More pernicious course of bipolar disorder in the United States than in many European countries: implications for policy and treatment.

BACKGROUND: There is some controversy but growing evidence that childhood onset bipolar disorder may be more prevalent and run a more difficult course in the United States than some European countries. METHODS: We update and synthesize course of illness data from more than 960 outpatients with bipolar disorder (average age 40) from 4 sites in the U.S. and 3 sites in Netherlands and Germany. After giving informed consent, patients reported on parental history, childhood and lifetime stressors, comorbidities, and illness characteristics. RESULTS: Almost all aspects of bipolar disorder were more adverse in patients from the US compared with Europe, including a significantly higher prevalence of: bipolar disorder in one parent and a mood disorder in both parents; childhood verbal, physical, or sexual abuse; stressors in the year prior to illness onset and the last episode; childhood onsets of bipolar illness; delay to first treatment; anxiety disorder, substance abuse, and medical comorbidity; mood episodes and rapid cycling; and nonresponse to prospective naturalistic treatment. LIMITATIONS: Selection bias in the recruit of patients cannot be ruled out, but convergent data in the literature suggest that this does not account for the findings. Potential mechanisms for the early onset and more adverse course in the U.S. have not been adequately delineated and require further investigation. CONCLUSIONS: The data suggest the need for earlier and more effective long-term treatment intervention in an attempt to ameliorate this adverse course and its associated heavy burden of psychiatric and medical morbidity.

Increased parental history of bipolar disorder in the United States: association with early age of onset.

OBJECTIVE: Early-onset bipolar (BP) disorder and other poor prognosis characteristics are more prevalent in patients from the United States than from the Netherlands and Germany (abbreviated as Europe). We explored the impact of parental loading for affective illness on onset and other characteristics of BP disorder. METHOD: Parental history for unipolar (UP) and bipolar (BP) depression and course of illness characteristics were obtained from self-report in adults (average age 42) with BP disorder. Illness characteristics were examined by χ2 and multinomial logistic regression in relationship to the degree of parental loading: i) both parents negative; ii) one UP disorder; iii) one with BP disorder; and iv) both affected. RESULTS: After controlling for many poor prognosis factors, compared with those from Europe, patients from the United States had more iii) one parent with BP disorder and iv) both parents affected. An early age of onset of BP disorder was independently associated with this increased parental loading for affective disorder. CONCLUSION: Parental history of BP disorder and both parents with a mood disorder were more common in the United States than Europe and were associated with an early onset of bipolar disorder and other poor prognosis characteristics. These findings deserve replication and exploration of the potential mechanisms involved and their therapeutic implications.

Tranylcypromine vs. lamotrigine in the treatment of refractory bipolar depression: a failed but clinically useful study.

OBJECTIVE: To compare the efficacy and tolerability of tranylcypromine vs. lamotrigine in bipolar depression not responding to conventional antidepressants. METHOD: Bipolar depressed patients received open randomized treatment with tranylcypromine or lamotrigine as add-on to a mood stabilizer during 10 weeks. In a second treatment phase, non-responding patients could receive the opposite drug. Outcome criteria were response (measured with CGI-BP and IDS-C), switch into mania, and completion of the study. RESULTS: Only 20 of 70 planned patients were randomized, due to problems with recruitment, and 19 patients received any medication. During the first treatment phase 5/8 patients (62.5%) responded to tranylcypromine without switch into mania, compared with 4/11 patients (36.4%) on lamotrigine with two switches (statistically not significant). Over both treatment phases, 8/10 patients (80%) receiving tranylcypromine completed the study vs. 5/13 (38.5%) on lamotrigine (likelihood 0.02). CONCLUSION: There still appears to be a role for tranylcypromine in the treatment of refractory bipolar depression. Larger controlled studies are demanded.

The role of psychosocial stress in the onset and progression of bipolar disorder and its comorbidities: the need for earlier and alternative modes of therapeutic intervention.

Psychosocial stress plays an important role at multiple junctures in the onset and course of bipolar disorder. Childhood adversity may be a risk factor for vulnerability to early onset illness, and an array of stressors may be relevant not only to the onset, recurrence, and progression of affective episodes, but the highly prevalent substance abuse comorbidities as well. A substantial group of controlled studies indicate that various cognitive behavioral psychotherapies and psychoeducational approaches may yield better outcomes in bipolar disorder than treatment as usual. Yet these approaches do not appear to be frequently or systematically employed in clinical practice, and this may contribute to the considerable residual morbidity and mortality associated with conventional treatment. Possible practical approaches to reducing this deficit (in an illness that is already underdiagnosed and undertreated even with routine medications) are offered. Without the mobilization of new clinical and public health approaches to earlier and more effective treatment and supportive interventions, bipolar illness will continue to have grave implications for many patients' long-term well being.

Mood switch in bipolar depression: comparison of adjunctive venlafaxine, bupropion and sertraline.

BACKGROUND: Few studies have examined the relative risks of switching into hypomania or mania associated with second-generation antidepressant drugs in bipolar depression. AIMS: To examine the relative acute effects of bupropion, sertraline and venlafaxine as adjuncts to mood stabilisers. METHOD: In a 10-week trial, participants receiving out-patient treatment for bipolar disorder (stratified for rapid cycling) were randomly treated with a flexible dose of one of the antidepressants, or their respective matching placebos, as adjuncts to mood stabilisers. RESULTS: A total of 174 adults with bipolar disorder I, II or not otherwise specified, currently in the depressed phase, were included. All three antidepressants were associated with a similar range of acute response (49-53%) and remission (34-41%). There was a significantly increased risk of switches into hypomania or mania in participants treated with venlafaxine compared with bupropion or sertraline. CONCLUSIONS: More caution appears indicated in the use of venlafaxine rather than bupropion or sertraline in the adjunctive treatment of bipolar depression, especially if there is a prior history of rapid cycling.

Use of quetiapine in bipolar disorder: a case series with prospective evaluation.

Quetiapine, a new atypical antipsychotic, was added to ongoing treatment of bipolar I outpatients (n=15) for symptoms of illness (mood lability, irritability, psychosis and/or difficulty sleeping). All evaluations were prospectively obtained, with the majority of patients (n=9) showing much or very much improvement on the Clinical Global Impression for Bipolar Disorder (CGI-BP). Somatic complaints were limited. Mean (SD) duration before changes in medication regimens was 134 (100) days. Studies of the use of quetiapine in maintenance treatment of bipolar disorder are warranted.

Relationship between prior course of illness and neuroanatomic structures in bipolar disorder: a preliminary study.

OBJECTIVE: In this preliminary study, we examined the relationships between prior course and severity of illness and size of the hippocampus, temporal lobes, and third and lateral ventricles in patients with bipolar disorder. BACKGROUND: The few studies that have investigated relationships between course of illness measures and neuroanatomic structures in patients with bipolar disorder found divergent results. METHOD: Twenty-six outpatients, who met Diagnostic and Statistical Manual, Third Edition - Revised (DSM-III-R) criteria for bipolar disorder, received a magnetic resonance imaging (MRI) scan, from which volumes of the temporal lobes, hippocampi, third ventricle, and areas of the lateral ventricles were calculated. Prior course of illness variables were determined using the NIMH Life-Chart Method and were correlated to the volumetric measures of neuroanatomic structures using multiple regression analyses. RESULTS: A longer duration of illness was paradoxically associated with a larger left temporal lobe volume whether patients with a history of substance abuse were removed from the analyses. CONCLUSIONS: Additional studies are needed to both replicate and further examine the association of prior course of illness and larger hippocampal and ventricular volumes in bipolar disorder.

Developmental vulnerabilities to the onset and course of bipolar disorder.

Different types of psychosocial stressors have long been recognized as potential precipitants of both unipolar and bipolar affective episodes and the causative agents in posttraumatic stress disorder (PTSD). New preclinical data have revealed some of the neurobiological mechanisms that could convey the long-term behavioral and biochemical consequences of early stressors. Depending on the timing, quality, quantity, and degree of repetition, maternal deprivation stress in the neonatal rodent can be associated with lifelong anxiety-like behaviors, increases in stress hormones and peptides. and proneness to drug and alcohol administration, in association with acute changes in the rate of neurogenesis and apoptosis (preprogrammed cell death) and decrements in neurotrophic factors and signal transduction enzymes necessary for learning and memory. Patients with bipolar illness who have a history of early extreme adversity (physical or sexual abuse in childhood or adolescence), compared with those without, show an earlier onset of illness, faster cycling frequencies, increased suicidality, more Axis I and Axis II comorbidities (including alcohol and substance abuse), and more time ill in more than 2 years of prospective follow-up. These findings are subject to a variety of interpretations, but to the extent that the more severe course of bipolar illness characteristics are directly and causally related to these early stressful experiences, early recognition and treatment of high-risk children could be crucial in helping to prevent or ameliorate the long-term adverse consequences of these stressors.

The Stanley Foundation Bipolar Network. I. Rationale and methods.

BACKGROUND: The Stanley Foundation Bipolar Network (SFBN) was created to address the paucity of help studies in bipolar illness. AIMS: To describe the rationale and methods of the SFBN. METHOD: The SFBN includes five core sites and a number of affiliated sites that have adopted consistent methodology for continuous longitudinal monitoring of patients. Open and controlled studies are performed as patients' symptomatology dictates. RESULTS: The reliability of SFBN raters and the validity of the rating instruments have been established. More than 500 patients are in continuous daily longitudinal follow-up. More than 125 have been randomised to one of three of the newer antidepressants (bupropion, sertraline and venlafaxine) as adjuncts in a study of mood stabilizers and 93 to omega-3 fatty acids. A number of open clinical case series have been published. CONCLUSIONS: Well-characterised patients are followed in a detailed continuous longitudinal fashion in both opportunistic case series and double-blind, randomised controlled trials with reliable and validated measures.

The Stanley Foundation Bipolar Network. 2. Preliminary summary of demographics, course of illness and response to novel treatments.

BACKGROUND: The Stanley Foundation Bipolar Network (SFBN) evaluates treatments, course and clinical and neurobiological markers of response in bipolar illness. AIMS: To give a preliminary summary of emerging findings in these areas. METHOD: Studies with established and potentially antimanic, antidepressant and mood-stabilising agents range from open case series to double-blind randomised clinical trials, and use the same core assessment methodology, thereby optimising the comparability of the outcomes. The National Institute of Mental Health Life Chart Method is the core instrument for retrospective and prospective longitudinal illness description. RESULTS: The first groups of patients enrolled show a considerable degree of past and present symptomatology, psychiatric comorbidity and functional impairment. There are associations of both genetic and early environmental factors with more severe courses of illness. Open case series with add-on olanzapine, lamotrigine, gabapentin or topiramate show a differential spectrum of effectiveness in refractory patients. CONCLUSIONS: The SFBN provides important new data for the understanding and treatment of bipolar disorder.

Axis I psychiatric comorbidity and its relationship to historical illness variables in 288 patients with bipolar disorder.

OBJECTIVE: Bipolar disorder often co-occurs with other axis I disorders, but little is known about the relationships between the clinical features of bipolar illness and these comorbid conditions. Therefore, the authors assessed comorbid lifetime and current axis I disorders in 288 patients with bipolar disorder and the relationships of these comorbid disorders to selected demographic and historical illness variables. METHOD: They evaluated 288 outpatients with bipolar I or II disorder, using structured diagnostic interviews and clinician-administered and self-rated questionnaires to determine the diagnosis of bipolar disorder, comorbid axis I disorder diagnoses, and demographic and historical illness characteristics. RESULTS: One hundred eighty-seven (65%) of the patients with bipolar disorder also met DSM-IV criteria for at least one comorbid lifetime axis I disorder. More patients had comorbid anxiety disorders (N=78, 42%) and substance use disorders (N=78, 42%) than had eating disorders (N=9, 5%). There were no differences in comorbidity between patients with bipolar I and bipolar II disorder. Both lifetime axis I comorbidity and current axis I comorbidity were associated with earlier age at onset of affective symptoms and syndromal bipolar disorder. Current axis I comorbidity was associated with a history of development of both cycle acceleration and more severe episodes over time. CONCLUSIONS: Patients with bipolar disorder often have comorbid anxiety, substance use, and, to a lesser extent, eating disorders. Moreover, axis I comorbidity, especially current comorbidity, may be associated with an earlier age at onset and worsening course of bipolar illness. Further research into the prognostic and treatment response implications of axis I comorbidity in bipolar disorder is important and is in progress.

Rate of switch in bipolar patients prospectively treated with second-generation antidepressants as augmentation to mood stabilizers.

INTRODUCTION: Bipolar patients with breakthrough major depressive episodes despite ongoing adequately-dosed mood stabilizer medication were randomized in a double-blind manner to one of three antidepressants with different mechanisms of action: bupropion, sertraline, or venlafaxine. Preliminary data are presented on the switch rates into hypomania or mania for the antidepressants as a group prior to unblinding the specific individual drug efficacy and tolerability data in this ongoing clinical trial. METHODS: Subjects included 64 bipolar patients who participated at five sites in a 10-week double-blind trial for depression and a 1-year blinded continuation maintenance phase for responders. Nonresponders were re-randomized such that there were 95 acute treatment phases. In the acute phase, doses were titrated to clinical response, side effects, or maximum dose of bupropion (450 mg/day), sertraline (200 mg/day), or venlafaxine (375 mg/day). Daily ratings on the National Institute of Mental Health-Life Chart Methodology (NIMH-LCM) were inspected for the degree of improvement on the Clinical Global Impressions scale as revised for bipolar illness (CGI-BP) and the occurrence of hypomania or mania. RESULTS: Thirty-five (37%) of the 95 acute treatment phases were associated with a much or very much improved rating in depression on the CGI-BP. Thirteen (14%) of these 95 acute trials of antidepressants as adjuncts to mood stabilizers were associated with switches, seven into hypomania and six into mania. Forty-two patients elected to go into the continuation phase in 48 instances. Sixteen (33%) of the continuation phase trials were associated with mood switches, 10 into hypomania and six into mania. CONCLUSIONS: In this randomized double-blind prospective study of three second-generation antidepressants (bupropion, sertraline, and venlafaxine) in bipolar patients whose depression broke through ongoing treatment with mood stabilizers, switches into hypomania or mania occurred in 14% of the acute phases and 33% of the continuation phases. Individual data on each drug will be assessed in the next phase of the study after more subjects are recruited and the blind is broken.

The Stanley Foundation Bipolar Treatment Outcome Network. I. Longitudinal methodology.

The NIMH-Stanley Foundation Bipolar Treatment Outcome Network, a multisite clinical trials network, has been established to address many of the neglected areas of research in bipolar illness. The Network was designed so that it would be able to conduct randomized clinical trials at several different levels of methodologic rigor (blinded and open-label) both in academic and community practice settings in order to better assess long-term efficacy of existing treatments and develop new ones. In this fashion, large numbers of representative patients with bipolar disorder have been enrolled with an additional focus of elucidating possible clinical and biological predictors of treatment response. The unique focus of the Network is its systematic longitudinal approach to illness so that patients can be assessed comprehensively over the long-term in sequential randomized clinical trials at critical clinical decision points where data on relative efficacy are inadequate. Bipolar I and bipolar II patients with a range of illness variants and comorbidities are included. Daily prospective ratings of severity of mania and depression and associated degree of functional impairment are completed on the NIMH-Life Chart Method and a modified Clinical Global Impressions Scale for Bipolar Illness (CGI-BP) is utilized. More detailed cross-sectional ratings for depression (Inventory of Depressive Symptomatology), mania (Young Mania Rating Scale), and psychosis (Positive and Negative Syndrome Scale) are additionally used at academic centers. This article describes the rationale for the Network, its guiding principles, methods, and study design to systematically assess the highly variable course of bipolar illness and its response to current and future treatments.

The Stanley Foundation Bipolar Treatment Outcome Network. II. Demographics and illness characteristics of the first 261 patients.

BACKGROUND: Since recent NIMH Bipolar Disorder Workshops highlighted the dearth of longitudinal and controlled studies of bipolar illness, the Stanley Foundation Bipolar Network (SFBN) has recruited a large cohort of patients with bipolar disorder to begin to address these issues. This report describes the demographics and course of illness characteristics of this study population. METHODS: The first 261 outpatients to be diagnosed by the Structured Clinical Interview for DSM-IV (SCID) and complete a detailed patient and a brief clinician questionnaire are described. All patients met DSM-IV criteria for bipolar I (n=211), bipolar II (n=42), or NOS (n=5) or schizoaffective (n=3), bipolar type. Chi-square and t-tests were used to examine statistically significant associations among important demographic and descriptive items. RESULTS: The general demographic and illness characteristics were similar to those in many bipolar clinical samples and not dissimilar from those reported in epidemiological surveys. The majority of patients had been hospitalized, with almost half reporting a worsening of illness over time, and two-thirds were not asymptomatic between episodes. First treatment for patients had been delayed by an average of 10 years from illness onset (by SCID). Almost a third of patients had attempted suicide at least once, and 30% reported current suicidal ideation at study entry. A total of 62% reported moderate to severe impact of the illness on occupational functioning. Early onset bipolar illness (< or =17 years old) was associated with increased frequency of mood switches, worsening course of illness, and history of early abuse (physical, verbal, or sexual). CONCLUSION: The SFBN represents a sample of predominantly BP I patients largely recruited from the community who will be followed in detail longitudinally, participate in clinical trials, and thus help advance our understanding and treatment of this life-threatening medical disorder. While there is a broad range of illness characteristics and severity, the majority of patients have been severely impacted by their illness despite the availability of multiple conventional treatment approaches in the community. These data further underscore the need for development of new and earlier treatment interventions. LIMITATION: The SFBN population is limited by the lack of random selection and represents a cohort willing to be treated and followed intensively in academic tertiary referral centers. While its characteristics are similar to many clinical study populations, the generalizability to non-clinic populations remains uncertain.

A placebo-controlled study of lamotrigine and gabapentin monotherapy in refractory mood disorders.

There is a pressing need for additional treatment options for refractory mood disorders. This controlled comparative study evaluated the efficacy of lamotrigine (LTG) and gabapentin (GBP) monotherapy versus placebo (PLC). Thirty-one patients with refractory bipolar and unipolar mood disorders participated in a double-blind, randomized, crossover series of three 6-week monotherapy evaluations including LTG, GBP, and PLC. There was a standardized blinded titration to assess clinical efficacy or to determine the maximum tolerated daily dose (LTG 500 mg or GBP 4,800 mg). The primary outcome measure was the Clinical Global Impressions Scale (CGI) for Bipolar Illness as supplemented by other standard rating instruments. The mean doses at week 6 were 274 +/- 128 mg for LTG and 3,987 +/- 856 mg for GBP. Response rates (CGI ratings of much or very much improved) were the following: LTG, 52% (16/31); GBP, 26% (8/31); and PLC, 23% (7/31) (Cochran's Q = 6.952, df = 2, N = 31, p = 0.031). Post hoc Q differences (df = 1, N = 31) were the following: LTG versus GBP (Qdiff = 5.33, p = 0.011); LTG versus PLC (Qdiff = 4.76, p = 0.022); and GBP versus PLC (Qdiff = 0.08, p = 0.70). With respect to anticonvulsant dose and gender, there was no difference between the responders and the nonresponders. The agents were generally well tolerated. This controlled investigation preliminarily suggests the efficacy of LTG in treatment-refractory affectively ill patients. Further definition of responsive subtypes and the role of these medications in the treatment of mood disorders requires additional study.

Validation of the prospective NIMH-Life-Chart Method (NIMH-LCM-p) for longitudinal assessment of bipolar illness.

BACKGROUND: Systematic and accurate depiction of a patient's course of illness is crucial for assessing the efficacy of maintenance treatments for bipolar disorder. This need to rate the long-term prospective course of illness led to the development of the National Institute of Mental Health prospective Life Chart Methodology (NIMH-LCM-p or LCM). The NIMH-LCM-p allows for the daily assessment of mood and episode severity based on the degree of mood associated functional impairment. We have previously presented preliminary evidence of the reliability and validity of the LCM, and its utility in clinical trials. This study is a further and more extensive validation of the clinician rated NIMH-LCM-p. METHODS: Subjects included 270 bipolar patients from the five sites participating in the Stanley Foundation Bipolar Network. Daily prospective LCM ratings on the clinician form were initiated upon entry, in addition to at least monthly ratings with the Inventory of Depressive Symptomatology-clinician rated (IDS-C), the Young Mania Rating Scale (YMRS) and the Global Assessment of Functioning (GAF). We correlated appropriate measures and time domains of the LCM with the IDS-C, YMRS and GAF. RESULTS: Severity of depression on the LCM and on the IDS-C were highly correlated in 270 patients (r = -0.785, P < 0.001). Similarly, a strong correlation was found between LCM mania and the YMRS (r = 0.656, P < 0.001) and between the LCM average severity of illness and the GAF (r = -0.732, P < 0.001). CONCLUSIONS: These data further demonstrate the validity and potential utility of the NIMH-LCM-p for the detailed daily longitudinal assessment of manic and depressive severity and course, and response to treatment.

Neuroleptic exposure in bipolar outpatients in a research setting.

The study purpose was to determine the extent of neuroleptic exposure in bipolar outpatients maintained on mood-stabilizing medications and any clinical correlates associated with this exposure. Data on medication and severity of illness were gathered from the records (prospective and retrospective) of 70 bipolar patients involved in outpatient research studies at the National Institute of Mental Health (NIMH). The percentage of patients requiring neuroleptic treatment, percentage of time on neuroleptics during the period of observation, total dose of neuroleptics in chlorpromazine (CPZ) equivalency, and number of neuroleptic trials were among the variables calculated. Regression analyses and analyses of variance (ANOVAs) were performed to assess the relationships between neuroleptic exposure and clinical course. Forty-five patients (64.3%) had a neuroleptic trial during the prospective study. Subjects exposed to neuroleptics spent, on average, 15.4% (median, 6.0%) of the time in study on neuroleptic treatment, and were administered, on average, a total of 11,770.5 mg (median, 1,621.9 mg) of neuroleptics (in CPZ equivalency) per year in the prospective study. As expected, bipolar I compared with bipolar II patients had significantly higher neuroleptic exposure by a number of measures. The number of hospitalizations for mania prior to study entry was associated with greater prospective neuroleptic use during the study. Despite maintenance treatment with one or more moodstabilizing agents, we found a relatively high need for adjunctive neuroleptic medication even in this sample of high-functioning bipolar outpatients. These results highlight the need for the study of alternatives, as well as more effective primary mood-stabilizing agents.

Open-label adjunctive topiramate in the treatment of bipolar disorders.

BACKGROUND: To preliminarily explore the spectrum of effectiveness and tolerability of the new antiepileptic drug topiramate in bipolar disorder, we evaluated the response of 56 bipolar outpatients in the Stanley Foundation Bipolar Outcome Network (SFBN) who had been treated with adjunctive topiramate in an open-label, naturalistic fashion. METHODS: In this case series, response to topiramate was assessed every 2 weeks for the first 3 months according to standard ratings in the SFBN, and monthly thereafter while patients remained on topiramate. Patients' weights, body mass indices (BMIs), and side effects were also assessed. RESULTS: Of the 54 patients who completed at least 2 weeks of open-label, add-on topiramate treatment, 30 had manic, mixed, or cycling symptoms, 11 had depressed symptoms, and 13 were relatively euthymic at the time topiramate was begun. Patients who had been initially treated for manic symptoms displayed significant reductions in standard ratings scores after 4 weeks, after 10 weeks, and at the last evaluation. Those patients who were initially depressed or treated while euthymic showed no significant changes. Patients as a group displayed significant decreases in weight and BMI from topiramate initiation to week 4, to week 10, and to the last evaluation. The most common adverse side effects were neurologic and gastrointestinal. CONCLUSIONS: These preliminary open observations of adjunctive topiramate treatment suggest that it may have antimanic or anticycling effects in some patients with bipolar disorder, and may be associated with appetite suppression and weight loss that is often viewed as beneficial by the patient and clinician. Controlled studies of topiramate's acute and long-term efficacy and side effects in bipolar disorder appear warranted.

The increasing use of polypharmacotherapy for refractory mood disorders: 22 years of study.

BACKGROUND: Few studies have approached the subject of polypharmacotherapy systematically. This retrospective review of 178 patients with refractory bipolar disorder or unipolar depression (Research Diagnostic Criteria or DSM-III-R criteria) discharged from the National Institute of Mental Health (NIMH) Biological Psychiatry Branch between 1974 and 1996 was conducted to assess the degree and efficacy of "add-on" pharmacotherapy. METHOD: Following completion of formal structured blinded research protocols, patients entered a treatment phase (often again on a blind basis) in which all agents available in the community could be utilized. Each patient's retrospective life chart and all prospective double-blind nurse- and self-rated NIMH data were reviewed. The overall degree of improvement at discharge was assessed by rating on the Clinical Global Impressions scale (CGI) as modified for bipolar illness (CGI-BP). RESULTS: A 78% improvement rate (moderate or marked on the CGI) was achieved at the time of discharge. There was a significant relationship between number of medications utilized at discharge as a function of discharge date (r = 0.45, p < .0001). The percentages of patients discharged on treatment with 3 or more medications were 3.3% (1974-1979), 9.3% (1980-1984), 34.9% (1985-1989), and 43.8% (1990-1995). No correlation was found between polypharmacy and age (r = -0.03, p = .66). Patients more recently discharged from the NIMH had an earlier age at illness onset, more lifetime weeks depressed, and a higher rate of rapid cycling than patients in the earlier cohorts. CONCLUSION: Increasing numbers of medications in more recent NIMH cohorts were required to achieve the same degree of improvement at hospital discharge. More systematic approaches to the complex regimens required for treatment of patients with refractory mood disorder are clearly needed.