The treatment of depressive disorders in epilepsy: What all neurologists should know


  • Andres M. Kanner

    Corresponding author
    • Comprehensive Epilepsy Program and Department of Neurology, University of Miami, Miller School of Medicine, Miami, Florida, U.S.A
    Search for more papers by this author

Address correspondence to Andres M. Kanner, Director, Epilepsy Center, University of Miami, Miller School of Medicine, 1120 NW 14th Street, Room 1324, Miami, FL 33136, U.S.A. E-mail:


One of every three patients with epilepsy (PWE) will experience a depressive disorder in the course of their life, often associated with anxiety symptoms or a full blown anxiety disorder. Clearly, the high prevalence of these psychiatric comorbidities calls for their early identification and management. This article provides practical strategies in the management of depressive episodes in PWE. Contrary to long-held beliefs, the use of antidepressant drugs are safe in PWE when used at therapeutic doses. Antidepressant drugs of the selective serotonin reuptake inhibitor (SSRI) or serotonin-norepinephrine reuptake inhibitor (SNRI) families are the first line of therapy in depressive disorders, and failure to achieve complete symptom remission after a trial of an SSRI or SNRI at optimal doses should be followed by a second trial with a drug from the other antidepressant family. In developing countries, antidepressant drugs of these two antidepressant families are not always available, and tricyclic antidepressants (TCAs) are the drugs of choice. Although there are no differences in efficacy among the three families of antidepressants, TCAs have a lower tolerability and higher toxicity, with greater mortality risk associated with cardiotoxic effects in overdoses. Cognitive behavior therapy is another treatment modality that has been shown to be effective in the treatment of depressive disorders in patients with and without epilepsy. Its use should be considered together with pharmacotherapy or by itself.

Depressive disorders are the most common psychiatric comorbidity in patients with epilepsy (PWE) (Tellez-Zenteno et al., 2007). They are part of the family of mood disorders, which are divided into unipolar (e.g., only depressive) and bipolar disorders (e.g., depressive and manic and/or hypomanic episodes) (DSM-IV-TR, 2000). Their clinical presentations in PWE may be indistinguishable from the primary depressive disorders, and as such they are classified into depressive episodes and depressive disorders. When depressive disorders are devoid of any manic or hypomanic symptoms, they are classified as unipolar, a category that has been subdivided into major depressive disorder (MDD), dysthymic disorder (DD), and depressive disorder not otherwise specified (DNOS) in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV-TR, 2000). A diagnosis of MDD is established following a major depressive episode (MDE), which consists of a cluster of symptoms of at least 2-week duration (see Table 1), and it is the most severe expression of depressive disorders. Dysthymic disorder consists of a cluster of symptoms of milder severity but of at least 2-years’ duration (see Table 1). A cautionary note is in order, however. Several of the listed symptoms of primary depressive episodes can be the expression of toxicity to antiepileptic drugs (AEDs, e.g., fatigue, sleep disturbances, cognitive problems) or of the underlying neurologic disorder associated with the epilepsy.

Table 1. Efficacy of SSRIs and SNRIs in primary depression and anxiety disorders
Antidepressant drugDepressionPanic disorderGeneralized anxietyStarting doseMaximal dose
  1. a


  2. b


Sertralinea++ 25200
Fluoxetinea++ 1080
Citaloprama+  1060

In a significant percentage of patients, MDD is a chronic condition. Indeed, following a first MDE, patients have a 50% chance of subsequent MDEs; after two MDE, the risk of subsequent MDE increases to 70% and to 90% after three episodes. Often, MDEs may occur in the midst of a DD, in which case the patient is considered to have a “double depression.”

In more than 50% of PWE, however, depressive disorders tend to have atypical clinical manifestations with respect to the duration and type of symptoms, which often cluster into a pleomorphic clinical picture of symptoms of depression, irritability, and anxiety and fail to meet any diagnostic classification listed in the DSM-IV-TR (Mendez et al., 1986; Kanner et al., 2000). For example, this was the case of 69% of 100 consecutive patients with a depressive episode sufficiently severe to merit referral for psychiatric treatment (Kanner et al., 2000).

Depressive disorders in epilepsy may be mediated by neurobiologic, iatrogenic, and psychosocial pathogenic mechanisms and often they result from the interplay of all of these mechanisms. Therefore, the treatment of depressive disorders has to be based on the type of episode and its operant potential pathogenic mechanism(s). The purpose of this review article is to provide clinicians with a pragmatic approach to the management of this psychiatric comorbidity in PWE.

What Is Unique about Depressive Episodes in Epilepsy?

In PWE, psychiatric symptoms and/or episodes are typically classified according to their temporal relation with the occurrence of seizures into two major classes: interictal (e.g., occur unrelated to seizure occurrence) and periictal, which in turn can be subclassified into preictal (preceding the onset of the seizure), ictal (e.g., as an expression of the actual seizure), or postictal (following seizures). Postictal episodes can occur immediately following a seizure (immediate postictal episode), although more commonly there is an asymptomatic period ranging from several hours to up to 7 days (delayed postictal period). Clinically, however, interictal and periictal episodes may be identified in the same patient. For example, postictal depressive phenomena may often be the expression of an exacerbation in severity of interictal depressive episodes (Kanner et al., 2004), or preictal episodes may persist into the postictal period (Blanchet & Frommer, 1986; Mula et al., 2008). The need to recognize the type of depressive episode is of the essence as the type of treatment differs for interictal and periictal depressive episodes. Indeed, total seizure remission is required for suppression of periictal events, whereas interictal depressive episodes may require pharmacotherapy with psychotropic drugs and/or psychotherapy.

Occasionally, depressive episodes may “paradoxically” occur following a total seizure remission in patients with a long history of poorly controlled seizures. This phenomenon is known as alternative psychopathology or forced normalization (Trimble, 1998). The actual prevalence of “alternative depression” is not known as it goes often unrecognized and/or unreported and the pathogenic mechanisms operant in this type of depressive episodes has yet to be identified. Furthermore, whether some of the de novo postsurgical depressive episodes are an expression of this phenomenon remains a source of debate.

As stated earlier, depressive episodes can also be the expression of an iatrogenic effect associated with the introduction or discontinuation of certain AEDs or may be a complication of epilepsy surgery. Therefore, depressive episodes may result from the following:

  1. The administration of AEDs with negative psychotropic properties in patients susceptible to psychiatric illness (e.g., history of a prior psychiatric illness and/or family psychiatric illness).
  2. Discontinuation of AEDs with mood-stabilizing properties, or antidepressant and/or anxiolytic properties in patients with a prior mood disorder in remission.
  3. Addition of AEDs with enzyme-inducing AEDs (e.g., phenobarbital, primidone, phenytoin, carbamazepine, high-dose topiramate, and oxcarbazepine) in patients taking antidepressant drugs that had yielded a remission of a depressive episode. The increased clearance of the antidepressant drug results in lower serum concentration and potential loss of efficacy.

What Is the Target of Treatment?

As stated above, among the various expressions of depressive disorders, MDE is the most severe type and often has become the only target of therapy in patients with and without epilepsy by many clinicians. Yet, depressive episodes of milder severity (e.g., DD) need to be recognized, as they also have a negative effect on the patients’ quality of life. Furthermore, in the last decade, investigators of primary mood disorders have made emphasis on the relatively high prevalence and significant clinical impact of subsyndromic forms of depressive episodes (SSDEs) (van Praag et al., 2004). This has been the case for PWE as well. For example, in a study of 188 consecutive outpatients from five tertiary epilepsy centers, 26 met criteria for SSDE (Kanner et al., 2010). The effect on health-related quality of life measures was comparable to that of patients with MDE and anxiety disorders.

Complete symptom remission should be the aim of any treatment of depression. Indeed, recognition of less severe forms of depression and residual symptoms is of the essence as their persistence is associated with an increased risk of recurrence and/or development of MDE.

When Treating Depression, Always Look for Comorbid Symptoms of Anxiety or Anxiety Disorders

In patients with and without epilepsy, anxiety symptoms and full blown anxiety disorders are more often than not present (see also article by Mula, 2013 in this issue). In fact, a category of mixed depression and anxiety disorders has been added to the latest edition of the DSM classification (DSM-V) due to appear in 2013. In patients with primary depressive disorders, the comorbid occurrence of an anxiety and depressive disorder has been estimated to range between 40% and 75% (Kessler et al., 1994). Identification of comorbid symptoms of anxiety in depressed patients is of importance, as they are predictive of a higher risk of treatment-resistant depression, including a higher risk of suicidality (Keller, 2005; Young et al., 2008). For example, patients with panic disorder and depression were found to have a 15-fold increased incidence of suicidal ideation (Fawcett et al., 1990; Goodwin et al., 2001).

In a Canadian population study, a lifetime prevalence of comorbid depression and anxiety disorders was identified in 35% of patients (Tellez-Zenteno et al., 2007). Likewise, in a study of 188 outpatients with epilepsy from five tertiary epilepsy programs, 31 had a depressive episode that met criteria for a MDE; 21 of these patients were also experiencing one or more anxiety disorders (Kanner et al., 2010). In this study, patients with MDE and more than one comorbid anxiety disorder had worse scores in health-related quality of life measures. However, whether the co-occurrence of depressive and anxiety disorders in PWE is associated with a worse course as in patients with primary mood disorders is yet to be demonstrated.

Treatment Strategies

Clearly, the aim of any treatment of depressive disorders in PWE is the achievement of complete symptom remission, including symptoms of depression and anxiety. Yet, despite the relatively high prevalence of depressive disorders in PWE, there has been a handful of controlled studies published to date, one with psychotropic drugs and one with cognitive behavior therapy (CBT). Accordingly, the treatment strategies continue to have to rely on data derived from the management of primary depressive and anxiety disorders (Barry et al., 2008; Kerr et al., 2012). Treatment modalities include pharmacotherapy with psychotropic drugs, occasionally AEDs, as well as psychotherapy or a combination of pharmacotherapy and psychotherapy.

Interictal depressive/anxiety episodes have been the typical targets of therapy, whereas remission of seizures has been considered to be the strategy to achieve full remission of periictal depressive episodes, with ictal depression being the obvious example. Yet, in patients with treatment-resistant epilepsy, postictal depressive episodes can be seen in up to 43% of patients, with postictal suicidal ideation identified in 13% (Kanner et al., 2004). Unfortunately, there have been no published data on the treatment of preictal and postictal depressive episodes, although anecdotal reports have suggested that postictal symptoms of depression do not respond to pharmacologic treatment.

Prevention of depressive episodes

Iatrogenic depressive episodes are not infrequent in PWE. Several studies of psychiatric adverse events of AEDs have indicated a significant association between a past psychiatric history and/or family psychiatric history and the development of psychiatric adverse events (Brent et al., 1987; Trimble et al., 2000; Mula et al., 2003; Mula & Sander, 2007; Mula et al., 2009). Therefore, the following steps can assist in preventing or minimizing the occurrence of iatrogenic depressive episodes:

  1. Avoid if possible the introduction of an AED with negative psychotropic properties in patients at risk for psychiatric disorders (e.g., patients with a family or personal psychiatric history). Clearly, an investigation of a personal and family psychiatric history is of the essence before considering the use of this type of AEDs. These include the γ-aminobutyric acid (GABA)ergic AEDs such as barbiturates, benzodiazepines, tiagabine, vigabatrin, and less frequently gabapentin. The inhibition of secretion of serotonin has been postulated as a potential mechanism mediating these depressive episodes. Other AEDs with the potential to cause depressive episodes include topiramate, zonisamide, and levetiracetam. Depressive episodes may be dose-related with some (e.g., barbiturates, benzodiazepines) but not others (e.g., levetiracetam), whereas a rapid titration has been found to facilitate the development of depressive episodes with topiramate (Mula et al., 2009).
  2. Monitor the discontinuation of an AED with mood-stabilizing properties (e.g., valproic acid, carbamazepine, oxcarbazepine, and lamotrigine) in patients with a past and/or a family psychiatric history, as these AEDs may be keeping an underlying mood disorder in remission. This applies as well to the discontinuation of AEDs with anxiolytic properties (e.g., gabapentin, pregabalin, benzodiazepines). Although these observations do not imply that introductions of AEDs with negative psychotropic drugs or discontinuation of AEDs with mood-stabilizing properties are contraindicated in PWE with a previous or family psychiatric history, patients and family members need to be warned of the risks and advised to monitor the patient for any changes in mood and/or the development of anxiety symptoms.
  3. Adjust the dose of concomitant antidepressant drugs when introducing AEDs with enzyme-inducing properties in patients taking this type of drugs, lest they yield lower serum concentrations of the psychotropic drug.
  4. Between 20% and 30% of patients who undergo epilepsy surgery experience a depressive episode in the first 6 months. A majority of these patients had a presurgical history of a mood disorder. Therefore, presurgical recognition and management of this comorbidity may avert its occurrence and/or mitigate its severity postsurgically.


The pharmacologic treatment of depressive disorders has been based on the monoaminergic theory of mood disorders, which suggests a disturbance in neurotransmitter secretion including serotonin, norepinephrine, and dopamine (Charney et al., 1998; Stahl, 2000). Accordingly, the various classes of antidepressant drugs target serotonergic and/or noradrenergic and/or dopaminergic neurotransmission, with the aim of increasing their synaptic concentrations. More recent studies conducted in patients with treatment-resistant depression have recognized the important pathogenic role of glutamate in depression, which has led to the successful testing of N-methyl-d-aspartate (NMDA) antagonist drugs (e.g., ketamine) (Zarate et al., 2006). The use of this class of drugs remains experimental, but their availability in the future may have greater significance in the management of depressed PWE, given the significant pathogenic role played by glutamate in the seizure disorder as well.

To date, the available families of antidepressant drugs include (1) tricyclic antidepressants (TCAs), which was the first one to be introduced; (2) selective serotonin reuptake inhibitors (SSRIs); (3) serotonin-norepinephrine reuptake inhibitors (SNRIs); (4) serotonin/antagonist reuptake inhibitors (i.e., trazodone, nefazodone); (5) dopamine and norepinephrine agonists (i.e., bupropion); and (6) α-2 antagonist (mirtazapine). Most of the available published data in PWE include open trials with TCAs and SSRIs. In fact, to date there has been only one double-blind placebo-controlled trial of antidepressant drugs in adults with PWE, published in 1975, which compared the efficacy of amitriptyline and mianserin (Robertson, 1985). A randomized double-blind controlled trial comparing the efficacy of the SSRI sertraline and CBT was presented at the annual meeting of the American Academy of Neurology in 2009 but has not been published yet beyond its abstract from the meeting (Gilliam et al., 2009). Given its importance, it deserves to be reviewed in detail (see below). Because there are no data available on the therapeutic expectation of antidepressant drugs in the treatment of depressive disorders in PWE, we have to extrapolate from that obtained in the management of primary depressive disorders.

Therapeutic expectations of pharmacotherapy

Primary major depressive episodes that are left untreated may last between 6 and 24 months in 90–95% of cases, whereas in the remaining 5–10%, they could last more than 2 years (Paykel et al., 1995 Charney et al., 1998; Stahl, 2000). Response to pharmacotherapy in depressive disorders has been divided into three phases: (1) Response, manifested by an improvement of 50% or more in the severity of the depressive symptomatology; this is typically observed in the first 8 weeks of pharmacotherapy. (2) Remission, reflected in a complete symptom remission. Approximately 50% of patients are expected to reach remission within the first 6 months and about two thirds within 2 years of the start of pharmacotherapy. (3) Recovery, defined as the state when remission has lasted for a period of 6–12 months. Symptom relapse is considered to occur when symptoms recur before a full remission has been achieved or before a state of remission has turned into a recovery. If on the other hand symptom recurrence takes place several months into a recovery, it is referred as a recurrence.

The response to pharmacotherapy in primary MDD is illustrated in the STAR*D study, which included 3,671 patients who were initially treated with the SSRI citalopram (Sackeim et al., 2006). Of note, close to 45% of patients were found to have a comorbid anxiety disorder as well as a medical disorder. Two thirds of patients entered remission, but only 37% after the first trial of citalopram. A change to a second antidepressant drug was necessary to achieve remission in an additional 30.6%. A third and fourth trial were necessary to yield remission in 13.7% and 13.0%, respectively. Yet, among all patients, sustained remission was maintained in only 43%.

The variables predictive of relapse include the following: (1) multiple prior episodes; (2) severe episodes; (3) long-lasting episodes; (4) episodes with psychotic or bipolar features; and (5) incomplete recovery between two consecutive episodes. As stated above, complete symptom remission should always be the goal of therapy, as persistence of residual symptoms is a strong predictor for recurrence of MDE (Paykel et al., 1995). Furthermore, premature discontinuation of antidepressant medication can result in relapse.

Pharmacotherapy in PWE

Before starting patients on an antidepressant drug, clinicians must ensure that the patient does not have a bipolar disorder, as the use of antidepressant drugs in a bipolar disorder can cause manic and hypomanic episodes and (even worse!) facilitate the development of a rapid cycling bipolar disorder, defined as a bipolar disorder with four or more depressive and/or manic episode within a 12-month period (Hirschfield et al., 2002). Rapid-cycling bipolar disease has been found in approximately 15% of patients with bipolar disorders and is relatively resistant to pharmacotherapy. The initial manifestations of a bipolar disorder can often consist of recurrent major depressive episodes, and several years can go by before the first manic or hypomanic episode occurs. Accordingly, before starting antidepressant medication for a depressive episode, clinicians must always inquire about any personal or family history of manic or hypomanic episodes, which is a strong risk factor for the development of this disorder in the patient. Finally, the occurrence of a first major depressive disorder before the age of 18 suggests a 40–50% risk that the patient may have bipolar illness. Given the greater severity and worse course of bipolar illness, this condition should be managed only by psychiatrists.

Available data

Following a review of the available literature, two groups of experts reached a consensus that antidepressant drugs of the SSRI family should be considered as first line for the treatment of depression in PWE, given their better safety and tolerability (Barry et al., 2008; Kerr et al., 2011). Yet, there has been only one double-blind study with SSRIs in PWE. It consisted of a comparison of sertraline and CBT in 144 PWE with MDE (Gilliam et al., 2009). At the end of the 16-week trial, 60% of patients in each treatment arm achieved full symptom remission. No recurrences were noted by 52 weeks. Of note, in this trial, the response to treatment with either sertraline or CBT appeared to be better than that published for patients with primary MDD (Sackeim et al., 2006). No further comparisons can be made, however, as Gilliam's study has yet to be published and the presence of comorbid anxiety was not included in the initial presentation.

Open trials with SSRIs have been published, suggesting efficacy in the treatment of the depressive disorder. For example, Kanner et al. (2000) found total symptom remission in 57 (59%) of 96 consecutive PWE treated with sertraline in monotherapy. Twenty-nine of these patients met criteria for MDE, whereas the other 67 had an atypical type of depression.

Hovorka et al. (2000) examined the efficacy of citalopram in 43 PWE with a depressive episode established on the 21-item Hamilton Depression Scale (HAMD 21; with a total score >15). The total score on the HAMD 21 decreased from a mean baseline of 21.5 ± 2.9 (range 17–26) to 14.5 ± 2.9 (range 10–19) (p < 0.001) after 4 weeks of treatment and to 9.9 ± 3.1 (range 4–19) (p < 0.001) after 8 weeks of treatment. Nine patients (20.9%) were considered to be responders after 4 weeks of treatment and 28 (65.1%) after 8 weeks. Specchio et al. (2004) conducted a 4-month multicenter open trial of citalopram at a dose of 20 mg/day to assess its safety in 45 PWE and depression. Response of the depressive symptomatology was a secondary outcome measured with the Montgomery-Asberg and Zung depression rating scales. Six patients withdrew from the study. At the end of the trial, 18% of patients experienced full symptom remission and 67% were considered to have experienced a moderate or marked improvement. Kühn et al. (2003) examined the effect of citalopram, mirtazapine, and reboxetine on the depressive episode of 75 PWE. The Hamilton Rating Scale for Depression was used to quantify the symptom severity at admission and after 4 and 20–30 weeks. By the end of the trial, symptom remission had been achieved in 21.2% of patients on citalopram, 14.8% on mirtazapine, and 20% on reboxetine, whereas 36%, 52%, and 53% had a >50% reduction in symptoms severity on the HDRS, respectively. With respect to TCAs, Blumer reported a positive response to the use of low-dose imipramine alone and/or in combination with low-dose sertraline in the treatment of interictal dysphoric disorder, an atypical form of depression in epilepsy (Blumer & Zielinski, 1988). Unfortunately, these findings need to be replicated in double-blind placebo-controlled trials given the relatively high placebo response to pharmacotherapy in primary depressive disorders.

Choice of antidepressant drug

Among the six families of antidepressant drugs, most of the published data in PWE include open trials with SSRIs, SNRIs, and TCAs and the norepinephrine and dopamine reuptake blocker (e.g., bupropion), which is one of the antidepressant drugs that is not recommended in these patients because of a proconvulsant effect in nonepileptic patients (Horne et al., 1988; Davidson, 1989; Jefferson et al., 2005; Alper et al., 2007). In developed countries, SSRIs and SNRIs have become the first line of pharmacotherapy for primary MDD, DD, and anxiety disorders, whereas in developing countries, drugs of the TCA family continue to be the first line of therapy, mainly because of economic reasons. Of note, no difference in efficacy has been demonstrated among SSRIs, SNRIs, and TCAs, but the two former are favored over TCAs because of their lower toxicity risk (particularly in the case of overdose) and better overall tolerance. Furthermore, the potential for cardiac arrhythmias mediated by a sodium channel blocking effect of high serum concentrations of TCAs needs to be factored in the choice of antidepressant drugs in PWE, given the increased risk of sudden death in PWE.

The SSRI family includes six drugs (fluoxetine, fluvoxamine, sertraline, paroxetine, citalopram, and escitalopram), which have as primary effect to increase the synaptic concentration of serotonin (Tollefson & Rosenbaum, 1998). As stated above, two separate consensus papers have supported the choice of this family of drugs as a first line of antidepressant pharmacotherapy in PWE (Barry et al., 2008; Kerr et al., 2011).

The SNRI family of antidepressants includes four drugs (venlafaxine and its active metabolite desvenlafaxine, duloxetine, and milnacipran), which act by increasing the synaptic concentration of serotonin and norepinephrine. Yet, the effect on these two neurotransmitters varies among the drugs, as it is observed at low doses of duloxetine, whereas it occurs only at moderately high doses of venlafaxine (>150–200 mg/day) (Kent & Gorman, 1998). There are no published data on the use of desvenlafaxine in PWE and milnacipran, which is also not available in the United States. Table 1 summarizes doses and their use with comorbid anxiety disorder:

If at maximal doses the patient continues to be symptomatic, a switch to another class of antidepressant is suggested (i.e., SNRI if the prior trial was with an SSRI). Failure to obtain full remission of symptoms after two trials at optimal doses should lead to referral to a psychiatrist.

One cautionary note is in order, however: the therapeutic effect of SSRIs and SNRIs may not be identified for 3–6 weeks after the start of the drug. Because SSRIs can at times cause restlessness and mild anxiety at the start of therapy, a short course of a benzodiazepine such as clonazepam (0.5–1 mg/day) should be considered in patients with comorbid anxiety and depression.

Adverse effects of SSRIs and SNRIs have been reported in about 20–30% of patients with primary mood disorders and include anxiety and agitation during the acute phase of treatment, gastrointestinal symptoms (i.e., nausea, abdominal cramping, and diarrhea), and changes in appetite and weight, sexual disturbances, and rarely involuntary movements. Furthermore, SSRIs can compound on a weight gain caused by AEDs such as valproic acid, gabapentin, pregabalin, and carbamazepine.

Among the SNRIs, hypertension is a potential adverse event identified in patients taking venlafaxine, but not duloxetine (Kent & Gorman, 1998). Therefore, blood pressures need to be monitored closely and venlafaxine and duloxetine should be used with great caution among patients with a risk or with existing hypertension. Duloxetine should be used with great care in patients with a history of liver disease and should be avoided in those with glaucoma.

All SSRIs and SNRIs can cause sexual disturbances, although the effect is less frequent with the latter. Accordingly, it is essential for clinicians to investigate the existence of sexual disturbances before starting an SSRI, as these are relatively common among PWE, either, as a direct consequence of the seizure disorder, as an adverse event in response to an AED, or a combination of the two factors. The most frequent common sexual disturbances include decreased libido, anorgasmia, impotence, and disturbances in ejaculation, and in women, dyspareunia. Citalopram and escitalopram and the SNRIs have been reported to have a lower incidence of sexual adverse events.

Several population-based studies have suggested that SSRIs can cause osteopenia and/or osteoporosis, which could worsen the effect of enzyme-inducing AEDs (Liu et al., 1998; Warden et al., 2005; Diem et al., 2007) .

TCAs were the first available agents and include 12 drugs, of which 7 are among the most frequently used (imipramine, desipramine, amitriptyline, doxepin, nortriptyline, protriptyline, and clomipramine) (Stahl, 2000). Their efficacy has been demonstrated in major depressive episodes, dysthymia, and generalized anxiety disorder and imipramine was the first drug with prophylactic efficacy in panic disorder, whereas clomipramine was the first agent with proven efficacy in the treatment of obsessive-compulsive disorders; yet, the relatively high proconvulsant risk of clomipramine precludes its use in PWE (Rosenstein et al., 1993).

Although the TCAs intended pharmacodynamic effects include an increase in synaptic serotonin and/or norepinephrine by blocking their reuptake (see Table 2 below), these agents have effects at muscarinic cholinergic receptors, histamine1 (H1) receptors, and alpha1-adrenergic receptors, which mediate several of their adverse events. For example, anticholinergic effects mediate dry mouth, constipation, urinary hesitation/retention, blurred vision, tachycardia, exacerbation of narrow angle glaucoma, problems with memory, and confusion. Blockade of H1 receptors has been associated with weight gain and sedation. Blockade of the alpha1 adrenoreceptor has been associated with postural hypotension, dizziness, and reflex tachycardia, and can potentiate the effects of antihypertensive drugs. The reuptake blocking of serotonin has been associated with sexual disturbances and gastrointestinal symptoms, as well as myoclonic jerks, and diaphoresis, whereas the reuptake blocking of norepinephrine can cause tremors and tachycardia and can exacerbate the pressor effects of sympathomimetic amines. Likewise, the blocking of sodium channels is associated with an increased risk of cardiac arrhythmias and epileptic seizures in case of overdoses.

Table 2. Summary of the target doses and magnitude of the effect in various receptors of the five most frequently used TCAs

As stated above, the use of TCAs has been relegated to third place in North America and Europe given its significantly greater toxicity risks, particular as it pertains to cardiovascular adverse events that significantly increase the lethality potential of these drugs in the case of overdoses. It is recommended that the patient be started on TCAs at a low dose, which should be titrated slowly (Table 2). Furthermore, serum concentrations of most antidepressants of this class are available, and it is recommended that they be measured once the desired dose is reached to identify those patients that may be “slow metabolizers.” Finally, baseline electrocardiography (ECG) studies and a repeat study, once the target dose of the TCA has been reached, are recommended to identify any electrocardiographic disturbance.

Clinicians must factor in the following variables in the process of choosing a first antidepressant drug among SSRIs, SNRIs, and TCAs:

  1. Whether the depressive disorder is retarded or agitated. Therefore, the use of an SSRI is suggested in case of a depressive disorder associated with anxiety, whereas an SNRI may be more beneficial in a retarded depressive episode in which the patient reports fatigue and slow thinking. Among TCAs, noradrenergic agents like desipramine may be more appropriate in cases of retarded depressive episodes, whereas amitriptyline should be considered in the case of agitated depressive episodes.
  2. The presence of comorbid anxiety disorder. The SSRIs and SNRIs with anxiolytic effect are listed on Table 1(see also article by Mula, 2013 in this issue). In the case of TCA, imipramine may be ideal in the treatment of depressive episodes with comorbid panic attacks, whereas for general anxiety disorder (GAD), imipramine, nortriptyline, desipramine, and doxepin in case of GAD.
  3. The potential adverse event profile of the specific antidepressant drug that could worsen underlying medical comorbidities associated with the seizure disorder or other concurrent medical condition, in particular adverse events caused by concomitant AEDs such as weight gain and sexual disturbances. In addition, SSRIs, SNRIs, and TCAs should not be discontinued abruptly as patients may experience withdrawal symptoms, including sensory disturbances and ataxia. These symptoms are less frequent in SSRIs with the longer half-lives, such as fluoxetine.
  4. Potential pharmacokinetic interactions with concurrent AEDs. Several drugs of the SSRI family are inhibitors of one or more cytochrome P450 (CYP) isoenzymes (Patsalos et al., 2005; Trimble & Mula, 2005). For example, fluoxetine has been shown to inhibit several CYP isoenzymes including CYP3A4, CYP2C9, CYP2C19, CYP2D6, and CYP1A2 (Patsalos et al., 2005; Trimble & Mula, 2005). The active metabolite of fluoxetine, norfluoxetine, has also been shown to inhibit CYP2D6, with inhibition of CYP3A4, CYP2C9, and CYP2C19 yielding the most significant potential interactions with AEDs. Furthermore, several case reports have suggested that SSRIs such as fluoxetine and sertraline have caused increased phenytoin and carbamazepine serum concentrations (Nelson et al., 2001). The antidepressant fluvoxamine, is an inhibitor of CYP1A2, CYP3A4, and a potent inhibitor of CYP2C9 and CYP2C19. Comedication therefore is likely to cause marked increases in phenytoin concentrations (Mamiya et al., 2001). In settings where higher SSRI doses are used, or perhaps in elderly patients who may have reduced clearance of both phenytoin and the SSRI, the potential for a clinically meaningful interaction may be increased.

The SSRIs with the least potential for causing inhibitory interactions are citalopram and escitalopram, followed by sertraline with mild inhibitory effects, paroxetine, fluoxetine with moderate effects and fluvoxamine with maximal effects. Although definitive studies are lacking, it has also been suggested that venlafaxine and duloxetine are unlikely to cause significant interactions with currently available AEDs.

Conversely, first-generation AEDs such as phenytoin, carbamazepine, and phenobarbital are potent inducers of the CYP enzyme system (Patsalos & Perucca, 2003). Oxcarbazepine and topiramate at high doses are much weaker inducers of CYP3A4. AEDs including gabapentin, pregabalin, lamotrigine, levetiracetam, tiagabine, zonisamide, ezogabine, clobazam, vigabatrin, and perampanel do not appear to interfere with CYP activity or other metabolizing pathways (i.e., glucuronidation). Because most SSRIs, SNRIs, and TCA are substrates for one or more or the CYP isoenzymes (Spina & Perucca, 2002; Trimble & Mula, 2005), comedication with an enzyme-inducing AED would be expected to increase their systemic clearance resulting in lower serum concentrations. This effect has been demonstrated in antidepressants such as sertraline, paroxetine, citalopram, as well as in most TCAs, requiring an adjustment of their dose (by 25–30%) to avert recurrence of psychiatric symptoms.

In contrast to the enzyme-inducing drugs, the AED sodium valproate can inhibit certain CYP (2C9) and UDP-glucuronyltransferase enzymes, and may cause significant increases (50–60%) in serum concentrations of antidepressants such as amitriptyline or nortriptyline (Spina & Perucca, 2002).

Do Antidepressant Drugs Cause or Worsen Seizures?

The belief that SSRIs, SNRIs, and TCAs have proconvulsant properties is a common misconception held by clinicians and remains a big obstacle in providing pharmacologic treatment of depressive and anxiety disorders to PWE. Yet, a careful review of the literature has revealed that seizure occurrence has been associated with high serum concentrations caused by overdoses or their slow metabolism (Preskorn & Fast, 1992; Rosenstein et al., 1993). Four antidepressants can cause seizures at therapeutic doses: clomipramine, maprotiline, amoxapine, and bupropion, although the evidence of the latter is still the subject of debate. Furthermore, the occurrence of epileptic seizures has to be interpreted in the context of a bidirectional relation between depressive (and/or anxiety disorders) and epilepsy, whereby not only are patients with epilepsy at greater risk of developing these two psychiatric comorbidities, but patients with these psychiatric disorders have a twofold to sevenfold higher risk of developing epilepsy (Hesdorffer et al., 2000, 2006, 2012). Therefore, the occurrence of a seizure in a depressed patient may be the expression of the natural course of the mood disorder and not of an adverse event of the antidepressant drug. This is also illustrated in a study by Alper et al. (2007) who compared the incidence of epileptic seizures in depressed patients randomized to placebo or one of several SSRIs, the SNRI venlafaxine, or the α2- antagonist mirtazapine in the course of phase II and III clinical regulatory trials. The incidence of seizures was significantly lower among patients randomized to antidepressants compared to placebo (standardized incidence ratio = 0.48; 95% confidence interval [955 CI] 0.36–0.61). Of note, in both patients assigned to antidepressants and placebo, the seizure incidence was 19-fold higher than the published incidence of unprovoked seizures in the general population. Furthermore, SSRIs have been found to be safe in open trials in PWE, some of which have even suggested that they may exert a reduction in seizure frequency (Favale et al., 2003, 2005). For example, sertraline was found to definitely worsen seizures in only 1 of 100 consecutive patients with pharmacoresistant epilepsy (Kanner et al., 2000).


Among the various forms of psychotherapy, CBT has been proven to be effective in the treatment of depression and anxiety disorders in patients with and without epilepsy in controlled studies. This form of therapy should be considered for patients who refuse to take antidepressant drugs or who cannot tolerate the side effects of these drugs. In addition, it should be considered in PWE who have a chronic depressive disorder, in particular if associated with an anxiety disorder.

Among the three studies published on the efficacy of CBT in PWE, the double-blind, controlled study by Gilliam et al. (2009) cited above yielded complete symptom remission in 60% of PWE with a major depressive episode, which was identical to the efficacy of sertraline. One controlled and two open trials supported the efficacy of CBT in the treatment of depression and anxiety in PWE. The controlled study consisted of a 6-week group CBT to assess its effect on seizure frequency and changes in mood and psychosocial well-being (McLaughlin & McFarland, 2011). The study was conducted in 37 elderly PWE aged 60 years and older. Patients were randomized to receive group CBT or be part of a control group. Measures of depression, dysthymia, psychosocial functioning, and seizure frequency were recorded at baseline and at the end of the intervention. Seizure frequency was significantly reduced in the CBT group compared to the control group. The CBT and control groups improved significantly from baseline, and there were no significant group differences on measures of depression and psychosocial functioning. The two other studies were not controlled. In one, 18 PWE underwent a 10-week group-CBT uncontrolled intervention for symptoms of depression and/or anxiety (Macrodimitris et al., 2011). At the end of the trial, patients exhibited improvements in depression, anxiety, negative automatic thoughts, and cognitive therapy knowledge and skills. The second study also consisted of uncontrolled group CBT for the treatment of major depressive episode in 23 patients with temporal lobe epilepsy (Crail-Melendez et al., 2012). Patients were divided into two groups to receive 16 weekly sessions of CBT. The primary outcome measures were depression severity (assessed with the Beck Depression Inventory) and quality of life (measured with the Quality of Life in Epilepsy-31). Sixteen patients (70%) completed at least 80% of the sessions. Efficacy was identified as of week 8 with respect to severity of depression; this effect persisted until the end of treatment. A significant improvement in quality of life was also observed.

When patients should be referred to a psychiatrist

Automatic referrals to a psychiatrist include the presence of depressive episodes associated with an increased suicidal risk, with psychotic symptomatology, in depressive episodes that are part of a bipolar disorder and in patients whose depressive episode have failed to reach complete symptom remission after two trials of antidepressant drugs (with a different class of drug) at optimal doses.

Concluding Remarks

The relatively high prevalence of depressive disorders in PWE requires that all neurologists be able to recognize them and when appropriate begin therapy or refer for treatment. Comorbid anxiety symptoms and/or disorders are frequent and their presence should be investigated, as their treatment should also be incorporated in the treatment plan. Although sparse controlled data are available on the efficacy of antidepressant drugs in PWE, reports from open trials have suggested that these drugs appear to be effective. Contrary to long-held beliefs, the use of antidepressant drugs are safe in PWE when used at therapeutic doses. Antidepressant drugs of the SSRI or SNRI families are the first line of therapy in depressive and anxiety disorders, and failure to achieve complete symptom remission after a trial of an SSRI or SNRI at optimal doses should be followed by a second trial with a drug from the other antidepressant family. Failure to reach symptom remission after these two trials may be suggestive of a treatment-resistant depressive disorder, in which case patients must be referred for treatment by an experienced psychiatrist.

In developing countries, antidepressant drugs of the SSRI and SNRI families are not always available, and TCAs are the drugs of choice. Although there are no differences in efficacy among the three families of antidepressants, TCAs have a lower tolerability and higher toxicity, with greater mortality risk associated with cardiotoxic effects.

Cognitive behavior therapy is another treatment modality that has been shown to be effective in the treatment of depressive and anxiety disorders in patients with and without epilepsy. Its use should be considered together with pharmacotherapy or by itself. In summary, the treatment of comorbid depressive and anxiety disorders in PWE should be incorporated in the overall management of these patients.


The author has received research funding from Pfizer Laboratories. I confirm that I have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. The contents of this supplement reflect the opinions of the individual authors and do not necessarily represent official policy or position of the ILAE.