From Cornell University Medical College, New York, NY. Drs Klerman and Perry are deceased.
This randomized clinical trial compared 16-week interventions with interpersonal psychotherapy, cognitive behavioral therapy, supportive psychotherapy, and supportive psychotherapy with imipramine for human immunodeficiency virus (HIV)-positive patients with depressive symptoms.
Subjects (N=101; 85 male, 16 female) with known HIV seropositivity for at least 6 months were randomized to 16 weeks of treatment. Inclusion criteria were 24-item Hamilton Depression Rating Scale score of 15 or higher, clinical judgment of depression, and physical health sufficient to attend outpatient sessions. Therapists were trained in manualized therapies specific for HIV-positive patients. Treatment adherence was monitored.
Subjects randomized to interpersonal psychotherapy (n=24) and supportive psychotherapy with imipramine (n=26) had significantly greater improvement on depressive measures than those receiving supportive psychotherapy (n=24) or cognitive behavioral therapy (n=27). Similar results appeared in the completer subsample.
Depressive symptoms appear treatable in HIV-positive patients. Interpersonal psychotherapy may have particular advantages as a psychotherapy for patients who have experienced the significant life events of HIV infection.
IN 1988 the late Samuel Perry, MD, inaugurated a 4-cell treatment study for human immunodeficiency virus (HIV)-positive patients with depressive symptoms, modeled on the National Institute of Mental Health Treatment of Depression Collaborative Research Program (TDCRP) for medically healthy patients.1 Although depression prevalence among HIV-positive individuals was known to be elevated and response of HIV-positive patients to antidepressant medication was being established,2 little was known about the efficacy of psychotherapy for this population.3 Perry decided to compare interpersonal psychotherapy (IPT),4 cognitive behavioral therapy (CBT),5 supportive psychotherapy (SP), and supportive psychotherapy with imipramine (SWI). When the study began, HIV infection appeared more acutely fatal than it does today. Clinicians wondered whether antidepressant psychotherapies would help patients with seeming "reasons to be depressed."
Our study design diverged from the TDCRP in offering SP rather than pill placebo plus clinical management as its presumed least active treatment. Supportive psychotherapy was included to control for nonspecific effects of psychotherapy, and SWI as standard antidepressant medication treatment. We have reported preliminary data finding IPT (n=16) more efficacious than SP (n=16); however, symptoms diminished in both groups of patients.3 No other published trials have examined psychotherapy for HIV patients with depressive symptoms.
We hypothesized that focal antidepressant psychotherapies (IPT and CBT) would rival pharmacotherapeutic efficacy, and might perhaps be better accepted by HIV-positive patients who already took many pills and might not want or be able to tolerate additional medication. Supportive psychotherapy was designed to approximate a control condition, yet not be a minimal or empty treatment.3
Subjects were recruited through advertisement and referral. Advertisements offered HIV-positive individuals free treatment for depression. Inclusion criteria required known HIV-positive status of 6 months or more, a score of 15 or higher on the 24-item Hamilton Depression Rating Scale (Ham-D),6 and clinical judgment of significant depressive symptoms. Physical health had to permit outpatient treatment. Most subjects at entry fit Centers for Disease Control and Prevention stages II to III for HIV infection7 (relatively medically asymptomatic). Over time, however, increasingly seriously ill subjects were accepted. Exclusion criteria were significant non-HIV medical disease, schizophrenia, bipolar disorder, contraindication to imipramine, current substance abuse, significant cognitive impairment (Mini-Mental State Examination8 score <25), inability to speak English, and concurrent psychiatric treatment aside from HIV self-help or support groups.
Four IPT (2 psychiatrists, 2 social workers), 3 CBT (all PhD psychologists), 9 SP (2 psychiatrists, 4 social workers, 3 registered nurses), and 6 SWI (3 psychiatrists, 3 registered nurses) therapists each treated between 1 and 15 cases (mean ± SD, 5.6 ± 4.1). Experts in each modality trained therapists during a year-long phase, certifying them after at least 3 pilot cases. To avoid drift, therapists used treatment manuals and were monitored throughout with individual supervision of audiotaped sessions and blinded ratings of random sessions by independent, reliable adherence monitors.
Separate manuals, treatment teams, and team leaders encouraged uniform treatment and therapist morale.
Interpersonal psychotherapy helps patients relate changes in mood to events in their environment and consequent changes in social roles.4 Depression is defined as a medical illness. The therapist gives the patient a depressive diagnosis and the sick role, engages the patient on affectively laden current life issues, and frames the patient's difficulties within an interpersonal problem area: grief, role dispute, role transition, or interpersonal deficits. Strategies address these problem areas, focusing in the present on what the patient wants and what options exist to achieve this. The manual9 modified IPT to particular psychosocial concerns of depressed HIV-positive patients. Therapists told patients they had 2 medical illnesses: depression and HIV.
Cognitive behavioral therapy5 is another antidepressant treatment validated by numerous controlled trials. Cognitive behavioral therapists help patients identify irrational, negative thoughts associated with depression. Patients learn to record and examine these thoughts, weigh their evidence, and challenge rather than believing and acting on them. The therapist helps the patient test hypotheses based on negative cognitions with the aim of disproving them. Techniques include cognitive restructuring and refocusing and formal homework.
Defined in its manual as non-IPT and non-CBT, SP resembled the client-centered therapy of Rogers10 with added psychoeducation about depression and HIV. Supportive psychotherapy was inherently less structured than IPT or CBT, and unlike them did not offer patients a framework for therapy or focus on specific themes. Although possibly hampered by the proscription of interpersonal and cognitive techniques, SP was by no means nontreatment. Supportive therapists were empathic, skillful, and experienced.
The SWI cell added imipramine and a biochemical rationale11 to SP. Imipramine therapy was begun at 50 mg/d and increased as tolerated to 300 mg/d for 3 to 4 weeks unless limited by adverse effects.
Sixteen 50-minute IPT or CBT sessions were scheduled within a 17-week period. Both SP and SWI conditions ranged between 8 and 16 sessions, determined by patient need, of 30 to 50 minutes' duration. All sessions were audiotaped or videotaped.
Subjects were assessed on study entry using the Structured Clinical Interview for DSM-III-R, nonpatient version (SCID-NP12), the Personality Disorder Examination (PDE13), and demographic questionnaires. Outcome assessments were the 24-item Ham-D and Beck Depression Inventory (BDI14), to measure depressive symptomatology; CD4 cell count; and the clinician-administered, 100-point Karnofsky scale15 to assess physical functioning. Medication adverse effects were assessed by rater checklist. For comparison with other studies, we calculated Ham-D 17-item scores.
Raters monitored therapist adherence using a 104-item version, modified for SP, of the 96-item Collaborative Study Psychotherapy Rating Scale (CSPRS16) developed for the TDCRP. This scale differentiated the 4 therapies. Raters were 4 predoctoral psychology graduate students who developed reliability (intraclass correlation coefficient range, 0.26-0.90, generally 0.62-0.90) after approximately 40 hours of training. Training was conducted by J.C.M. and later by initial trainees and included rating and discussing up to 16 randomly chosen pilot tapes covering all interventions. Raters met for ongoing supervision to review tapes and prevent drift. All therapist-patient dyads were rated as adherent to treatment protocols based on CSPRS ratings of 2 randomly chosen tapes, 1 early (sessions 3-6) and 1 late session (9-12).17
Following telephone screening, subjects met with psychiatric nurses, who completed the Ham-D. Eligible subjects gave informed written consent for participation. Subjects were retested for HIV and discussed test results before treatment began. At intake interviewers administered clinical ratings, the SCID-NP, and the PDE. Subjects were randomly assigned to treatment in a balanced design using a computer-generated random number sequence sealed in individual envelopes.
Subjects completed BDIs before each session. At baseline, midpoint (week 8), and just before termination (week 15), independent raters repeated the Ham-D and Karnofsky scales. Therapists also rated the Ham-D at alternate sessions.
Outcomes were compared across the 4 treatments for intent-to-treat (N=101) and completer (n=69) samples. Intent-to-treat analyses compare the outcomes of all study entrants, including subjects who refused randomization (n=4) or received minimal treatment (≥1 but <4 sessions, n=15). Completer analyses test treatment efficacy only in the subsamples receiving full treatment dosage, and hence are less generalizable to treatment at large.
Demographic and diagnostic variables were compared using χ2 and t tests for discrete and continuous variables, respectively, with 2-tailed α=.05. Outcome was assessed using analysis of covariance (ANCOVA) for Ham-D and BDI, controlling for initial scores, and by repeated-measures analysis of variance (ANOVA). Clinician-rated Ham-D scores were used for the last observation in the intent-to-treat sample. Remission rates were analyzed using χ2 tests.
One hundred one subjects were randomized to treatment (IPT=24, CBT=27, SP=24, SWI=26). Table 1 illustrates no significant demographic differences across groups at baseline. Subjects were predominantly male (85%), gay or bisexual (80%), and white (58%). Mean ± SD age was 36.9 ± 6.9 years (range, 24-59 years). Eighty-three percent reported some college education. Eighty-four percent had known their HIV-positive status for 1 year or more; 71% for more than 2 years (n=9).
Most subjects were not severely medically ill at study entry. The sample shifted over time, reflecting the HIV epidemic, to greater proportions of heterosexual, ethnic minority, and medically sicker subjects. Baseline mean Karnofsky score was 80 ± 6.5; CD4 cell count, 280 ± 222 (n=99).
Although all subjects were judged to have clinically significant depressive symptoms, only 53% met DSM-III-R criteria for a current mood disorder (Table 1). Most subjects met lifetime mood disorder criteria. About half of the subjects met criteria for an Axis II disorder and for lifetime substance abuse history. Comorbidity did not statistically differ across treatments.
Interpersonal psychotherapy and CBT subjects attended nonsignificantly more sessions than SP or SWI subjects (IPT, 11.5 ± 5.7; CBT, 11.5 ± 6.5; SP, 9.0 ± 5.3; SWI, 9.9 ± 5.0). Subjects randomized to the SWI group (n=26) received a mean imipramine dosage of 210 ± 66 mg/d (n=23; range, 50-300 mg/d), with a mean imipramine/desipramine blood level of 177 ± 76 mg/dL (n=18; range, 54-339 mg/dL).
Subjects in the intent-to-treat samples did not differ across cells at baseline in depressive severity on the Ham-D or BDI. Scores on the Ham-D decreased significantly for all treatments by midpoint and at termination. The BDI scores fell significantly for IPT and SWI by midpoint, and for all cells by termination.
Across treatment groups, findings converged both for Ham-D and BDI, and for intent-to-treat and completer analyses. Table 2 and Table 3 and Figure 1 demonstrate a bifurcation in treatment outcomes. The IPT and SWI groups clustered with similarly low outcome scores; CBT and SP also moved together but showed less improvement. The intent-to-treat ANCOVA showed a main effect on Ham-D (ANCOVA F3,96=2.67; P=.05; n=101): IPT and SWI were each superior to CBT, with SP a distant but not statistically different third. On the BDI, SWI was superior to CBT and SP by ANCOVA, and IPT was superior to SP (ANCOVA F3,95=4.26; P=.007; n=100). Repeated-measures ANOVA found main effects for intervention and time, but no significant intervention by time interaction.
Outcome: Beck Depression Inventory. For interpersonal psychotherapy (IPT), n=24; for cognitive behavioral therapy (CBT), n=27; for supportive psychotherapy (SP), n=23; and for supportive psychotherapy with imipramine (SWI), n=26. Analysis of covariance F3,95=4.26; P=.007; IPT>SP; SWI>CBT and SP.
Among completers (Table 2; n=69), IPT was superior to both CBT and SP on Ham-D (ANCOVA F3,64=3.22; P<.03); SWI did not differ from other treatments. On BDI (Table 3), IPT was superior to CBT and SP, while SWI was superior to SP alone (ANCOVA F3,64=5.01; P=.003). In no analyses did IPT differ from SWI, or CBT from SP.
To ascertain whether medical status influenced antidepressant outcome, we repeated analyses using as covariates intake Karnofsky score and CD4 cell count in addition to initial Ham-D or BDI score. Findings were unchanged except among completers, where one additional difference emerged. In this analysis alone, intake CD4 cell counts (t=−2.1, P<.04) and Karnofsky severity (t=−2.2, P<.04) were associated with endpoint BDI score, and the post hoc Tukey least significant difference test showed a trend for CBT to outperform SP (P=.06). Analyses of the 17-item Ham-D using ANCOVA failed to show statistically significant differences for the intent-to-treat sample. Differences persisted for the completer subsample (F3,64=3.70; P=.02; n=69).
Remission from depression was assessed using the strict TDCRP criterion of 17-item Ham-D score of 6 or greater. Groups did not significantly differ, although remission rates followed already described trends: IPT, 46%; CBT, 30%; SP, 21%; and SWI, 50% in the intent-to-treat sample; and 59%, 35%, 30%, and 56%, respectively, among the completer sample. Remission rates using the 24-item Ham-D at a threshold score of 8 or greater were similar: IPT, 46%; CBT, 30%; SP, 17%; and SWI, 42% for the intent-to-treat sample; and 59%, 41%,24%, and 50%, respectively, for the completer sample.
Karnofsky scores rose with mood improvement (baseline, 80 ± 6.5; termination, 86 ± 8.3), as might be expected with alleviation of depression3 (Table 4). Whereas Karnofsky scores did not differ at baseline, the IPT and SWI groups had higher termination scores than the CBT and SP groups (ANCOVA F3,62=4.60; P=.006). The CD4 cell count did not change significantly by group over time.
Neither Axis I diagnosis nor Axis II personality clusters mediated outcome by intervention. Presence of personality disorder was nonsignificantly associated with elevated depressive severity (Cluster A: 5 points on the Ham-D and BDI; Cluster B: 0 points on the Ham-D, 1 point on the BDI; Cluster C: 3 points on the Ham-D, 1 point on the BDI). Initial Karnofsky score correlated with the final Ham-D (r=−.27, P=.009, n=94) and BDI (r=−.25, P=.02, n=95) scores, suggesting patients with greater physical dysfunction responded best; CD4 cell count was not significantly correlated. Medication adverse effects at midtreatment, whether measured by severity or total number, were not significantly correlated with outcome. Therapist experience and professional training also did not affect outcome.
This first study of individual antidepressant psychotherapy for HIV-positive patients reveals that they, like other medically ill patients with depressive symptoms,2,18 warrant and respond to specific antidepressant treatments. Depressive symptoms diminished across treatments. High prevalences of Axis II disorders and lifetime substance abuse did not preclude successful treatment of mood symptoms. Depressive symptoms should never be viewed as normal, even when HIV is involved: they deserve vigorous treatment. The improvement in Karnofsky functional scores, particularly in treatments of greatest antidepressant potency, suggests that patients had previously attributed to HIV symptoms for which depression was culpable.
Treatment with IPT and SWI yielded the best outcomes, while CBT and SP fared less well. Statistical significance of outcome analyses varied by treatment sample and assessment instrument considered, but results clearly converged. The differential treatment effects invite comparison with the TDCRP, in which post hoc analyses found imipramine and IPT each superior to placebo and clinical management among more severely depressed patients. The TDCRP remission rates (IPT, 43%; CBT, 36%; placebo and case management, 21%; and imipramine and case management, 42%; n=239)1 closely resembled ours. Psychotherapy adherence findings were also similar, and the 32% attrition rate was identical.17 The TDCRP post hoc analysis of more severe (17-item Ham-D score ≥20)1 vs milder depression was not undertaken because only 10% of subjects met the severity criterion.
This study, which is only the second direct comparison of IPT and CBT as antidepressant treatments, revealed differences favoring IPT in some analyses. Research finding differences between active psychotherapies has been rare indeed. Possible explanations for this include the quality of therapy and differences in specificity for the target disorder. Therapists across treatments were uniformly well-trained, empathic, technically adherent, and esteemed by their patients.
Interpersonal psychotherapy may have particular advantages over CBT for HIV-positive patients with depressive symptoms. Interpersonal psychotherapy connects life events to mood episodes (1) to help patients mourn life upheavals while (2) pragmatically and optimistically encouraging them to find new life goals and adjustments. Both halves of this formula seemed important to patients. Having suffered a surfeit of HIV-related life events—multiple bereavements, role disputes, and role transitions—they responded to IPT therapists' supportive encouragement to "live out your fantasies": to change their lives and seek whatever they desired for however much time remained to them. This seemed a tailored fit of therapy and patient. By contrast, CBT addresses patients' exaggeration of hopeless thoughts, a relatively disadvantageous stance in treating patients with objectively negative life events. Even with optimistic cognitive restructuring and refocusing, CBT may fit HIV patients' situations less well. We speculate that IPT may be indicated for depressive patients who have experienced recent distressing life events and are likely to experience more in the future. Patients who report few life events (the IPT "interpersonal deficits" category) may be better CBT candidates.19 Despite (nonsignificantly) fewer sessions, SP equaled CBT in efficacy, suggesting no advantage for CBT beyond so-called "nonspecific" psychotherapy effects. The efficacy of pharmacotherapy was again validated.2
This study has numerous limitations. Relative to the TDCRP, sample size was small and depression scores were lower, reducing statistical power to find treatment differences (particularly on the 17-item Ham-D). Forty-seven percent of subjects lacked current SCID mood disorders, although all were deemed clinically depressed by supervising psychiatrists. Most subjects had lifetime mood disorder histories, and many barely missed the SCID threshold for a current episode. The criterion of clinical impression of depression rather than SCID imitates clinical practice and may widen the generalizability of these findings.
Another design limitation, potentially less treatment exposure for SP, could have handicapped SP outcome. Finally, subjects were overwhelmingly male. Treatment of depressed HIV-positive women, who face different psychosocial and socioeconomic pressures than men, remains relatively unexplored.20
Psychotherapy of HIV-positive patients deserves greater attention than it has received. Interpersonal psychotherapy may be a particularly beneficial psychotherapy for HIV patients with depressive symptoms. Yet, as the sole study of its kind, our results require replication. Future reports will examine whether psychotherapy yielded benefits to offset its greater costs in time and therapist effort relative to pharmacotherapy.
Accepted for publication October 31, 1997.
This study was supported by grant MH-46250 from the National Institute of Mental Health, Rockville, Md, and from a fund established in The New York Community Trust by DeWitt-Wallace, New York, NY.
Presented in part at the American Psychopathological Association Annual Meeting, New York, NY, February 29, 1996, and in a New Research Session at the Annual Meeting of the American Psychiatric Association, New York, NY, May 7, 1996.
This article is dedicated to the late Samuel W. Perry, MD, who developed this research study, and to the late Gerald L. Klerman, MD, pioneer in interpersonal psychotherapy, who also played an active role in this research. Both were Professors of Psychiatry and Associate Chairmen for Research at Cornell University Medical College, New York, NY.
The authors gratefully acknowledge the comments of Judith Rabkin, PhD, on a draft of the manuscript, the dedication of the many study therapists, and the dedication of Dean Haglin, Kathy Dodd, and Alyce Kuklinski, without whom the study could never have been completed.
Reprints: John C. Markowitz, MD, Cornell University Medical College, 445 East 68th St, Suite 3N, New York, NY 10021 (e-mail: firstname.lastname@example.org).
Thank you for submitting a comment on this article. It will be reviewed by JAMA Psychiatry editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Register and get free email Table of Contents alerts, saved searches, PowerPoint downloads, CME quizzes, and more
Subscribe for full-text access to content from 1998 forward and a host of useful features
Activate your current subscription (AMA members and current subscribers)
Purchase Online Access to this article for 24 hours
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 101
Customize your page view by dragging & repositioning the boxes below.
The Rational Clinical Examination
Make the Diagnosis: Depression
The Rational Clinical Examination
Original Article: Is This Patient Clinically Depressed?
All results at
and access these and other features:
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.