95)|| 35.44 (11.96)|| 35.24 (12.30)||0.54|
|Gender|| || || || || |
| Female||173 (56.2)|| 68 (59.2)||59 (55.6)||46 (40.0)||0.19|
|Civil status|| || || || || |
| Married||110 (35.7)|| 40 (34.8)||42 (39.6)||28 (32.2)||0.56|
|Education|| || || || || |
| Primary education||189 (61.3)|| 67 (58.6)||64 (60.9)||57 (65.3)||0.82|
|Occupation|| || || || || |
| Unemployed||111 (35.7)|| 36 (31.3)||38 (34.9)||37 (42.5)||0.23|
|Localization of lesionb|| || || || || |
| Temporal||163 (73.7)||115 (100)||48 (45.3)|| || |
| Extratemporal|| 58 (26.2)|| ||58 (54.7)|| |
|Lateralization of lesionb|| || || || || |
| Left||103 (46.6)|| 51 (44.3)||52 (49.1)|| ||0.69|
| Right|| 92 (41.6)|| 51 (44.3)||41 (38.7)|| |
| Bilateral|| 26 (11.8)|| 13 (11.3)||13 (12.3)|| |
|Duration of illness (in years)a|| 21.22 (12.01)|| 23.60 (12.10)|| 22.09 (12.86)|| 19.89 (12.09)|
|Seizure frequency|| || || || || |
| >5 per month||186 (60.3)||68 (58.2)||66 (61.6)||52 (58.2)||0.11|
Lifetime and last-year prevalence of anxiety and depression of the total sample is shown in Table 2. Lifetime prevalence of any Axis I psychiatric disorder was 54.8%. Thirty-six patients (11.7%) fulfilled criteria for more than one Axis I diagnosis. The highest comorbidity was found between mood and anxiety disorders. Among patients with affective disorders, lifetime prevalence of an anxiety disorder was 24.4%.
Table 2. Comparison of lifetime and last year prevalence of psychopathology among groups
|Variable||Total sample (N = 308) N (%)||Lesional MTS (N = 101) N (%)||Lesional non-MTS (N = 123) N (%)||Nonlesional (N = 184) N (%)||p-Value|
|Lifetime||Last year||Lifetime||Last year||Lifetime||Last year||Lifetime||Last year||Lifetime||Last year|
|Any AXIS I disorder||170 (54.8)||122 (39.4)||70 (60.9)||44 (38.3)||53 (50.0)||41 (38.7)||47 (54.0)||37 (42.5)||0.25||0.80|
|Mood disorders||107 (34.5)|| 66 (21.3)||45 (39.1)||23 (20.0)||34 (32.1)||24 (22.6)||28 (32.2)||19 (21.8)||0.45||0.88|
| Major depression|| 67 (21.6)|| 35 (11.3)|| 30 (26.1)A||16 (13.9)||19 (17.9)B||13 (12.3)|| 10 (11.5)C||6 (6.9)|| 0.03*||0.27|
| HAD-Aa|| 6.31 (4.3)|| 11.62 (2.5)|| 6.32 (4.4)|| 7.65 (5.0)|| 6.85 (4.3)|| 6.90 (4.2)|| 5.66 (4.3)|| 5.87 (4.7)||0.55||0.44|
| HAD-D|| 8.27 (4.9)|| 6.80 (4.7)|| 8.48 (4.5)|| 11.71 (2.9)|| 8.62 (5.0)|| 12.30 (2.3)|| 9.07 (4.1)|| 10.91 (1.8)||0.35||0.16|
| HADS total|| 14.90 (7.3)|| 18.42 (6.0)|| 14.03 (6.6)|| 19.39 (6.7)|| 15.47 (6.9)|| 19.20 (5.6)|| 14.72 (6.6)|| 16.78 (5.6)||0.60||0.27|
|Anxiety disorders|| 67 (21.6)|| 60 (19.4)||25 (21.7)||22 (19.1)||19 (17.9)||19 (17.9)||23 (26.4)||19 (21.8)||0.86||0.86|
| GAD|| 39 (12.6)||30 (9.7)||15 (13.0)||12 (10.4)||12 (11.3)||9 (8.5)||12 (13.8)|| 9 (10.3)||0.63||0.87|
| HAD-Aa|| 10.31 (4.4)|| 12.67 (3.0)|| 10.10 (4.2)|| 12.57 (2.6)|| 11.29 (4.5)|| 12.87 (3.2)|| 9.06 (4.4)|| 11.82 (2.8)||0.24||0.91|
| HAD-D|| 5.72 (4.5)|| 5.92 (4.7)|| 5.52 (5.0)|| 6.71 (5.7)|| 5.82 (3.8)|| 5.30 (4.2)|| 5.89 (4.3)|| 6.21 (3.4)||0.09||0.13|
| HADS total|| 15.54 (6.5)|| 18.40 (6.6)|| 16.62 (7.8)|| 19.29 (7.8)|| 16.11 (1.44)|| 17.22 (6.4)|| 14.94 (6.1)|| 17.36 (5.4)||0.66||0.34|
|Prescription of:|| || || || || |
| Psychotropic drugs||60 (19.4)||25 (21.7)||20 (18.9)||15 (17.2)||0.71|
| Antidepressants||40 (12. 9)||15 (13.0)||15 (14.2)||10 (11.5)||0.86|
Only the HADS scores of patients with mood or anxiety disorders are displayed in Table 2. Patients who had no history of anxiety or depression (n = 152) obtained significantly lower ratings (p < 0.00) on the total HADS (7.1 ± 3.4) as well as on the anxiety (4.62 ± 2.8) and depression (2.86 ± 2.09) subscales, compared to patients with a lifetime history of anxiety or depression (n = 156) (results not shown in the table).
When the three groups were examined using univariate analysis, only prevalence of lifetime MDD was found to be significant (χ2 = 6.96; p = 0.03). Post hoc analysis revealed that at univariate level the “lesional MTS” group showed higher rates of lifetime MDD than the “nonlesional” group (χ2 = 6.64; p = 0.007) and the “lesional non-MTS” group (χ2 = 1.54; p = 0.05), respectively. On the other hand, the “nonlesional” and “lesional non-MTS” groups did not differ in MDD rates. Furthermore, prevalence of neither lifetime nor any other current Axis I psychiatric disorder was found to be significantly different when the three groups were compared. Similarly, scores of HADS total or depression and anxiety subscales did not differ between groups.
After controlling for different possible confounding factors in the multivariate analysis, MTS remained significantly associated with a higher lifetime prevalence of MDD (OR 2.46; CI 95% 1.28–4.73 p = 0.006). No interaction between independent variables was identified to be significant. In the model, female gender, primary education, comorbidity with anxiety disorders, and antidepressant treatment were also independently associated with lifetime prevalence of MDD. In addition, marriage decreased the risk of MDD (Table 3).
Table 3. Logistic regression analysis for variables associated with lifetime major depression
| ||OR (95% CI)||p-Value|
|MTS versus non-MTS (other lesions and absence of identifiable lesion)||2.46 (1.28–4.73)||0.006|
|Female gender||2.74 (1.36–5.52)||0.005|
|Not married (single/widow/divorced)||2.18 (1.04–4.56)||0.03|
|Primary education||2.30 (1.11–4.75)||0.02|
|Antidepressant treatment||7.58 (3.34–17.24)||0.000|
|Lifetime anxiety disorder||10.38 (2.84–37.97)||0.000|
The sample was also subdivided into patients with lesional (“MTS lesional” and “non-MTS lesional”) and nonlesional epilepsy, revealing no significant differences between groups with regard to lifetime or current prevalence of mental health disorders.
- Top of page
This study demonstrates association between MTS, a specific limbic system lesion, and “postseizure onset” lifetime MDD, in a large sample of patients with refractory epilepsy, thereby confirming our primary hypothesis.
Our results contradict the previous report by Adams et al. (2008), in which patients with nonlesional epilepsy were more susceptible to depression than patients with lesional epilepsy. Sample size, recruitment, and classification of epilepsy are similar in both studies, although discrepancies with regard to the diagnostic formulations and the statistical analysis could have influenced the final findings. It is worth noting that our study has a number of methodologic strengths: first, psychiatric assessments were performed using a standardized interview, blinded to epilepsy subtype, whereas in the previous investigation by Adams et al., the psychiatrist was not blinded and no standardized instrument was used. Second, we set out to perform a more comprehensive logistic regression analysis controlling for potential confounders and their interactions, in order to make our results as robust as possible.
Our findings also contradict those by Heuser et al. (2009). In their study, depression was more prevalent in the non-MTS group; this could be due to differences in inclusion criteria and in diagnostic procedures. The 218 patients with TLE who were included were recruited from seven different hospitals, and self-reported scales were used for the assessment of depression. In our study, when depression and anxiety were measured using the self-administered HADS scale, although scores were higher in patients with depression or anxiety than in those without, the overall ratings did not differ across the three groups. However, when a validated instrument based on a semistructured DSM-IV interview was employed, lifetime MDD was found to be associated with MTS. It is widely accepted that self-reported questionnaires have a tendency to reflect emotional states, but that they are less accurate at predicting diagnoses of psychiatric disorders.
In concert with the HADS scores, last-year prevalence of mood and anxiety disorders was similar across groups. This concurs with the hypothesis that the association of MTS with depression is more of a chronic than a state-dependent condition. This is supported by the notion that structural changes associated with MTS are generally stable over time, and have been observed during the early stages of development in this form of epilepsy. Furthermore, volumetric reductions of the hippocampus have been associated with past, as opposed to recent onset, MDD (MacQueen et al., 2003).
On the other hand, our findings are in keeping with those of the very first study by Quiske et al. (2000), in which the relationship between depression and MTS was also reported to be independent of focus lateralization. Two studies (Adams et al., 2008; Salgado et al., 2010) also failed to find an association using a psychiatric interview. However, the study by Salgado suggested involvement of the left hippocampus in the severity of symptoms, when measured using the BDI. de Oliveira et al. (2010) recently reported that patients with left MTS were at greater risk of depression. However, only 73 TLE patients, subdivided into four groups (right mesial, left mesial, bilateral mesial and temporal without mesial sclerosis) were included, thus increasing the likelihood of type II error.
It is clear that the role of epileptic loci in the pathogenesis of emotional symptoms remains controversial. Recent neuroimaging studies investigating the influence of depression on volumetric changes in limbic areas have been inconsistent. Although some studies (Helmstaedter et al., 2004) have found no correlation between depression and lateralization, others (Baxendale et al., 2005; Shamim et al., 2009) have described a significant left hippocampal volume reduction in depressed MTS patients compared with nondepressed MTS patients.
We have failed to find a specific target region for anxiety disorders. Once again, this lies in contrast with the results of the Brazilian study (de Oliveira et al., 2010), where an increase of anxiety in patients with left MTS was reported. The methodologic differences between both studies may account for this discrepancy. Moreover, there is an important overlap between depressive and anxiety symptoms in clinical practice, which is why we decided to include anxiety disorders as an independent variable in the logistic regression model. We found that a lifetime anxiety disorder significantly increases the risk of having an MDD. To our knowledge, the remaining studies did not include anxiety as a confounding factor. We consider this to be an important shortcoming that should be taken into account when interpreting their results.
The association encountered between MDD and other variables such as female gender (Cavanna et al., 2009) and primary education (Muntaner et al., 2004) have been widely described in the literature. Moreover, marriage has been reported as a protective factor in epidemiologic studies in MDD (Trivedi et al., 2005; Holzel et al., 2011). In contrast, other seizure-related variables including duration of epileptic illness and seizure frequency were not found to be associated with an increased risk of MDD. Although one may expect that duration of epilepsy could influence the risk for depression (Briellmann et al., 2007), prior studies have not supported this association (Salgado et al., 2010). The challenges involved in the collection of reliable data related to the onset of epilepsy could have influenced the final findings in this regard.
Forty patients (12.9%) were taking antidepressants at the time of assessment; no differences were found in rates of treatment between groups. Low rates of psychopharmacologic treatments have been consistently reported in epileptic samples (Kanner et al., 2010). “In spite of the low proportion of patients treated in our sample, antidepressant treatment was a strong predictor of lifetime diagnoses of MDD. On one hand we can speculate that the association of MDD and antidepressants could be due to the fact that depressed patients are more likely to be treated with antidepressants, but on the other hand, antidepressant treatment could have potentially underestimated the rates of current mood and anxiety disorders. Unfortunately, as treatment was naturalistic, the specific role of antidepressants in this study cannot be ascertained.”
The present study has several limitations: First, the cross-sectional and retrospective assessment of past Axis I disorders. There is a need of further longitudinal studies with periodic follow-up psychiatric assessments from the time of epilepsy onset, so as to increase the reliability of these associations. Second, the instrument used for assessment (SCID-I) has not been validated in epilepsy, and does not assess personality disorders or their influence on psychopathology. Finally, the generalizability of the findings was limited by the fact that the patients were drug-resistant.
On the other hand, the strengths of this study lie in its large sample size and the robust methodology and statistical analysis employed. The classification of epilepsy strictly adhered to the International League Against Epilepsy (ILAE) recommendations. Finally, a broad psychiatric assessment using both self-reported scales and diagnostic standardized interview was performed by a psychiatrist blinded to epilepsy subtype.