Michael A. Posternak

Untreated short-term course of major depression: a meta-analysis of outcomes from studies using wait-list control groups

INTRODUCTION While the short-term response rates to antidepressant medication and placebo are well established, very little is known about the short-term course of untreated depression. Knowledge of the course of untreated depression can serve as a benchmark for assessing the true benefits of active treatment. METHOD A meta-analysis was performed analyzing the outcomes of all psychotherapy studies that randomized adult outpatients with major depressive disorder to a wait-list control group. RESULTS Nineteen studies involving 221 subjects were reviewed. The mean decrease in Hamilton Rating Scale for Depression scores over 2-20 weeks was 11.9%, while the mean decrease in Beck Depression Inventory scores was 15.7%. Using a subsample of studies that reported individual outcomes, we estimated that 15 of 76 subjects (19.7%) improved to a degree comparable to what would be considered a positive response in antidepressant trials. CONCLUSIONS In the short-term, depressive symptomatology can be expected to decrease by about 10-15% on average without treatment. As many as 20% of subjects who participate in a short-term antidepressant trial may experience a spontaneous remission.

A review of studies of the Hamilton depression rating scale in healthy controls: implications for the definition of remission in treatment studies of depression.

The Hamilton Rating Scale for Depression (HRSD) is the most commonly used symptom severity scale to evaluate the efficacy of antidepressant treatment. On the basis of an expert consensus panel, an HRSD score of ≤7 was recommended as a cutoff to define remission. Since that recommendation, little empirical work has been conducted to confirm the validity of this threshold. One approach toward determining a cutoff score for defining remission is to establish the range of values for healthy controls. We therefore conducted a literature review of studies of the HRSD in healthy controls to determine the normal range of values. Studies of the HRSD in healthy control groups were identified in two ways. First, a MEDLINE search for the years 1966 to 2002 was conducted using the key words Hamilton, depression, and controls, and articles were reviewed. Second, the 69 studies included in two review articles written by the authors were examined. We identified 27 studies that included data on the HRSD for 1014 healthy controls. Across all studies, the weighted mean (SD) HRSD score, adjusting for sample size, was 3.2 (3.2; 95% CI, 3.0 to 3.4). HRSD scores were similar in geriatric and nongeriatric samples, and in men and women. Because HRSD scores in healthy controls are more likely to follow a skewed than a normal distribution, based on a mean of 3.2 and a SD of 3.2, at least 84% of healthy controls scored 7 or less on the HRSD, and 97.5% scored 10 or less. Thus, these results can be taken as support for the recommended cutoff of 7 on the HRSD to define remission. The results can also be used for normative comparisons in which posttreatment group mean scores are compared with mean scores from normative samples.

Empirical testing of two models for staging antidepressant treatment resistance.

Background: An increasing amount of attention has been paid to treatment resistant depression. Although it is quite common to observe nonremission to not just one but consecutive antidepressant treatments during a major depressive episode, a relationship between the likelihood of achieving remission and ones degree of resistance is not clearly known at this time. This study was undertaken to empirically test 2 recent models for staging treatment resistance. Materials and Methods: Psychiatrists from 2 academic sites reviewed charts of patients on their caseloads. Clinical Global Impressions-Severity (CGI-S) and Clinical Global Impressions-Improvement (CGI-I) scales were used to measure severity of depression and response to treatment, and 2 treatment-resistant staging scores were classified for each patient using the Massachusetts General Hospital staging method (MGH-S) and the Thase and Rush staging method (TR-S). Results: Out of the 115 patient records reviewed, 58 (49.6%) patients remitted at some point during treatment. There was a significant positive correlation between the 2 staging scores, and logistic regression results indicated that greater MGH-S scores, but not TR-S scores, predicted nonremission. Conclusions: This study suggests that the hierarchical manner in which the field has typically gauged levels of treatment resistance may not be strongly supported by empirical evidence. This study suggests that the MGH staging model may offer some advantages over the staging method by Thase and Rush, as it generates a continuous score that considers both number of trials and intensity/optimization of each trial.

The Naturalistic Course of Unipolar Major Depression in the Absence of Somatic Therapy

The goal of the study was to describe the naturalistic course of unipolar major depression in subjects not receiving somatic therapy for their depressive illness. Affectively ill individuals were recruited into the Collaborative Depression Study and followed prospectively for up to 15 years. One hundred thirty subjects who recovered from their intake episode of major depression subsequently experienced a recurrence that went untreated for at least 4 weeks following onset of the recurrence. The duration of the recurrent episode was examined using survival analytic techniques. Of the 130 subjects, 46 obtained somatic therapy at some time during the course of their depressive illness, while 84 subjects received no somatic therapy throughout their entire depressive episode. Survival analysis, which accounts for these 46 individuals by censoring their episodes at the time treatment was obtained, yielded a median time to recovery of 23 weeks. In the subsample of 84 subjects whose depressive illness went untreated from its inception through its resolution, the median time to recovery was 13 weeks. These results suggest that there is a high rate of recovery in individuals not receiving somatic treatment of their depressive illness, particularly in the first 3 months of an episode. Because treatment-seeking behavior is known to be associated with a worse prognosis, 23 weeks probably represents a lower-limit approximation of the median duration of an untreated depressive episode.

A review of studies of the Montgomery-Asberg Depression Rating Scale in controls: implications for the definition of remission in treatment studies of depression.

The Montgomery–Asberg Depression Rating Scale (MADRS) is one of the most commonly used symptom severity scales to evaluate the efficacy of antidepressant treatment. Various cut-offs have been employed in antidepressant efficacy trials to define remission, although little empirical work has been carried out to determine the validity of various thresholds. One approach towards deriving a valid cut-off score for defining remission is to determine whether a patients level of symptoms falls within the normal range of values after treatment. We therefore conducted a literature review of studies of the MADRS in healthy controls to determine the normal range of values. We identified 10 studies of 14 samples that included data on the MADRS for 569 controls. Across all studies, the mean (±SD) weighted MADRS score, adjusting for sample size, was 4.0 (5.8) (95% confidence interval 3.5–4.5). These results are consistent with the findings of our study of the validity of different cut-offs to define remission on the MADRS—based on a narrow definition of remission, which required a complete absence of clinically significant symptoms of depression, the optimal MADRS cut-off was ≤4 whereas based on a broader definition, the optimal cut-off was ≤9. The findings can be used for normative comparisons in which post-treatment group mean scores are compared to mean scores from normative samples. A limitation of the review is that none of the studies was based on a randomly selected sample from the general population. In addition, the rigor of the screening used to exclude individuals with psychopathology in most studies is unknown; thus, some of the controls may have had diagnosable depression, thereby elevating the mean scores in the presumptively healthy control group.

A Brief Assessment of Psychosocial Functioning of Subjects with Bipolar I Disorder: The LIFE-RIFT

Those afflicted with bipolar disorder often suffer from substantial functional impairment both when in episode and when in remission. This study examined the psychometric properties of a brief assessment of psychosocial functioning, the Range of Impaired Functioning Tool (LIFE-RIFT), among subjects with bipolar I disorder. The study sample consisted of 163 subjects who presented with bipolar I disorder at intake into the NIMH Collaborative Depression Study (CDS). All LIFE-RIFT items come from the Longitudinal Interval Follow-up Evaluation (LIFE). Follow-up data that were used to examine the reliability and validity of the scale come from assessments of psychosocial functioning that were conducted 6, 12, 18, and 24 months after intake into the CDS. The results of factor analyses indicate that the scale items are measures of one construct, psychosocial functioning. The interrater agreement on the scale score was very good with an intraclass correlation coefficient was 0.94. The internal consistency reliability among the scale items was uniformly satisfactory over the four assessment periods, with coefficient alpha ranging from 0.78 to 0.84. Mixed-effect regression analyses showed that during mood episodes subjects were significantly more impaired than those in recovery. In conclusion, the psychometric properties of the LIFE-RIFT were examined in subjects with bipolar I disorder. The analyses from this longitudinal, observational study provide empirical support for the reliability and validity of the scale. The LIFE-RIFT provides a brief, inexpensive alternative to scales currently used to assess psychosocial functioning and can be easily added to semistructured assessments that are used in clinical and treatment outcome studies.

Is the cutoff to define remission on the Hamilton Rating Scale for Depression too high

The Hamilton Rating Scale for Depression (HRSD) is the most frequently used measure of outcome in antidepressant efficacy trials. More than a decade ago, a consensus panel recommended that remission be defined on the 17-item version of the HRSD as a cutoff ≤7. This recommendation was accompanied by a call for research to validate this cutoff value; however, little research in this area has been performed since then. The goal of the present study was to compare the validity of different HRSD cutoffs for defining remission. Three hundred three depressed psychiatric outpatients were rated on the 17-item HRSD, the Global Assessment of Functioning (GAF) scale, and the Standardized Clinical Outcome Rating for Depression, an index of DSM-IV remission status. We examined the sensitivity, specificity, and overall classification rate of the HRSD for identifying a broad and narrow interpretation of the DSM-IV definition of remission, and the association between the breadth of the definition of remission on the HRSD and self-report ratings of global psychosocial functioning and quality of life. Based on a narrow definition of DSM-IV remission, which requires an absence of clinically significant symptoms of depression, the optimal 17-item HRSD cutoff was ≤2. Compared with patients scoring 3 through 7 on the HRSD, those who scored 0 to 2 reported significantly less psychosocial impairment and better quality of life. Our results thus support the use of a lower cutoff on the HRSD than has been traditionally used to define remission.

Exclusion criteria used in antidepressant efficacy trials: consistency across studies and representativeness of samples included.

The inclusion and exclusion criteria used to select subjects for participation in antidepressant efficacy trials (AETs) vary from study to study. It is unknown how much impact different sets of exclusion criteria have on the representativeness of subjects treated in AETs. In the present study, we applied the inclusion and exclusion criteria used in 39 recently published AETs to patients evaluated in routine clinical practice to evaluate the range and extent of the representativeness of samples treated in AETs. Nearly 600 patients with DSM-IV major depressive disorder (MDD) or bipolar depression underwent a thorough diagnostic evaluation. Inclusion and exclusion criteria used in AETs were applied to determine how many patients from our sample would have qualified for each AET had they applied. Approximately one sixth of the 596 depressed patients would have been excluded from an efficacy trial because they had a bipolar or psychotic subtype of depression. In the remaining 503 outpatients with nonpsychotic, unipolar MDD, the rates of exclusion ranged from 0% to 95.0% (mean = 65.8%). Thus, the findings suggest that there is much variability in the generalizability of AETs, although, in general, subjects treated in AETs represent only a minority of patients treated for MDD in a community-based psychiatry outpatient practice.

Symptoms of atypical depression

Studies examining the demographic and clinical features of depressed patients who meet criteria for the atypical features subtype have often yielded conflicting results. The present study sought to evaluate the demographic and clinical correlates associated with each of the five symptoms (mood reactivity, hypersomnia, hyperphagia, leaden paralysis and rejection sensitivity) that constitute the DSM-IV criteria set of atypical depression. Symptom prevalence rates were determined for 661 psychiatric outpatients diagnosed with a major depressive disorder, and were analyzed as a function of age, sex, severity, and episode duration. We found that: (1) younger age was positively associated with hypersomnia and negatively associated with leaden paralysis, while middle age was positively associated with both hyperphagia and rejection sensitivity; (2) female sex was associated with all of the atypical symptoms except rejection sensitivity; (3) a greater severity of illness was positively associated with leaden paralysis and rejection sensitivity, and negatively associated with mood reactivity; and (4) a duration of illness of greater than 3 months was positively associated with hyperphagia, leaden paralysis, and rejection sensitivity. Thus, the five atypical features do not appear to be associated with the same clinical profiles.

Symptom severity and exclusion from antidepressant efficacy trials.

It is the standard practice in antidepressant efficacy trials (AETs) to exclude potential participants with major depressive disorder (MDD) who score below a threshold on the Hamilton Rating Scale for Depression (HAM-D). It is unknown to what extent various cutoff scores impact on the generalizability of these trials. In the present report, we sought to determine how many patients with MDD presenting to an outpatient practice would fail to qualify for an AET because their symptoms were not sufficiently severe, and to what extent the variability in HAM-D cutoff scores impacts exclusion rates. Fifteen hundred individuals presenting for an intake at a psychiatric outpatient practice underwent an evaluation with semistructured diagnostic interviews. Five hundred and three patients received a principal diagnosis of nonbipolar, nonpsychotic MDD. Thirty-nine AETs published in five leading journals during the past 7 years were reviewed, 36 of which required a minimum score on the HAM-D for inclusion. We applied the HAM-D cutoffs used in these AETs to the 503 depressed patients to determine how many would qualify for each AET. Based on the least and most restrictive cutoff scores, between 11.3% and 71.0% of the depressed patients from our practice had an insufficient HAM-D score to qualify for an AET. The two most commonly used cutoff scores would lead to the exclusion of almost half of our sample. AETs tend to include the subset of depressed individuals with moderate to severe MDD and exclude a significant proportion of depressed patients who have mild MDD. The implications of these findings are discussed.