example, found that in-hospital poststroke depression was the most important variable that predicted poor recovery in activities of daily living over a 2-year period [13]. Of note, the score of in-hopsital activities of daily living was not associated with the 2-year recovery. Chemerinski et al. found that successful treatment of poststroke depression with nortriptyline was significantly associated with recovery in activities of daily living [14]. A significant relationship between the presence of poststroke depression and a higher mortality risk following stroke was reported by Wade et al. in a study of 976 patients followed for 1 year [15]. Patients with poststroke depression had 50% higher mortality than those without. Robinson et al. found that patients treated with fluoxetine or nortriptyline had an increased survival probability at 6 years (61%) compared with patients given placebo (34%) [16]. Robinson also found that treatment with antidepressant medication was an independent predictor of increased survival [17].

In patients with epilepsy the impact of depressive disorders in their quality of life is equally significant. In a study of 56 patients carried out in Germany, Lehrner et al. found that depression was the single strongest predictor for each domain of health-related quality of life instrument [18]. The significant association of depression with health-related quality of life measures persisted after controlling for seizure frequency, seizure severity, and other psychosocial variables. Gilliam et al. also found that mood status was the strongest clinical predictor of the patients' assessment of their own health status in a group of 125 patients more than 1 year after temporal lobe surgery [19]. In a separate cohort of 194 epilepsy clinic patients, they also found that a depressed mood and neurotoxicity to antiepileptic drugs were the only variables with a significant correlation with poorer self-reported health status [20].

Given these data, we would expect a very close collaboration between neurologists and psychiatrists in the evaluation and management of their neurologic patients. It would follow from the above-cited studies that patients with stroke, epilepsy, Parkinson's disease, or multiple sclerosis, to cite only some neurologic disorders, would be screened for psychiatric comorbidity with the aim of incorporating the appropriate psychiatric treatment into the overall management. This is not the case, however, as illustrated in a study of 226 patients seen at a neurology clinic in which depressive disorders were identified in 88 (40%) patients, 54 of whom (26%) had major depression. Eight months later, 69 (78%) patients continued to display symptoms of depression; 46 (85%) of the 54 patients with major depression continued suffering from major depression [21].

A review of the literature on the treatment of psychiatric comorbid disorders in the major neurologic disorders reveals a common finding: psychiatric comorbid disorders remain unrecognized and untreated in a large

percentage of patients [22]. The purpose of this article is to examine some of the more obvious obstacles that get in the way of neurology patients' receiving the appropriate psychiatric evaluation and treatment. While I restricted the discussion to the evaluation of patients in an epilepsy clinic, the conclusions are applicable to other fields of neurology.

2. An example of poor communication

One of every four to five patients referred to a video- EEG monitoring unit with a diagnosis of intractable epilepsy does not have epilepsy. The majority of these patients have psychogenic nonepileptic events (PNES). Ideally, after the diagnosis is established patients are expected to be enrolled in some type of psychiatric treatment. A recently published study of 174 patients with PNES showed that more often than not, this does not happen: almost 82% of patients were readmitted to a neurologic ward and 40.7% continued on antiepileptic drugs after the diagnosis was reached (this percentage excluded patients who had epileptic seizures in addition to psychogenic events) [23].

The recurrent admission to a neurology service after diagnosis and the continuous intake of antiepileptic drugs (AEDs) in the absence of concurrent epileptic seizures clearly suggest a lack of communication between neurologists and psychiatrists (or other mental health professionals). The evidence of such lack of communication is illustrated by a recent study aimed at identifying the opinions of neurologists and psychiatrists (and their trainees) on the evaluation and management of PNES. Seventy-five psychiatrists and fifty neurologists were surveyed regarding the diagnostic significance of video-EEG data in the evaluation of patients with PNES [24]. Only 18% of psychiatrists stated that video- EEG is an accurate diagnostic method "most of the time" for patients with PNES (in contrast to 70% of neurologists). This difference in opinion was also found between psychiatry and neurology trainees. These data speak for themselves.

3. The postsurgical psychiatric risks that continue to be ignored

The prevalence of psychiatric comorbid disorders is higher among patients with refractory epilepsy. Thus, various series have identified depression in up to 50% of patients, to name only one comorbid disorder [22]. A significant percentage of patients refractory to AED therapy undergo a presurgical evaluation to establish their eligibility for surgical treatment. In the majority of surgical centers in the United States, the evaluation for an anterotemporal lobectomy includes high-resolution brain

MRI, video-EEG studies, a neuropsychological evaluation, and an intracarotid sodium amytal test (Wada test). However, only a minority of centers perform a psychiatric evaluation as part of their presurgical evaluation. Furthermore, in less than 25% of major epilepsy centers surveyed the epilepsy team includes a psychiatrist who is available to evaluate every patient undergoing a presurgical evaluation. And yet, de novo psychiatric complications have been identified following temporal lobectomies, the most frequent of which are depressive disorders during the first year after surgery [25]. It is not unusual to see ‘‘mood lability’’ within the initial 6 weeks of surgery in more than 50% of patients. Often these symptoms subside, but in 30% of patients, overt depressive episodes become apparent within the first 6 postoperative months that vary in severity from mild to very severe, including suicidal attempts. Patients with a prior history of depression are at greater risk. Of note, this risk is independent of the postsurgical control of seizures.

Postoperative psychosis has also been reported after temporal lobectomies for epilepsy. In a series of 100 of Falconer's patients, Taylor reported 7 with de novo postoperative psychosis [26]. Jensen and Vaernet reported de novo psychotic disorders in 9 of 74 patients [27] and Trimble calculated postoperative de novo psychoses to range between 3.8 and 35.7% (mean, 7.6%) of patients [28]. The obvious question is: Why is a psychiatric evaluation not included in the presurgical protocols of these centers given the relatively high prevalence of (presurgical) psychiatric comorbidity and the risk of de novo (or exacerbation of existing) depression and psychosis postsurgically? After all, all epilepsy centers perform a neuropsychological evaluation in every patient who undergoes a presurgical evaluation to assess the postsurgical risk of memory deficits.

I have heard different opinions with respect to this issue. The most frequent are the following two: (1) A neuropsychological evaluation is sufficient to screen for psychiatric comorbidity. (2) A psychiatric evaluation is necessary only in patients with known psychiatric history.

Can a neuropsychological evaluation replace a neuropsychiatric evaluation? In my opinion, neuropsychological and neuropsychiatric evaluations provide different data that complement each other. Thus, a neuropsychological evaluation cannot and should not substitute for a psychiatric evaluation. The former provides primarily data on the patient's performance in the various cognitive domains. While some neuropsychologists include a variety of screening instruments of psychopathology (Minnesota Multiphasic Personality Inventory [MMPI], the Beck Depression Inventory, etc.), we must recall, however, that such instruments are not diagnostic but merely screening tools. The neuropsychiatric evaluation provides detailed clinical data on the presence and type of comorbid psychiatric disease, its risk factors (i.e., iatrogenic and genetic), the longterm course of the

disease, its relationship to (or independence from) the epileptic seizure disorder, and the need to begin pharmacologic treatment or other form of therapy before or after the surgical procedure.

A common example of how relevant clinical data may be missed without a proper psychiatric evaluation is illustrated here. A patient may be euthymic at the time of the presurgical evaluation and for the previous year, but may have suffered one or more depressive episodes several years before. That history would not be obtained without a carefully taken psychiatric history; a screening instrument would fail to detect these data. Their recognition would make the clinician aware of the higher risk this patient has of experiencing a postsurgical depressive episode.

It should be noted that neurologists have relied with increasing frequency on screening instruments of depression to decide on the need to start pharmacotherapy with antidepressant medication. Yet, disastrous consequences may result in some cases when this decision is reached without a psychiatric evaluation. For example, a patient with bipolar disease who is found to be depressed and is started on antidepressant therapy on the basis of a high score on the Beck Depression Inventory can be at significant risk of developing a manic episode triggered by the antidepressant drug.

Also, various authors have recognized atypical clinical manifestations of depressive and psychotic disorders in epilepsy [29–31]. These are not identified with diagnostic instruments developed for nonepileptic patients, however, and can be detected only in the course of a psychiatric interview. The sole reliance on neuropsychological evaluations has relegated a psychiatric evaluation for patients with the more severe psychopathology. We have to recognize as well that in many epilepsy centers in the United States the psychiatrist has been displaced by the neuropsychologist.

Should psychiatric evaluations be restricted to patients with the more severe psychiatric disease? As stated above, several studies have documented that psychiatric comorbid disorders of epilepsy and specifically depression are very often underrecognized and undertreated. Thus, reliance on patients' reporting the presence of a psychiatric disorder without having done a proper neuropsychiatric evaluation is likely to result in a failure to detect a significant number of symptomatic patients during presurgical evaluations. Patients with a chronic history of dysthymia may deny being depressed, as they assume that their dysphoric state is ‘‘a normal’’ mood of patients with epilepsy. Such patients are at risk of a postsurgical worsening of their mood disorder.

If the goal of any presurgical evaluation is to recognize all postsurgical risks, why are psychiatric evaluations not performed in all patients? Clearly, the arguments against the inclusion of a psychiatric evaluation as part of any presurgical evaluation are another

example of poor communication between neurologists and psychiatrists.

4. Other common consequences of the lack of communication between neurologists and psychiatrists

As stated above, psychiatric comorbid disorders in epilepsy are underrecognized and undertreated. However, underrecognition does not account for all the causes behind the failure to treat these disorders. Many clinicians continue to have misconceptions on the potential impact of psychotropic drugs in lowering the seizure threshold, which leads them to recommendations against the use of psychotropic drugs in epilepsy patients. We cite the most frequent misconceptions encountered daily.

Attention deficit hyperactivity disorder (ADHD) has been recognized in approximately 30% of children with epilepsy [32]. The use of central nervous system stimulant drugs like methylphenidate has been proven to be safe and effective in the treatment of ADHD in epilepsy. Yet, there is a wide misconception that these drugs lower the seizure threshold and hence many clinicians refuse to prescribe them to these children.

A similar concern exists with respect to the use of antidepressant drugs. While antidepressant drugs can lower the seizure threshold at toxic doses in nonepileptic patients, the proper doses of most antidepressants belonging to the selective serotonin reuptake inhibitor (SSRI), selective serotonin norepinephrine reuptake inhibitor, tricyclic antidepressant, and monoamine oxidase inhibitor classes are safe in epilepsy patients. This is also a frequent reason for undertreatment of depression in epilepsy. Better communication between neurologists and psychiatrists would eliminate such misconceptions.

5. Who is to blame?

There is a lack of proper training of both neurologists and psychiatrists. Psychiatry residents have very limited training in neurology while neurology residents are not required to have any formal training in psychiatry. Indeed, in residency programs in the United States, psychiatry residents are required to spend only 2 months of their training in a neurology service, while neurology residents are not required at all to rotate through a psychiatry service during the course of their residency. And yet, the authorities in charge of regulating the curricula of these two specialties recognize the need for neurologists to have a certain level of training in psychiatry and vice versa. Indeed, the Psychiatry Board examination includes a section of neurology questions while the Neurology Board has a psychiatry section that all candidates must pass to be board certified. How do we reconcile these two facts?

Furthermore, it is very difficult for me to understand the rationale for limiting the training of psychiatry residents in neurology to 2 months given the direction that modern psychiatry has taken in attempting to identify the neurologic substrates of psychiatric disease. Also, the abolition of a required rotation in psychiatry by neurology trainees makes no sense, given the frequent psychiatric comorbidity in neurologic disorders and the data cited above on the impact of psychiatric comorbid disorders on the neurologic recovery of patients.

6. Neurologists and psychiatrists have a lot to talk about

The recent advances of modern psychiatry clearly show that if anything, the fields of neurology and psychiatry are converging to common areas. To cite one example, recent MRI studies of patients with major depression have revealed a decrease in the volume of hippocampus, prefrontal cortex, and basal ganglia and the presence of bifrontal areas of increased signal in white matter of frontal lobes [33]. The proposed mechanisms for the decrease in hippocampal volume include the development of atrophy mediated by high glucocorticoid exposure and/or an alteration in brain neurotrophic factors as a result of serotonergic disturbances associated with the mood disorder. Can these structural MRI changes explain the negative impact that poststroke depression has on the cognitive recovery of stroke patients alluded to above?

7. Concluding remarks

The few data cited in this article clearly establish that the separation between the fields of neurology and psychiatry reflected in a ‘‘mind–body duality’’ that began in the 19th century should be relegated to the history books. Neurologists are as responsible as psychiatrists in eliminating the stigmatization of mental illness and the misconception (widely prevalent in many cultures of developed and underdeveloped countries alike) that psychopathology is a character flaw and not the result of biological dysfunction. By the same token, neurologists must stop considering neurologic illnesses as the expression of only somatic and/or cognitive disturbances. Comprehensive treatment of people with neurologic diseases requires that their psychiatric manifestations be recognized and their treatment be incorporated into the overall management. The fact is that neurologists and psychiatrists are talking the same language. . . : They just have to become aware of it!

7.1. On a personal note

Since the completion of my residency trainings in psychiatry and neurology I have been impressed by the

rather distant professional and scientific relationship between neurologists and psychiatrists. I expected such a relationship to be significantly closer between psychiatrists and epileptologists, since epilepsy is the neuropsychiatric disorder ‘‘par excellence.’’ Yet, I was wrong again. I have inquired whether this was unique to America. It is not. Neurologists and psychiatrists from countries like Canada, Spain, France, Mexico, Argentina, Chile, and Peru, to name a few, have echoed similar observations.

I have wanted to write this article for a long time but could not decide on the best way of presenting the concerns I had without taking the risk that neurologists or psychiatrists, or both, would take offense. In the end, I came to the conclusion that it is impossible to write this article without making neurologists and psychiatrists alike uncomfortable, because the issues I raise are indeed disturbing. I amsure that many will disagree with some or all the views presented in this article. I welcome those dissenting opinions and hope that they will be voiced in letters to the editor. The one thing I am sure of is that it is time that we start confronting the way neurologists and psychiatrists collaborate with each other as such a relationship is having an impact on the management of patients.

References