Although schizophrenia and depression have historically been considered to be separate diagnostic disorders, the observation has repeatedly been made that, from a symptom perspective, depression-like phenomenology occurs quite frequently in schizophrenia (1–4). In the early decades of psychopharmacology, however, it seemed most important to differentiate schizophrenia and depression in order to define most clearly which patients should be treated with antipsychotic agents and which with antidepressants. Nevertheless, the occurrence of the phenomenology of depression in a substantial percentage of patients with schizophrenia (as well as psychosis in patients with depression) has kept alive the twin issues of the appropriate descriptive boundaries between the two disorders and the best approaches to treatment. Indeed, these issues assume added importance, since it has been found that the occurrence of depression in schizophrenia has often been associated with worse outcome (5), impaired functioning, personal suffering (6), higher rates of relapse or rehospitalization (7–10), and even suicide (8, 11, 12)—a tragic event that terminates the lives of an estimated 10% of patients with schizophrenia (11, 13).
This article considers our current conceptualizations of depression-like symptoms in patients with schizophrenia, documents their common occurrence, and reviews the differential diagnosis of conditions that such symptoms might reflect. It also presents a potential integrating model that may facilitate the understanding of depression in at least some individuals with schizophrenia. Last, this article explores the implications for treatment as we enter a new era in schizophrenia pharmacotherapy occasioned by the advent of the so-called atypical antipsychotic agents.
Affect, Symptom, or Syndrome?
Some of the confusion attending the term "depression" in schizophrenia is attributable to the question of whether it is the affect of depression, the symptom of depression, or the syndrome of depression that is being considered. This confusion is not limited to patients with schizophrenia, of course, but it certainly has confounded the schizophrenia literature. Depression as an affect reflects an individual’s momentary mood state on the subjective experience spectrum from happiness to sadness as he or she interacts with his or her internal and external environments. It is not, in itself, pathological as long as it is situationally appropriate (e.g., a touching movie may elicit sad affect). Depression as a symptom is a sad mood state that causes a person distress. It is an unwanted painful feeling and can be a source of complaint. However, it is not necessarily enduring or accompanied by other features that are required for the diagnosis of the syndrome of depression. The depression syndrome is a complex of features that typically involves the symptom of depression but also includes cognitive and vegetative features such as pessimism, guilt, impaired concentration, lack of confidence, loss of interest or pleasure, and disturbances in sleep, appetite, and energy level. Unfortunately, the literature on depression in schizophrenia is often imprecise as to whether the affect, symptom, or syndrome of depression is involved, and these three meanings have often been employed loosely and interchangeably. The literature needs to be read carefully in this regard, and, most important, patients need to be listened to closely to understand the type of "depression" that they are experiencing. Differences in definition may account for some of the discrepancies in the reported occurrence rates and treatment responses of depression in schizophrenia, and this situation has contributed to the creation of the proposed syndromal definition "postpsychotic depressive disorder of schizophrenia" in Appendix B of DSM-IV (pp. 711–712).
Rates of Occurrence of Depression in Schizophrenia
Over three dozen published studies have examined rates or dimensions of depression in schizophrenia. These have been reviewed elsewhere (4, 6, 14–18). Investigations have varied considerably in terms of the definitions employed for schizophrenia and depression, the observed interval, the assessment methodology, and the patient status. Nevertheless, at least some meaningful point prevalence of depression was observed in each study in which it was assessed. Proportions of patients manifesting depression ranged from a high of 75% to a low of 7%. The 75% figure reflected at least one positive assessment of depression according to either of two criteria (one syndromal, the other based on a rating scale score) among patients with first-episode schizophrenia evaluated longitudinally on a weekly to monthly basis for up to 5 years (19). The 7% observation involved a single cross-sectional assessment of a rating scale item in chronically hospitalized patients with schizophrenia in whom an effort was made to distinguish depression from negative symptoms (20). The modal rate of depression for all reports was 25% (3, 4, 7, 14).
Symptom factor analysis studies that have attempted to go beyond the basic positive/negative symptom distinction by expanding to five dimensions have identified a depression or anxiety/depression factor, whereas those studies that used fewer dimensions have not (15–18). Results of factor analyses, of course, are dependent on exactly what symptom items are employed, and in this way initial assumptions can impact results substantially.
A number of medical/organic factors can present as depression in patients with schizophrenia (21). These include cardiovascular disorders, pulmonary infections, autoimmune diseases, anemia, cancer, and metabolic, neurological, and endocrine disorders. Various pharmaceuticals used in medical treatment such as β blockers, other antihypertensive agents, sedative hypnotics, antineoplastics, barbiturates, nonsteroidal anti-inflammatory drugs, sulfonamides, and indomethacin can cause depression as a side effect. Depression can also accompany the discontinuation of other prescribed medications such as corticosteroids and psychostimulants. Used or abused substances such as alcohol, cannabis, cocaine, or narcotics can contribute to depression either on the basis of acute use, chronic use, or discontinuation. It is also important to note that the discontinuation of two very commonly used legal substances, nicotine and caffeine, can lead to withdrawal states that potentially mimic depression (22, 23). Recent "smoke-free" and "decaf" policies on many inpatient units may have led to higher rates of this form of depression.
Negative Symptoms of Schizophrenia
The negative symptom syndrome of schizophrenia overlaps with the syndrome of depression in a number of important respects (24–27). Diminished interest, pleasure, energy, or motivation along with psychomotor retardation and impaired ability to concentrate are relevant overlapping features. However, certain other symptoms may be more distinguishing (28–31). Blunted affect, for example, suggests negative symptoms, whereas distinct blue mood or cognitive features such as guilt or suicidal thoughts suggest depression. Unfortunately, differentiating these two states can sometimes be difficult if patients lack the interpersonal communication skills to articulate their internal subjective states well.
Dopamine synapses are involved in brain pathways mediating "reward" (32, 33). Therefore, dopamine blockade by a neuroleptic drug could theoretically lead to anhedonia and, perhaps, depression. Indeed, a state of dysphoria is commonly described by neuroleptic-treated patients (34), a number of older anecdotal reports have suggested a link between neuroleptic use and depression (35–39), and one study found more anhedonia and depression in maintenance-phase schizophrenic patients who were taking neuroleptics than in others who were not (33). Another study found a positive relationship between haloperidol plasma levels and depressive symptoms in the context of a positive association between extrapyramidal symptoms and depressive symptoms (40), and impairments of quality of life related to neuroleptic-induced dysphoria have been reported (41). Nevertheless, the preponderance of controlled-study evidence tends to refute the proposition that neuroleptic medication is regularly responsible for the development of depressive states in schizophrenia (4 , 42). Prospective observations that have followed schizophrenic patients through the treatment of acute psychotic episodes suggest that depressive symptoms were present before the neuroleptic was given and tended to subsequently subside (20, 42–46). Studies that compared patients with schizophrenia who were being treated with neuroleptic medications versus those who were not did not find that neuroleptic-treated patients manifested more depression (20, 47–49). And studies examining schizophrenic patients with and without depression found that these two groups did not show any differences in neuroleptic doses or blood levels (8, 50–54). Nevertheless, one biological possibility is that schizophrenia may represent a basic disorder of dopamine regulation (55) in which "brittle" patients are vulnerable to dopamine storms (psychosis) and droughts (negative symptoms). In this situation, the administration of more than the minimum required neuroleptic (dopamine-blocking) medication could exacerbate negative symptoms, thereby possibly contributing to the impression of neuroleptic-induced depression.
Rifkin et al. (56, 57) and Van Putten and May (58) went beyond the original "large muscle stiffness" definition of akinesia to describe a more subtle but equally debilitating extrapyramidal side effect of neuroleptic treatment involving impaired ability to initiate and sustain motor behavior. Patients with this form of akinesia may or may not have the classical parkinsonian feature of decreased accessory motor movements. However, they act "as if their starter motor is broken," and they consequently appear to lack spontaneity. Since so much of human interchange, such as holding a conversation or participating in activities, involves the initiation or maintenance of motor behavior, this side effect leads to exclusion from much of what is normal in life. Patients themselves may attribute this effect to "laziness," experiencing guilt or shame. Blue mood can also accompany this condition, possibly as a primary issue (58), making it virtually indistinguishable clinically from depression. Unfortunately, most studies of depression in neuroleptic-treated patients have not adequately considered this form of akinesia as a potentially confounding factor.
Akathisia is another extrapyramidal side effect of neuroleptic treatment that, in subtle presentation, can easily be confounded with depression (59). Patients with akathisia behave "as if their starter motor won’t stop" and often experience this state as substantially dysphoric (59, 60). Indeed, akathisia has been associated with both suicidal ideation and—perhaps as a consequence of a general tendency toward motor action—suicidal behavior (61, 62). As with akinesia, akathisia has seldom been considered as a possible confound in studies of depression in schizophrenia.
Reactions to Disappointment or Stress
Reactions to disappointments, a sense of loss or powerlessness, or awareness of psychotic symptoms or psychological deficits can certainly present as or contribute to depression, especially when depression follows closely after a stressful event or exacerbation of schizophrenia (3, 10, 63, 64). Such reactions are of two types: acute and chronic. An acute reaction to disappointment or stress is suggested by the parallel history of a recent compatible event and is sometimes validated by the patient making a psychological connection between that event and a mood change. Acute reactions are generally brief, lasting hours, days, or at most weeks, and may be responsive to supportive interventions or counterbalancing positive experiences. Chronic reactions to disappointment or stress have also been termed the demoralization syndrome (65–67). By definition, these last longer and are apt to be more difficult to distinguish from other forms of depression. Indeed, schizophrenic patients who experience less of a sense of control regarding their illness (one of the hallmarks of demoralization) have been found to be more likely to experience depression (10). Demoralized patients with extended histories of disappointment or failure can also develop the conviction that a useful or satisfying life is impossible, further blurring the distinction from other forms of depression. Demoralization is important to diagnose because it may be more responsive than other depressed states to appropriately targeted psychosocial interventions.
In earlier times, the term "postpsychotic depression" was used to describe a dysphoric state that immediately followed a psychotic episode (3). In the sense that this was reactive to disappointment or stress, such an episode might belong in the previous category. DSM-IV now suggests that the term "postpsychotic depression" be used to describe depression that occurs at any time after a psychotic episode in schizophrenia—even after a prolonged interval. This definition would include many of the other defined categories in this article.
Prodrome of Psychotic Relapse
Investigations into the course of psychotic decompensation in schizophrenia have frequently noted the appearance of depression-like symptoms (9, 20, 45, 68–72). Anxiety, withdrawal, guilt, and shame commonly accompany dysphoria in this situation. Clues to the underlying circumstance of psychotic relapse may be found, however, in signs or symptoms of early psychotic decompensation, such as hypervigilance, perceptual disturbances, or the overinterpretation of perceptions or events. The appearance of a depression-like state as a prodrome of a new psychotic episode is a short-lived phenomenon, often lasting only a couple days to a couple weeks, before being superseded by more prominent and definitive psychotic symptoms.
The term "schizoaffective disorder" was first used in the early 1930s to describe patients showing an overlap of features of schizophrenia and affective illness (73). Over time, schizoaffective disorder has been defined differently according to various diagnostic schemes (74–76), which has resulted in variations in the boundary between schizoaffective depression and depression in schizophrenia. In DSM-IV, schizoaffective disorder refers to patients in whom a full affective syndrome coincides with the florid psychotic syndrome but who also have substantial periods of psychosis in the absence of an affective syndrome. This definition does not resolve the etiology or pathophysiology of the condition, and debate continues as to whether schizoaffective disorder should be considered a type of schizophrenia, a type of affective disorder, sometimes one and sometimes the other, a distinct entity, a domain on a continuum between schizophrenia and affective disorder, a co-occurrence of two distinct diatheses, or an erroneous concept altogether. The descriptive diagnosis of schizoaffective disorder, however, has value in defining phenomenologically an interesting patient group for which biological studies and clinical trials can be undertaken.
The well-known stress-diathesis model of schizophrenia (77, 78) depicts the psychosis of schizophrenia as a final common path of neuropsychiatric decompensation. The central notion is that vulnerability to schizophrenic psychosis occurs on a continuum in the population, from a tiny fraction of persons with such a strong vulnerability that psychosis is virtually inevitable to the overwhelming majority of the population in whom the vulnerability is so slight that the risk of psychosis is virtually negligible. In between, there is a small, yet meaningful, portion of individuals who could become psychotic if stressed enough but who could also survive without psychosis if not sufficiently stressed.
The stress in this model can be biological or psychosocial, and many relevant stresses have been described. Intrauterine viral infection, poor prenatal nutrition, birth injuries, and childhood head trauma would all be examples of early-life biological stressors. Substance abuse is an example of a biological stressor occurring more proximally to a first or subsequent psychotic episode. Traumatic interpersonal experiences, more longitudinal psychic stresses (such as chaotic, abusive, or high "expressed emotion" environments), and lack of opportunity to develop adequate compensating coping skills are examples of psychosocial stressors.
It is provocative to speculate that the activation of an affective diathesis, such as depression, could act as a sufficient stressor precipitating psychosis in people with otherwise modest vulnerabilities. Certainly depression is a highly stressful state psychologically, and it is possible that it is stressful in relevant biological ways as well (e.g., insomnia, hormonal changes, neurotransmitter effects). And since within the vulnerability continuum there are many more people with moderate than with extreme vulnerabilities, a reasonable percentage of manifest psychotic episodes could be precipitated by depressive episodes in individuals otherwise only moderately vulnerable to psychosis. If this is the case, it might help to explain some of the "depression" seen in the course of schizophrenia, including, notably, the depression-like symptoms observed early in the course of some psychotic decompensations. This hypothesis is consistent with the observation that dysphoria is associated more with positive symptoms than with negative symptoms in schizophrenia (79, 80). It might also help explain the finding that maintenance antidepressant treatment, administered for the purpose of averting depressive relapses in schizophrenic patients with histories of syndromal postpsychotic depression, also seemed to reduce their rate of psychotic exacerbations (81).
The advent of so-called atypical antipsychotic agents may be ushering in a new or "third" era with regard to dealing with schizophrenia. The first era was the preneuroleptic epoch, during which many detailed descriptions contributed to our understanding of the natural history of schizophrenia. After neuroleptic agents were introduced in the 1950s, it quickly became apparent that a remarkable curtailing of many dramatic psychotic manifestations was possible, even if many negative or cognitive symptoms persisted. Subsequently, almost all patients were treated with these agents, both acutely and in routine maintenance. A new and at least somewhat milder condition resulted: neuroleptic-treated schizophrenia. It was during this second era that most studies of "depression in schizophrenia" were undertaken, which form the basis of our current recognition and understanding of this condition.
There are several reasons to suspect that schizophrenia treated with atypical antipsychotics may prove to be at least a somewhat different condition than schizophrenia treated with conventional neuroleptics from the point of view of depression, although this, of course, still needs to be confirmed by appropriate careful investigations. First, atypical antipsychotic agents have a greatly reduced extrapyramidal side effect profile (82–86). Since akinesia and akathisia figure prominently in the differential diagnosis of what appears as depression in schizophrenia, this issue alone could be responsible for a notably different expression of depression in schizophrenia. Second, since atypical agents seem to rely much less exclusively on dopaminergic blockade for their therapeutic activity (83, 87–90), they might circumvent the mechanism of neuroleptic-induced dysphoria that could contribute to the depression syndrome. Third, atypical antipsychotics have frequently been reported to be superior to standard neuroleptics in the treatment of negative symptoms (83–86, 91–96), which can sometimes appear similar to depression.
Either through effects on negative symptoms or by some other means, atypical antipsychotics may lead to superior outcomes for patients with schizophrenia, as suggested by quality-of-life measures (97, 98). In this event, both acute and chronic reactions to disappointment or stress may be reduced. It is also possible that at least some atypical antipsychotic agents (so far demonstrated clearly only for clozapine in the treatment of refractory schizophrenia) may be superior to standard neuroleptics in the treatment of psychosis itself (93, 99). If this is the case, depression associated with the prodrome of psychotic relapse might possibly be reduced. Another mechanism whereby psychotic relapse may be averted through atypical antipsychotic use is by means of better medication compliance (95, 100–102), perhaps on the basis of a more favorable side effect profile (83, 85, 86, 103). Last, it is possible that atypical antipsychotic agents have direct antidepressant activity on their own. t1 reviews the controlled studies of atypical antipsychotic agents that make reference to the issue of depression. Several of these suggest a benefit for atypical antipsychotic agents in this regard. It is relevant to note, in this context, that conventional neuroleptics themselves have been found to have at least some degree of antidepressant activity (42, 107, 115–119), perhaps particularly in patients with agitated depression. Indeed, it may be features of agitation or excitement within the depression syndrome that may be particularly helped by atypical antipsychotics as well (120, 121), although studies have not been specifically designed to tease this out definitively. In addition to the studies cited in t1, a number of anecdotal reports also suggest that the atypical antipsychotics have superior antidepressant properties or, in the case of clozapine, antisuicidal properties when compared with conventional neuroleptics (87, 122–129), with but a single study suggesting the contrary (105).
Obviously, combinations of any or all of these effects could potentially be responsible for more favorable depression profiles in patients treated with atypical antipsychotics versus conventional neuroleptics. Further research is needed to clarify the shape and size of such an effect and to probe its possible mechanisms. Nevertheless, as summarized in t1, the controlled observations to date are intriguing with regard to potentially reduced rates of depression-like morbidity among schizophrenic patients receiving atypical antipsychotics.
Thirteen of the studies in t1 either are, or are reanalyses of, prospective randomized trials. Twelve of the 13 present evidence, of various strengths, that olanzapine, risperidone, or ziprasidone may have an antidepressant spectrum of activity in patients with schizophrenia or schizoaffective disorder. No published study could be found that tested the activity of quetiapine in this regard. One trial (105), the smallest of the group, presented data suggesting that risperidone may be inferior to haloperidol in terms of BPRS anxiety/depression factor improvement. Whether this result has to do with the relatively high dose range of risperidone employed (as much as 20 mg/day) or the lower dose range of haloperidol (as low as 2 mg/day) compared to the other studies is unclear. Altogether, eight of the studies in t1 used haloperidol as a comparison medication. The possibility, however, that extrapyramidal side effects of haloperidol created a spurious impression of superior antidepressant activity for the atypical antipsychotics in these studies is lessened by the sophisticated path-analysis examination in the largest of the studies (95, 96, 107). One study suggested that olanzapine was superior to risperidone in alleviating depression (71), and two studies found ziprasidone to be superior to placebo (83, 109). The final two studies in t1 present evidence suggesting that suicidality may be reduced in patients receiving clozapine. Suicidality is not a direct measure of depression, and other factors such as psychotic disorganization or terror can also be associated with self-destructive behavior in this population. However, it is not unreasonable to speculate that suicidality can still serve as a logical, although imperfect, proxy for depression in schizophrenia, especially since it represents such an important negative outcome.
The broad array of affinities for receptor sites attributable to the novel atypical antipsychotic agents (including a wide array of 5-hydroxytryptamine [5-HT], dopamine [other than D2], and muscarinic sites as well as α1-noradrenergic and histamine-1 receptor sites) suggest a variety of potential mechanisms through which atypical antipsychotics might exercise antidepressant effects (107, 130). Interest in both depression (131) and schizophrenia (132) has focused on the 5-HT2 binding site, but it would be premature to conclude that this is the key or exclusive site of action, and more research is needed. Useful clues in this area may also be gleaned from the existing experience involving the use of tricyclic antidepressants as adjunctive agents in the treatment of depression in schizophrenia (14). An extended consideration of mechanistic biochemical issues concerning depression, schizophrenia, and atypical antipsychotic agents goes beyond the scope of this article. Suffice it to say that there are many differences in receptor level activity among the different atypical antipsychotic agents that may prove to be relevant to their eventually being found to have differing clinical antidepressant profiles. They therefore should not necessarily be considered to be a homogeneous group of agents nor should the effect—or lack of effect—found with one agent necessarily be assumed to apply to other atypical agents.
A rational approach to treating depression in schizophrenia flows from considering the differential diagnosis. Excluding organic etiologies, the first consideration concerning a newly emergent depressive reaction in schizophrenia is whether it is a transient reaction to disappointment or stress or the prodrome of a new psychotic episode. The most prudent initial response is to increase surveillance and provide additional nonspecific support. A transient reaction to disappointment or stress will soon resolve spontaneously, and an incipient psychotic episode will soon declare itself. In the latter case, the increased surveillance will allow the best chance for the new episode to be detected promptly and "nipped in the bud" by appropriate interventions with antipsychotic medication.
If an episode of depression persists in a patient treated with a conventional neuroleptic, the question next arises whether the neuroleptic medication is responsible for the depression-like symptoms, either as an extrapyramidal side effect (akinesia or akathisia) or as a form of direct neuroleptic-induced dysphoria. There are three ways that such a situation could be approached: 1) neuroleptic dose reduction, if there is leeway to accomplish this safely; 2) introduction or upward titration of an antiparkinsonian medication (likely to be useful for akinesia but less likely for akathisia), a benzodiazepine, or β blocker (the latter two being likely to be useful for akathisia ); or 3) substitution of an atypical antipsychotic for the conventional neuroleptic.
If the episode of depression persists in a patient already being treated with an atypical antipsychotic, the existing literature gives less guidance. Again, dose reduction is a possibility if there is leeway for this, especially if the antipsychotic agent is risperidone, which has at least a modest degree of parkinsonian side effects in the higher dose range (88). Antiparkinsonian medication is another option, especially in conjunction with risperidone for the same reason. Anticholinergic antiparkinsonian medication may be an interesting option as well, since it is possible that it has its own antidepressant activity (27, 134) or anti-negative-symptom action (135). However, an anticholinergic antiparkinsonian agent would not be likely to be a felicitous choice in a patient receiving clozapine, since the combined anticholinergic activity might lead to too great a risk of autonomic side effects. Substituting one atypical antipsychotic agent for another is an additional possibility.
In antipsychotic-treated schizophrenic patients who are not flagrantly psychotic, persisting episodes of depression that do not respond to antiparkinsonian agents may respond to the addition of an adjunctive antidepressant medication (4, 14, 136–139). Most of the literature supportive of this intervention has focused on conventional neuroleptic-treated outpatients receiving adjunctive tricyclic antidepressants (4, 6, 14). The study that was most positive involved full syndromal depression criteria and continued antiparkinsonian medication throughout the antidepressant trial (136). It is certainly plausible, however, that selective serotonin reuptake inhibitor (SSRI) antidepressants would be useful as well, and some early results support this (140, 141). Adjunctive monoamine oxidase inhibitors (MAOIs) might also be useful for depression in schizophrenia, although the literature is sparse (142). It is of further interest that there have been some encouraging results involving adjunctive SSRIs (140, 141, 143–147), MAOIs (141, 148–150), and trazodone (151) in double-blind trials treating negative symptoms in schizophrenia. No prospective randomized study has yet been published, however, involving an adjunctive antidepressant added to an atypical antipsychotic agent in depressed patients with schizophrenia. Also of interest is that, although it may well be useful, ECT has not been specifically studied in the treatment of depression in patients with schizophrenia.
Additionally, lithium may be a useful adjunct in at least some cases of depression in schizophrenia, although definitive trials have not been published. Most reports concerning the use of lithium in schizophrenia have focused on its acute use during psychotic exacerbations rather than its extended use during maintenance-phase treatment (138, 152). The indicators most frequently cited in the literature as being of favorable prognostic significance for lithium response in schizophrenia have been excitement, overactivity, and euphoria. Nevertheless, depressive symptoms have occasionally been suggested as features indicative of a favorable adjunctive lithium response in schizophrenia (153). Other favorable prognosticators mentioned for lithium response in schizophrenia are the presence of previous affective episodes, a family history of affective illness, and an overall episodic nature to the clinical course (154).
Finally, although the focus of treatment discussed in this article has been primarily pharmacologic, any discussion of treatment strategies for depression in schizophrenia is incomplete without consideration of psychosocial interventions. Appropriately controlled studies of psychosocial interventions are, of course, notoriously complex and have not been carried out in a specifically depressed schizophrenic population. But this does not mean that they are not useful. There is little doubt that appropriately applied psychosocial strategies, such as stress reduction, psychoeducation, skill-building, problem-solving techniques, and family interventions designed to minimize excessive expressed emotion can be helpful (155). Strategies that foster hope, confidence, and self-esteem and interventions that contribute to real-life success experiences may be quite beneficial as well.
In summary, depression is well known to occur during the course of schizophrenia in many patients and contributes substantially to the morbidity—even mortality—of this disorder. "Depression" in schizophrenia, however, is heterogeneous, and the best approaches to its understanding and treatment are based on an appropriate differential diagnosis. Additional investigations will be necessary to fully appreciate the place of depressive symptoms in the conceptualization of schizophrenia and the best uses of contemporary therapeutic options, including atypical antipsychotic agents.
Received March 8, 1999; revisions received Oct. 26, 1999, and Feb. 15, 2000; accepted Feb. 25, 2000. From the Hillside Hospital Division of the North Shore-Long Island Jewish Health System. Address reprint requests to Dr. Siris, Kaufmann Building, Hillside Hospital, 75-59 263rd St., Glen Oaks, NY 11004; firstname.lastname@example.org (e-mail). The author thanks Pfizer Inc for its support in this project.