Predictors of improved mood over time in clinical trials for major depression

The objective of this exploratory analysis was to use a repeated measures modeling approach to identify potential predictors of improved mood over time in patients with major depression. Fifty-one subjects with major depressive disorder (MDD) were enrolled in a 1-week single blind placebo lead-in, followed by an 8-week, double-blind placebo-controlled treatment with either fluoxetine or venlafaxine. Hierarchical linear regression models were used to identify baseline and placebo lead-in predictors of change in repeated measures of the Hamilton Depression Rating Scale (HDRS) during treatment. Non-specific predictors of improved mood included decreased prefrontal activity during placebo lead-in as measured by quantitative electroencephalographic cordance, lower pretreatment depressed mood on the HDRS, shorter duration of current episode, increased verbalization of suicidal thoughts, no family history of mood disorders, less severe middle insomnia, higher guilt, lower somatic anxiety, and younger age. Moderators of improved mood included somatization, paranoid ideation, and self-reported depressed mood. We also found that change in prefrontal cordance after randomization mediated the effects of middle insomnia, suicidal thoughts, and family history of mood disorders. We recommend the use of repeated measures modeling, and the exploration of relationships among biological and psychological factors, for future analyses of clinical trial data.     

Placebo response and antidepressant response.

OBJECTIVE: Much of the response to an antidepressant is the result of placebo response. The placebo response embedded in drug response confounds studies seeking to identify brain mechanisms essential for pharmacologic response. Exclusion of patients who fail to meet entry criteria at the end of a placebo lead-in phase has been inadequate to reduce the effect of placebo during pharmacologic trials. This study focused on the placebo lead-in phase and examines whether change in severity of depression during placebo lead-in predicts change in depressive symptoms during antidepressant treatment. METHOD: The subjects were patients aged 60-85 years with nonpsychotic unipolar major depression not attributed to dementing disorders, medical illnesses, or drugs causing depression and had a 24-item Hamilton Depression Rating Scale score of 18 or greater. After a two-week placebo lead-in, subjects with Hamilton Depression Rating Scale score of 18 or greater received 10 mg escitalopram for 12 weeks. RESULTS: Worsening or limited change in depression during placebo treatment predicted improvement in depressive symptoms during escitalopram treatment. Limited change in anxiety, melancholia, helplessness, and paranoia during placebo treatment were the strongest predictors of improvement in depression while on escitalopram. CONCLUSIONS: These findings, if replicated, may be used to characterize depressed older patients likely to respond to the pharmacologic action of antidepressants rather than the placebo response embedded in drug trials and thus improve the methodology of biomarker studies of antidepressant response. On a clinical level, depressed older patients who improve during a prolonged evaluation may be candidates for nonpharmacologic treatments because their drug response may be limited.     

Transdermal selegiline in major depression: a double-blind,placebo-controlled,parallel-group study in outpatients

OBJECTIVE: The authors investigated the efficacy and safety of transdermal selegiline in adult outpatients with major depressive disorder. METHOD: Following a 1-week placebo lead-in, 177 adult outpatients with major depressive disorder were randomly assigned to receive transdermal selegiline (20 mg applied once daily by means of a 20-cm(2) patch) (N=89) or placebo (N=88) for 6 weeks. The patients followed a tyramine-restricted diet during the medication trial and for 2 weeks after completion of treatment. Response to medication or placebo was measured by using the 17-item and 28-item versions of the Hamilton Depression Rating Scale, the Montgomery-Asberg Depression Rating Scale, and the Clinical Global Impression (CGI) severity and improvement measures. RESULTS: Greater improvement was observed after 6 weeks in patients treated with transdermal selegiline than in those given placebo according to all measures. A statistically significant difference between drug and placebo was seen in Hamilton depression scale and Montgomery-Asberg Depression Rating Scale scores as early as week 1 of treatment. There were no differences in the adverse event profile of the patients given selegiline and those given placebo with the exception of application-site reactions, which were more common with the selegiline transdermal system. No orthostatic hypotensive or hypertensive reactions were observed. CONCLUSIONS: Transdermal selegiline (20 mg applied once daily by means of a 20-cm(2) patch) administered for 6 weeks was an effective and well-tolerated treatment for adult outpatients with major depression. The typical side effects commonly seen with traditional monoamine oxidase inhibitor antidepressants were not observed.     

Changes in brain function of depressed subjects during treatment with placebo.

OBJECTIVE: It has been proposed that 50%-75% of the efficacy of antidepressant medication represents the placebo effect, since many depressed patients improve when treated with either medication or placebo. This study examined brain function in depressed subjects receiving either active medication or placebo and sought to determine whether quantitative electroencephalography (QEEG) could detect differences in brain function between medication and placebo responders. Both QEEG power and cordance, a new measure that reflects cerebral perfusion and is sensitive to the effect of antidepressant medication, were examined. METHOD: Fifty-one subjects with major depression were enrolled in one of two independent, 9-week double-blind, placebo-controlled studies in which either fluoxetine (N=24) or venlafaxine (N=27) was the active medication. Serial QEEG recordings were performed during the course of treatment. After 9 weeks, the blind was broken and subjects were classified as medication responders, placebo responders, medication nonresponders, or placebo nonresponders. RESULTS: No significant pretreatment differences in clinical or QEEG measures were found among the four outcome groups. Placebo responders, however, showed a significant increase in prefrontal cordance starting early in treatment that was not seen in medication responders (who showed decreased cordance) or in medication nonresponders or placebo nonresponders (who showed no significant change). There was no significant change in QEEG power during treatment. CONCLUSIONS: These findings suggest that "effective" placebo treatment induces changes in brain function that are distinct from those associated with antidepressant medication. If these results are confirmed, cordance may be useful for differentiating between medication and placebo responders.     

The impact of medical comorbidity on acute treatment in major depressive disorder

OBJECTIVE: The authors investigated the impact of medical comorbidity on the acute phase of antidepressant treatment in subjects with major depressive disorder. METHOD: A total of 384 outpatients meeting DSM-III-R criteria for major depressive disorder enrolled in 8-week open treatment with fluoxetine, 20 mg/day. The authors used the Cumulative Illness Rating Scale to measure the burden of medical comorbidity and the 17-item Hamilton Rating Scale for Depression to assess changes in depressive symptoms. The outcome measures were response to treatment (>/=50% reduction in score) and clinical remission (score 相似文献   

Changes in QEEG prefrontal cordance as a predictor of response to antidepressants in patients with treatment resistant depressive disorder: a pilot study

INTRODUCTION: Previous studies of patients with unipolar depression have shown that early decreases of EEG cordance (a new quantitative EEG method) can predict clinical response. We examined whether early QEEG decrease represents a phenomenon associated with response to treatment with different antidepressants in patients with treatment resistant depression. METHOD: The subjects were 17 inpatients with treatment resistant depression. EEG data and response to treatment were monitored at baseline and after 1 and 4 weeks on an antidepressant treatment. QEEG cordance was computed at three frontal electrodes in theta frequency band. The prefrontal cordance combines complementary information from absolute and relative power of EEG spectra. Recent studies have shown that cordance correlates with cortical perfusion. Depressive symptoms were assessed using Montgomery-Asberg Depression Rating Scale (MADRS). RESULTS: All 17 patients completed the 4-week study. All five responders showed decreases in prefrontal cordance after the first week of treatment. Only 2 of the 12 nonresponders showed early prefrontal cordance decrease. The decrease of prefrontal QEEG cordance after week 1 in responders as well as the increase in nonresponders were both statistically significant (p-value 0.03 and 0.01, respectively) and the changes of prefrontal cordance values were different between both groups (p-value 0.001). CONCLUSION: Our results suggest that decrease in prefrontal cordance may indicate early changes of prefrontal activity in responders to antidepressants. QEEG cordance may become a useful tool in the prediction of response to antidepressants.     

No gender differences in placebo responses of patients with major depressive disorder.

BACKGROUND: This study was designed to compare placebo responses in men and women. METHODS: Data for 501 women and 375 men with major depressive disorder treated with placebo from seven investigational randomized double-blind trials comparing fluoxetine with placebo were analyzed. Changes in major depressive disorder symptoms with placebo administration were measured as changes in total Hamilton Depression Rating Scale scores and adverse (nocebo) effects were measured by comparing treatment-emergent signs and symptoms. RESULTS: Both women and men with major depressive disorder showed significant symptomatic improvement following placebo administration, similar in magnitude and time course of response. Women on placebo reported slightly more nocebo effects than men. CONCLUSIONS: The finding that women and men with major depressive disorder demonstrated a similar therapeutic outcome after placebo administration suggests that gender is not a predictor of placebo response.     

Efficacy of typical and atypical antipsychotics for primary and comorbid anxiety symptoms or disorders: a review

OBJECTIVE: The efficacy of antipsychotics in the treatment of primary or comorbid anxiety disorders or anxiety symptoms in major depressive disorder or bipolar disorder was reviewed. DATA SOURCES: English-language literature cited in MEDLINE from January 1, 1968, to December 31, 2005, was searched with the keywords anxiety disorder, anxiety symptoms, generalized anxiety disorder, panic disorder, obsessive-compulsive disorder, posttraumatic stress disorder, social phobia, bipolar disorder, major depressive disorder, Hamilton Rating Scale for Anxiety, antipsychotics, typical antipsychotics, atypical antipsychotics, fluphenazine, haloperidol, perphenazine, pimozide, thiothixene, trifluoperazine, loxapine, molindone, chlorpromazine, mesoridazine, thioridazine, fluspirilene, penfluridol, pipothiazine, flupenthixol, clozapine, olanzapine, risperidone, quetiapine, ziprasidone, aripiprazole, amisulpride, and clinical trial. Randomized, double-blind, placebo-controlled trials and open-label studies with a minimum of 20 subjects with a DSM-III/IV or ICD-10 diagnosis of anxiety disorder and studies without a DSM-III/IV or ICD-10 diagnosis of anxiety disorder but with Hamilton Rating Scale for Anxiety (HAM-A) scores as an outcome were prioritized. Studies on bipolar disorder or major depressive disorder with the analysis of changes in anxiety symptoms were reviewed. Early studies on neurosis/ anxiety or anxious depression without a HAM-A component were also reviewed. DATA SYNTHESIS: Six trials in primary generalized anxiety disorder (GAD), 15 in refractory obsessive-compulsive disorder (OCD), 8 in posttraumatic stress disorder (PTSD), 6 in neurosis with the HAM-A, 1 in social phobia, and 2 in anxiety symptoms in bipolar depression were identified. Low doses of trifluoperazine were superior to placebo in the treatment of GAD. Most of the less well-designed studies showed that other typical antipsychotics might be superior to placebo or as effective as benzodiazepines in the treatment of GAD and other anxiety conditions. In most studies, risperidone, olanzapine, and quetiapine augmentation to antidepressants was superior to placebo in treating refractory OCD and PTSD. Both olanzapine and quetiapine significantly reduced anxiety compared to placebo in studies of bipolar depression. CONCLUSION: Except for trifluoperazine, there is no large, well-designed study of antipsychotics in the treatment of primary or comorbid anxiety symptoms or disorders. The efficacy of these agents in various anxiety conditions needs to be further investigated with large, well-designed comparison studies.     

Changes in prefrontal activity characterize clinical response in SSRI nonresponders: a pilot study

Previous studies in unipolar depression have shown that early decreases in prefrontal values of the QEEG cordance measure identified responders to pharmacotherapy. These studies have all examined individuals who were drug-free prior to the first physiologic assessment, yet in the clinical management of treatment resistant depression (TRD), many patients undergo changes in treatment without a drug-free interval between treatments. Here, we investigated whether cordance decreases were associated with response in Stage I TRD subjects without wash-out between treatment trials. Awake EEGs were recorded from 12 adults with unipolar depression. Subjects were receiving naturalistic treatment, had failed SSRI monotherapy, and were starting a new treatment prescribed by their treating psychiatrists. EEG data were recorded before starting the new treatment and after approximately 1 week. Six of the 12 subjects responded to treatment after 8--10 weeks. Five of the six responders showed an early cordance decreases, compared with two of the six nonresponders (accurate characterization in 75% of the cases). Consistent with previous treatment trials, decreases in prefrontal cordance differentiated responders from nonresponders in this setting as well. These findings suggest that cordance biomarkers may be a useful tool in effectiveness trials that parallel clinical practices in SSRI nonresponders, and may not require a wash-out period between treatments.     

A randomized trial of citalopram versus placebo in outpatients with asthma and major depressive disorder: a proof of concept study.

BACKGROUND: The prevalence of asthma has increased in recent years and depression is common in this population. Minimal data are available on the treatment of depressed asthma patients. METHODS: Ninety adults with asthma and current major depressive disorder were randomized to receive citalopram or placebo for 12 weeks. At each visit, the Hamilton Rating Scale for Depression (HRSD), Inventory of Depressive Symptomatology - Self-Report, Asthma Control Questionnaire, and Asthma Quality of Life Questionnaire were administered, and oral corticosteroid use assessed. RESULTS: In the evaluable sample (n = 82), the primary outcome, a random regression analysis of HRSD scores, revealed no significant between-group differences. Bonferroni corrected secondary outcomes revealed HRSD scores decreased significantly in both groups with a significantly greater decrease in the citalopram group at week 6. Changes in asthma symptoms were similar between groups. The groups had similar rates of oral corticosteroid use at baseline, but the citalopram group had less corticosteroid use during the study. Changes in asthma symptom severity correlated with changes in depressive symptom severity. CONCLUSIONS: A reduction in depressive symptoms was associated with improvement in asthma. Corticosteroid use, an important measure of severe asthma exacerbations, was lower in the citalopram group. Larger clinical trials in this population are warranted.     

Achieving remission from depression with venlafaxine and venlafaxine extended release: a literature review of comparative studies with selective serotonin reuptake inhibitors

OBJECTIVE: To evaluate data supporting the ability of venlafaxine, an antidepressant with a dual mechanism of action, to produce remission from depression. METHOD: Review of multicentre, double-blind, randomized studies comparing venlafaxine or venlafaxine extended release (XR) with a selective serotonin reuptake inhibitor (SSRI), using Hamilton Depression Rating Scale total scores in the range of < or = 7 and < 10 as the final outcome measure, to evaluate the ability of venlafaxine/venlafaxine XR to produce full remission from depression. RESULTS: Venlafaxine/venlafaxine XR demonstrated higher rates of remission than did the SSRIs and placebo. CONCLUSION: With full remission rather than response as the measure of outcome, venlafaxine/venlafaxine XR demonstrated more robust antidepressant efficacy than the SSRIs and placebo. This finding suggests that venlafaxine/venlafaxine XR are appropriate standard-of-care therapies for the treatment of patients with major depressive disorder.     

Treatment of atypical depression with cognitive therapy or phenelzine: a double-blind, placebo-controlled trial

BACKGROUND: Patients with atypical depression are more likely to respond to monoamine oxidase inhibitors than to tricyclic antidepressants. They are frequently offered psychotherapy in the absence of controlled tests. There are no prospective, randomized, controlled trials, to our knowledge, of psychotherapy for atypical depression or of cognitive therapy compared with a monoamine oxidase inhibitor. Since there is only 1 placebo-controlled trial of cognitive therapy, this trial fills a gap in the literature on psychotherapy for depression. METHODS: Outpatients with DSM-III-R major depressive disorder and atypical features (N = 108) were treated in a 10-week, double-blind, randomized, controlled trial comparing acute-phase cognitive therapy or clinical management plus either phenelzine sulfate or placebo. Atypical features were defined as reactive mood plus at least 2 additional symptoms: hypersomnia, hyperphagia, leaden paralysis, or lifetime sensitivity to rejection. RESULTS: With the use of an intention-to-treat strategy, the response rates (21-item Hamilton Rating Scale for Depression score, < or =9) were significantly greater after cognitive therapy (58%) and phenelzine (58%) than after pill placebo (28%). Phenelzine and cognitive therapy also reduced symptoms significantly more than placebo according to contrasts after a repeated-measures analysis of covariance and random regression with the use of the blind evaluator's final Hamilton Rating Scale for Depression score. The scores between cognitive therapy and phenelzine did not differ significantly. Supplemental analyses of other symptom severity measures confirm the finding. CONCLUSIONS: Cognitive therapy may offer an effective alternative to standard acute-phase treatment with a monoamine oxidase inhibitor for outpatients with major depressive disorder and atypical features.     

Efficacy of desipramine in depressed outpatients. Response according to research diagnosis criteria diagnoses and severity of illness

The efficacy of desipramine for mild depression was tested in a double-blind, placebo-controlled study of outpatients with scores below 19 on the Hamilton Rating Scale for Depression (HAM-D). Of 103 such patients, 23 dropped out and 16 improved during a ten-day placebo period. Among 64 patients completing the randomized portion of the study, significantly more improved with desipramine that with placebo. The Research Diagnostic Criteria (RDC) category of major depressive disorder largely accounted for the drug-placebo response difference found for the entire sample. Patients with intermittent depressive disorder improved significantly less frequently with desipramine than patients with major depressive disorder. Independent of RDC diagnosis, severity of illness correlated with outcome. Thus, patients with pretreatment HAM-D scores at or above the median demonstrated significant drug effect, while patients with lower pretreatment HAM-D scores did not.     

Subject expectations of treatment effectiveness and outcome of treatment with an experimental antidepressant

OBJECTIVE: To evaluate the association between treatment expectations and response in a 9-week, single-blind experimental antidepressant treatment study. METHOD: Twenty-five adult subjects meeting DSM-IV criteria for major depressive disorder with Hamilton Rating Scale for Depression (HAM-D) scores of >/= 17 completed a treatment trial using the experimental antidepressant reboxetine. Following a 1-week placebo lead-in, subjects received single-blind treatment for 8 weeks with reboxetine 8 to 10 mg/day. During the screening visit, subjects were asked to self-rate their expectations of the effectiveness of the study medication. Forced-choice responses were "not at all effective," "somewhat effective," or "very effective." Response to treatment was defined as a final HAM-D score of 相似文献   

Efficacy of extended-release venlafaxine in nondepressed outpatients with generalized anxiety disorder

OBJECTIVE: This study evaluated the efficacy and safety of fixed doses of once-daily extended-release (XR) venlafaxine in outpatients with generalized anxiety disorder without concomitant major depressive disorder. METHOD: Adult outpatients with generalized anxiety disorder but not major depressive disorder with total scores of 18 or higher on the Hamilton Rating Scale for Anxiety and scores of 2 or higher on its anxious mood and tension factors were eligible. Patients were randomly assigned to receive placebo or venlafaxine XR (75, 150, or 225 mg/day) for 8 weeks. Primary efficacy variables were final total and psychic anxiety factor scores on the Hamilton anxiety scale and final severity and global improvement item scores on the Clinical Global Impression (CGI) scale. RESULTS: Of the 377 patients entering the study, 370 were included in a safety analysis and 349 in an efficacy analysis. Adjusted mean scores at 8 weeks (last-observation-carried-forward analysis) were significantly lower for one or more of the venlafaxine XR groups in four of four primary and three of four secondary outcome measures than for the placebo group. These included a change of 1.7 (versus 1.3) from baseline on CGI severity item scores and a final score of 2.2 (versus 2.6) on the CGI global improvement item. All doses of venlafaxine XR were well tolerated. CONCLUSIONS: Venlafaxine XR is an effective and well-tolerated option for the short-term treatment of generalized anxiety disorder in outpatients without major depressive disorder.     

Early reduction in prefrontal theta QEEG cordance value predicts response to venlafaxine treatment in patients with resistant depressive disorder

INTRODUCTION: Previous studies of patients with unipolar depression have shown that early decrease of prefrontal EEG cordance in theta band can predict clinical response to various antidepressants. We have now examined whether decrease of prefrontal quantitative EEG (QEEG) cordance value after 1 week of venlafaxine treatment predicts clinical response to venlafaxine in resistant patients. METHOD: We analyzed 25 inpatients who finished 4-week venlafaxine treatment. EEG data were monitored at baseline and after 1 week of treatment. QEEG cordance was computed at three frontal electrodes in theta frequency band. Depressive symptoms and clinical status were assessed using Montgomery-Asberg Depression Rating Scale (MADRS), Beck Depression Inventory-Short Form (BDI-S) and Clinical Global Impression (CGI). RESULTS: Eleven of 12 responders (reduction of MADRS >or=50%) and only 5 of 13 non-responders had decreased prefrontal QEEG cordance value after the first week of treatment (p=0.01). The decrease of prefrontal cordance after week 1 in responders was significant (p=0.03) and there was no significant change in non-responders. Positive and negative predictive values of cordance reduction for response were 0.7 and 0.9, respectively. CONCLUSION: The reduction of prefrontal theta QEEG cordance value after first week of treatment might be helpful in the prediction of response to venlafaxine.     

Morning light treatment hastens the antidepressant effect of citalopram: a placebo-controlled trial

BACKGROUND: Selective serotonin reuptake inhibitors are effective in approximately 70% of patients with a major depressive episode, but therapeutic changes usually require 2 weeks of administration to become clinically relevant. Adjunct light therapy has been proposed to hasten the effects of drug treatment. The purpose of the present study was to evaluate the effect of morning light therapy or placebo combined with citalopram in the treatment of patients affected by a major depressive episode without psychotic features. METHOD: Thirty inpatients (DSM-IV major depressive disorder [N = 21] and bipolar disorder [N = 9]) were treated with citalopram, 40 mg, and randomized in a 3:2 manner to receive 30 minutes of 400 lux green light treatment in the morning or placebo (exposure to a deactivated negative ion generator) during the first 2 weeks of drug treatment. Timing of light therapy was individually defined to obtain a 2-hour phase advance to morning light. Outcome was measured with the Hamilton Rating Scale for Depression and the Zung Self-Rating Depression Scale every week, and with a Visual Analogue Scale 3 times a day during the first week. RESULTS: All outcome measures showed significantly (p <.05) better mood improvement in light-treated patients, resulting in faster responses to antidepressant treatment. CONCLUSION: The combination of citalopram and light treatment was more effective than citalopram and placebo in the treatment of major depression. With an optimized timing of administration, low-intensity light treatment significantly hastened and potentiated the effect of citalopram, thus providing the clinical psychiatrists with an augmenting strategy that was found effective and devoid of side effects.     

A double-blind,randomized, placebo-controlled trial of augmentation with lamotrigine or placebo in patients concomitantly treated with fluoxetine for resistant major depressive episodes

BACKGROUND: Evidence of the antidepressant efficacy of lamotrigine is increasing, although there are no placebo-controlled trials of lamotrigine augmentation in depression. The aim of this study was to assess if augmentation with lamotrigine was superior to placebo in patients who were receiving fluoxetine for resistant major depressive episodes. METHOD: Twenty-three patients who had experienced at least 1 major depressive episode that was resistant to at least 1 prior trial of antidepressant therapy were selected. These patients were treated with fluoxetine, 20 mg/day, and concomitantly randomly assigned to receive either lamotrigine (N = 13) or placebo (N = 10) for 6 weeks. The dose of lamotrigine was titrated upward from 25 mg/day to 100 mg/day. Patients suffering from bipolar II disorder (N = 8) or from major depressive disorder (N = 15) (DSM-IV criteria) were enrolled, resulting in heterogeneity of the sample. The primary outcome measure was Hamilton Rating Scale for Depression score. Data were collected from 2000-2001. RESULTS: Lamotrigine was statistically superior to placebo on the Clinical Global Impressions scale at endpoint, both in absolute terms (mean +/- SD Clinical Global Impressions-Severity of Illness scores: lamotrigine, 2.15 +/- 1.28; placebo, 3.40 +/- 1.17; p =.0308) and using a responder analysis, with response defined as a Clinical Global Impressions-Improvement score of 2 or less (lamotrigine, 84.62% [N = 11]; placebo, 30.00% [N = 3]; p =.013). The effect of lamotrigine on Clinical Global Impressions scale scores was seen in both major depressive disorder and bipolar II disorder. Lamotrigine, however, failed to separate statistically from placebo on the Hamilton Rating Scale for Depression and Montgomery-Asberg Depression Rating Scale. This failure to differentiate on a primary outcome measure is essentially a negative study result. This result is most likely an artifact of the small sample size used and the resultant limited power of the study. CONCLUSION: The results of this trial add to the literature suggesting potential efficacy of the antidepressant profile of lamotrigine. In addition, this study points to a possible role of lamotrigine as an augmentation agent in depression.     

Cognitive reactivity to sad mood provocation and the prediction of depressive relapse

CONTEXT: Episode remission in unipolar major depression, while distinguished by minimal symptom burden, can also be a period of marked sensitivity to emotional stress as well as an increased risk of relapse. OBJECTIVE: To examine whether mood-linked changes in dysfunctional thinking predict relapse in recovered patients who were depressed. DESIGN: In phase 1 of this study, patients with major depressive disorder were randomly assigned to receive either antidepressant medication or cognitive behavior therapy. In phase 2, patients who achieved clinical remission underwent sad mood provocation and were then observed with regular clinical assessments for 18 months. SETTING: Outpatient psychiatric clinics at the Centre for Addiction and Mental Health, Toronto, Ontario. PARTICIPANTS: A total of 301 outpatients with major depressive disorder, aged 18 to 65 years, participated in phase 1 of this study and 99 outpatients with major depressive disorder in remission, aged 18 to 65 years, participated in phase 2. MAIN OUTCOME MEASURE: Occurrence of a relapse meeting DSM-IV criteria for a major depressive episode as assessed by the longitudinal interval follow-up evaluation and a Hamilton Depression Rating Scale score of 16 or greater. RESULTS: Patients who recovered through antidepressant medication showed greater cognitive reactivity following the mood provocation than those who received cognitive behavior therapy. Regardless of type of prior treatment, the magnitude of mood-linked cognitive reactivity was a significant predictor of relapse over the subsequent 18 months. Patients whose mood-linked endorsement of dysfunctional attitudes increased by a minimum of 8 points had a significantly shorter time to relapse than those whose scores were not as elevated. CONCLUSIONS: The vulnerability of remitted depressed patients for illness relapse may be related to the (re)activation of depressive thinking styles triggered by temporary dysphoric states. This is the first study to link such differences to prognosis following successful treatment for depression. Further understanding of factors predisposing to relapse/recurrence in recovered patients may help to shorten the potentially lifelong course of depression.