Efficacy of newer medications for treating depression in primary care patients

REVIEWS

Efficacy of Newer Medications for Treating Depression in Primary Care Patients Cynthia D. Mulrow, MD, MSc, John W. Williams, Jr., MD, MHS, Elaine Chiquette, PharmD, Christine Aguilar, MD, MPH, Polly Hitchcock-Noel, PhD, Shuko Lee, MS, John Cornell, PhD, Karen Stamm, BA PURPOSE: Several medications have recently been introduced for the treatment of depression. We reviewed the literature to summarize their efficacy in the treatment of depression in adult patients in primary care settings. METHODS: We searched the literature published from 1980 to January 1998 using the Cochrane Collaboration Depression Anxiety and Neurosis Group’s specialized registry of 8,451 clinical trials, references from trials and 46 pertinent meta-analyses, and consultation with experts. We included randomized controlled trials of at least 6 weeks’ duration that measured clinical outcomes and compared one of 32 newer medications with another newer antidepressant, an older antidepressant, a placebo, or a psychosocial intervention for the treatment of depressed patients in primary care settings. The primary outcome was response rate, defined as the proportion of patients experiencing a 50% or greater improvement in depressive symptoms. RESULTS: There were 28 randomized controlled trials involving 5,940 adult primary care patients with major depression, depression requiring treatment, dysthymia, or mixed anxiety depression. Newer agents, including selective serotonin re-uptake inhibitors, serotonin norepinephrine inhibitors, reversible

inhibitors of monoamine oxidase, and dopamine antagonists, were usually compared with tricyclic agents. Average response rates were 63% for newer agents, 35% for placebo, and 60% for tricyclic agents. Newer agents were significantly more effective than placebo [risk ratio ⫽ 1.6; 95% confidence interval (CI), 1.2 to 2.1), but similar to tricyclic agents (risk ratio ⫽ 1.0; 95% CI, 0.9 to 1.1). Response rates were similar in the different types of depressive disorders, except that two small trials in frail older patients showed no significant effects of newer agents compared with placebo. Dropout rates as a result of adverse effects were 8% with newer agents and 13% with tricyclic agents (P ⬍0.05). CONCLUSIONS: In primary care settings, newer antidepressants are more effective than placebo and have similar efficacy compared with tricyclic agents in the acute treatment of depression. Dropout rates as a result of adverse effects are lower with newer compared with tricyclic agents. Future studies should compare the effectiveness of different therapies among primary care patients with less severe depression and greater medical and psychiatric comorbidity. Am J Med. 2000;108:54 – 64. 䉷2000 by Excerpta Medica, Inc.

O

US guidelines for depression in primary care were published in 1993. These guidelines recommended antidepressant therapy for patients with major depression who were seen in these settings (4). In 1998, the North of England Evidence-based Guideline Development Project recommended tricyclic agents as the first-line treatment for depression in primary care patients (5). In making its recommendations, the US guidelines reviewed almost 200 trials, of which only 24 were conducted in primary care settings. The North of England guidelines reviewed 125 trials, of which only eight were conducted in general practice settings. Most of the trials reviewed in the US guidelines compared tricyclic antidepressants with placebo, whereas most of the trials reviewed in the North of England guidelines compared selective serotonin re-uptake inhibitors (SSRIs) with tricyclic agents. The applicability of guidelines for primary care patients has been questioned (6). Many patients in primary care with depressive symptoms tend to be at the milder end of major depression or have minor depression. Pa-

ne in 5 patients are affected by a mood disorder during their lifetime. Depressed persons have greater mortality and impairment in many areas of functioning compared with nondepressed persons (1,2). The estimated monetary costs associated with depression in the United States (US) exceeded $44 billion in 1990 (3). Of patients seeking help for depression, nearly 75% go to a primary care physician. Five percent to 10% of patients seen by primary care providers have major depression (2). From the Evidence-based Practice Center, Department of Medicine, the University of Texas Health Science Center at San Antonio, San Antonio, Texas, and the Veterans Affairs Cochrane Center, Audie L. Murphy Veterans Affairs Hospital, San Antonio, Texas. Supported by the John D. and Catherine T. MacArthur Foundation and the Agency for Health Care Policy and Research, San Antonio, Texas. Requests for reprints should be addressed to Cynthia D. Mulrow, MD, MSc, Audie L. Murphy Memorial Veterans Affairs Hospital, 7400 Merton Minter Boulevard (11C6), San Antonio, Texas 78284. Manuscript submitted November 20, 1998, and accepted in revised form August 25, 1999. 54

䉷2000 by Excerpta Medica, Inc. All rights reserved.

0002-9343/00/$–see front matter PII S0002-9343(99)00316-2

Newer Medications for Treating Depression/Mulrow et al

Table 1. Newer Antidepressants by Classification Generic Name Selective serotonin re-uptake inhibitors (SSRIs) Fluoxetine Fluvoxamine Paroxetine Sertraline Citalopram

Indalpine* Tomoxetine* Litoxetine* Femoxetine* Selective norepinephrine re-uptake inhibitors Venlafaxine

Milnacipran Mirtazapine Norepinephrine re-uptake inhibitors Viloxazine*

Reboxetin* Reversible inhibitors of monoamine oxidase A Moclobemide*

Medifoxamine* Brofaromine*

Toloxatone* 5-HT2 receptor antagonists Nefazodone Ritanserin*

5-HT1A receptor agonists Gepirone* Ipsapirone* Tandospirone* Felsinoxan* Gabamimetics Fengabine*

Examples of Trade Names

Usual Dosage Range (mg/day)

20–60 100–300 20–50 50–200 20–80

Prozac Luvox Paxil Zoloft Vitalopram Cipramil Seralgan Seropram Not determined Not determined Not determined Not determined

Selectively inhibit the re-uptake of 5-HT at the presynaptic neuronal membrane

Not determined Not determined Not determined 300–600

Effexor Efexor Trewilor Not determined Remeron Zispin

75–350

Potent inhibitors of 5-HT and norepinephrine uptake; weak inhibitors of dopamine re-uptake

Not determined 15–45

Vivalan Vivarint Vicilan Edronax

100–400

Auronix Manerix Moclamine Molcamine Cledial Gerdaxyl Drug unavailable (previously marketed by CIBA) Humoryl Umoril

300–600

Serzone Dutonin Not determined

300–600 Not determined

Not determined Not determined Not determined Not determined

Not determined Not determined Not determined Not determined

Not determined

Mechanism of Action

Selective norepinephrine re-uptake inhibitors without serotonin reuptake inhibition activity

4–12

Selective and reversible inhibitors of monoamine oxidase A resulting in increased concentrations of norepinephrine, 5-HT and dopamine

150 75–150†

400–800

900–1800†

5-HT2 antagonists and 5-HT re-uptake inhibitors In addition, nefazodone exhibits ␣1adrenergic blocking activity, and ritanserin is a long-acting 5-HT1c receptor antagonist 5-HT1A partial receptor agonists resulting in downregulation of 5HT1A or 5-HT2 receptors or both

GABAA and GABAB receptor agonists (continued)

January 2000

THE AMERICAN JOURNAL OF MEDICINE威

Volume 108 55

Newer Medications for Treating Depression/Mulrow et al

Table 1. Continued Generic Name Dopamine antagonists Amisulpride*

Sulpiride*

Dopamine re-uptake inhibitors Bupropion

Miscellaneous Minaprine* Herbal remedies Hypericum preparations (St. John’s wort)

Examples of Trade Names Deniban Solian Sulamid Arminol Eglonyl Meresa

Usual Dosage Range (mg/day) 50

Not determined

Mechanism of Action D2/D3 dopamine receptor antagonists that act preferentially on presynaptic neuronal membranes Mostly used as antipsychotics

Wellbutrin Zyban

300–450

Dopamine re-uptake inhibitor with minimal effect on the re-uptake of norepinephrine or 5-HT

Cantor Isopulsan

100–300

Facilitates serotonergic, cholinergic, and dopaminergic neurotransmission

Hyperforat Sedariston Psychotonin Jarsin Neuroplant Esbericum Kira

300–900 of hypericum extracts†

Unclear

* Not US FDA approved. † Doses usually studied. 5-HT ⫽ 5 hydroxytryptophan; GABA ⫽ gamma amino butyric acid.

tients with major depression who are treated by psychiatrists have greater psychiatric comorbidity but may have less medical comorbidity than primary care patients. They are less likely to recover and are at greater risk for psychiatric hospitalization than patients with major depression treated in primary care (7). The recent emphasis on recognizing and treating depression in primary care patients as well as many new antidepressants have contributed to the growth in the prescription of antidepressant medications. Newer antidepressants and herbal remedies have led to wider but sometimes confusing choices for primary care providers and their patients. This article updates evidence about the efficacy and effectiveness of newer medications for treating depression in primary care patients. It focuses on randomized trials of newer medications (antidepressant agents from nine classes other than tricyclic or tetracyclic agents, triazolopyridines, and nonselective monoamine oxidase inhibitors) that recruited primary care patients or were conducted in general practice settings.

METHODS This work was part of a large report that identified and reviewed 315 newer pharmacotherapy and herbal treat56

January 2000

THE AMERICAN JOURNAL OF MEDICINE威 Volume 108

ment trials regardless of setting, which is available at www.ahcpr.gov/clinic/htm.

Sources and Search Strategy We searched English and non-English literature sources including the Cochrane Collaboration Depression Anxiety and Neurosis Group’s specialized registry of 8,451 clinical trial articles, references from trial articles and 46 pertinent meta-analyses, and consultations with experts. The specialized registry includes trials from electronic databases, such as MEDLINE, EMBASE, PsycLIT, LILACS, Psyndex, SIGLE, CINAHL, Biological Abstracts, and the Cochrane Library; hand searches of conference proceedings and 69 psychiatry-related journals; contacts with 30 pharmaceutical companies; and routine reference checking of trial articles and systematic reviews. Sources were searched from 1980 to January 1998. The terms “depression,” “depressive disorder,” or “dysthymic disorder” were combined with 32 specific agents from nine classes of antidepressants: amisulpride, brofaromine, bupropion, citalopram, felsinoxan, femoxetine, fengabine, fluoxetine, fluvoxamine, gepirone, hypericum (herbal remedy), indalpine, ipsapirone, kava kava (herbal remedy), litoxetine, medifoxamine, milnacipran, minaprine, mirtazapine, moclobemide, nefazodone, paroxetine, re-

Newer Medications for Treating Depression/Mulrow et al

boxetine, ritanserin, sertraline, sulpiride, tandospirone, toloxatone, tomoxetine, valeriana (herbal remedy), venlafaxine, and viloxazine (Table 1).

Selection Process Trials were included if they studied depressed patients recruited from, or followed in, primary care settings; were randomized and controlled; compared a newer antidepressant with either another newer antidepressant, an older antidepressant, a placebo, or a psychosocial intervention; had an active intervention period of at least 6 weeks; and measured a clinical outcome such as depression, functional status, or quality of life. The original search, which included herbal therapies and other medications, identified 1,277 records, of which 418 were rejected by two independent reviewers as definitely not meeting inclusion criteria. Of the remaining 859 records, the texts of 748 studies were obtained; all but one (8) of the unavailable studies were identified from meta-analyses or pharmaceutical company records, none of which appeared to have been conducted in primary care settings. Of the 459 remaining studies, 315 unique trials were identified, including 27 studies in which more than 90% of patients were recruited from, or followed in, primary care or general practice settings. One additional study in a primary care setting was identified during ongoing searches of MEDLINE and PubMed from January to August 1998.

Data Abstraction Two independent investigators (a physician and a pharmacologist) abstracted data from each trial, including adequacy of randomization, whether the trial was single- or double-blind, cointerventions, and numbers of dropouts. Disagreements were resolved by consensus. A third independent investigator (a physician) verified information about type of depression diagnosis, number of dropouts and the reasons for dropping out, and outcomes.

Data Analysis The primary outcome measure was response rate, defined as the proportion of patients having a 50% or greater improvement in symptoms as assessed by a depression symptoms rating scale, or a rating of much or very much improved as assessed by a global assessment method. (Analyses using continuous scores from depression scales were also done, because a few trials did not provide categorical response data. These analyses were similar to those using the response rates and are not reported.) Because trials rarely collected symptom outcomes in patients who dropped out, analyses were not based on true intention-to-treat results. Instead, a modified intention-to-treat method was used (9), in which the denominators for the relative frequencies were the numbers of patients assigned to the treatment arm, and the numerators were the numbers of patients who stayed in treat-

ment and responded. To the extent that some patients who left treatment may have responded, these response rates are lower than those derived from a true intentionto-treat analysis. Thus, a sensitivity analysis was performed in which the denominator was the number of patients who completed follow-up or whose last response was carried forward (endpoint analysis). The DerSimonian and Laird empirical Bayes randomeffects estimator was used to estimate the pooled measures of treatment efficacy estimated as the log of the ratios of the response rates (10,11). Computations were performed with the STATA meta command (11,12). Outlier studies that might help explain heterogeneity in estimates of treatment efficacy were identified with Galbraith plots (13). Publication bias was assessed with funnel plots using Begg’s rank order correlation test.

RESULTS Of the 28 trials that evaluated newer agents in 5,940 adult primary care patients (14 – 41), 27 trials were conducted in Europe and one in the United States 41). Twenty trials had clearly reported funding sources (14,16 –23,27,29, 32–36,38,40); all were sponsored by pharmaceutical companies. Clinical characteristics of participants and design characteristics of the trials are given in Table 2. Seventy-one percent of all participants in the trials were women. The average age of participants was 45 years. Patients with alcohol abuse, cognitive impairment, and serious medical illnesses were typically excluded, except in the US trial, which included patients with medical illnesses. Two of the trials were limited to frail elderly patients, one of which included primary care patients who were receiving home health care (40) and the other of which enrolled general practitioners’ patients who had been admitted to a hospital for physical illness (38). Among trials that reported the Hamilton Rating Depression Scale scores at baseline, the average score was approximately 22. The newer agents that were evaluated included SSRIs, norepinephrine reuptake inhibitors, reversible inhibitors of monoamine oxidase, and dopamine antagonists. Most trials compared newer agents with first- or second-generation tricyclic agents. All but one trial used drug doses that were in usual dose ranges, although many were toward the lower range of standard doses. None of the trials compared newer pharmacotherapies with counseling, psychosocial, or psychotherapy interventions. Several depressive disorders were studied including major depression (n ⫽ 15 trials), “depression requiring treatment” as judged by the primary care provider (n ⫽ 4), dysthymia (n ⫽ 2), mixed anxiety depression (n ⫽ 3), and patients with more than one type of depressive disorder (n ⫽ 4). Treatment duration was typically 6 to 8 weeks (n ⫽ 22); January 2000

THE AMERICAN JOURNAL OF MEDICINE威

Volume 108 57

58

January 2000

THE AMERICAN JOURNAL OF MEDICINE威 Volume 108

Lecrubier et al (30)

Christiansen et al (29)

Skrumsager and Jeppesen (27) Lingjaerde et al (28)

Rosenberg et al (26)

Wheatley (25)

Wheatley (24)

Beaumont et al (23)

Kragh Sorensen et al (22)

Moon and Jesinger (21)

Doogan and Langdon (20)

Moon and Vince (19)

Hutchinson et al (18)

Corne and Hall (17)

Tylee et al (16)

Ekselius et al (15)

Patris et al (14)

Study (reference)

Citalopram (20 mg) vs Fluoxetine (20 mg) Citalopram (20–60 mg) vs Sertraline (50–150 mg) Fluoxetine (20 mg) vs Venlafaxine (37.5 mg) Fluoxetine (40–60 mg) vs Dothiepin (75–100 mg) Paroxetine (20 mg) vs Amitriptyline (100 mg) Paroxetine (20–30 mg) vs Lofepramine (140–210 mg) Placebo vs Sertraline (50–100 mg) vs Dothiepin (75–150 mg) Fluvoxamine (100–300 mg) vs Mianserin (60–180 mg) Moclobemide (400 mg) vs Clomipramine (150 mg) Moclobemide (450 mg) vs Dothiepin (150 mg) Minaprine (200 mg) vs Minaprine (200 mg) vs Amitriptyline (150 mg) Minaprine (200 mg) vs Minaprine (300 mg) vs Mianserin (60 mg) Citalopram (20–30 mg) vs Citalopram (20–60 mg) vs Imipramine (100–150 mg) Femoxetine (600–800 mg) vs Amitriptyline (150–200 mg) Moclobemide (400–600 mg) vs Doxepin (150–250 mg) Paroxetine (20–40 mg) vs Amitriptyline (100–150 mg) Placebo vs Venlafaxine (75–150 mg) vs Imipramine (75–150 mg)

Intervention (daily dose)

Table 2. Clinical and Design Characteristics of Primary Care Trials

90

Major depression

Mild moderate depression

229

13 weeks

8 weeks

6 weeks

56 144

6 weeks

6 weeks

6 weeks

6 weeks

6 weeks

6 weeks

6 weeks

6 weeks

6 weeks

6 weeks

6 weeks

12 weeks

24 weeks

8 weeks

Length

81

472

Depression requiring treatment Depression requiring treatment Depression requiring treatment “Depressive illness”

117

144

345

142

62

308

Major depression

Major depression

Major depression

Major depression

Major depression

Major depression

122

100

Major depression

Major depression

341

400

357

Sample Size

Mixed group

Major depression

Major depression

Diagnosis

HAMD, CGI

DSMIII, HAMD ⱖ18

RDC mild to moderate intensity score of 4–8 on RASKIN three areas scale*

(continued)

HAMD, CGI, Well Being* MADRS, CGI, Social Adjustment Anxiety Scale*

MADRS, CGI

HAMD

HAMD, MADRS, CGI

HAMD ⱖ14 and “in need of antidepression treatment” Depressive illness requiring treatment HAMD ⱖ17 No specific diagnosis required HAMD ⱖ15 (17 items)*

HAMD

HAMD, MADRS

HAMD, CGI

HAMD

DSMIII

DSMIIIR Newcastle HAMD ⬎10 DSMIIIR HAMD (17 items) ⬎13 DSMIIIR

MADRS, CGI

MADRS, CGI

HAMD, CGI

RDC HAMD ⱖ17

DSMIIIR, MADRS ⱖ22 & CGI Severity ⱖ4 DSMIII, MADRS ⱖ24

HAMD, MADRS, CGI

DSMIV, MDARS ⱖ19

MADRS, CGI

MADRS, CGI

DSMIIIR, MADRS ⱖ21

DSMIIIR

HAMD, MADRS, CGI

Reported Outcomes

DSMIIIR, MADRS ⱖ22

Inclusion Criteria

Newer Medications for Treating Depression/Mulrow et al

January 2000

Flupenthixol (1–2 mg) vs Ritanserin (5–10 mg) Placebo vs Sulpiride (100–200 mg)

Geisler et al (39) Mixed group depression Depression requiring treatment

Dysthymia

42

“Endogenous depression” Major depression

536

66

67

82

23

24 weeks

20 weeks

6 weeks

8 weeks

6 weeks

6 weeks

6 weeks

6 weeks

112 50

6 weeks

8 weeks

12 weeks

Length

106

1019

323

Sample Size

Major depression

Major depression and anxiety Mixed anxiety depression Mixed anxiety depression “Depressive illness”

Dysthymia

Diagnosis

DSMIIIR primary care MD decision to treat*

DSMIII ZUNG (45)

HAMD ⱖ25, ICD-8 296.2 or 300.4 DSMIIIR HAMD (17 items) ⱖ18; CGI ⱖ 4 Newcastle endogenous type and HAMD ⬎17 ELDRS ⱖ5; GMS-AGECAT case level of depression DSMIII HAMD ⱖ 15

Depression with anxiety; MADRS ⱖ20; CAS ⱖ11 DSMIIIR, HAMD ⱖ17 and HAMA ⱖ16 MADRS ⱖ21 and CAS ⱖ11

DSMIIIR, MADRS ⬍21

Inclusion Criteria

HAMD, SF-36,* HSCL

HAMD

HAMD, MADRS, CGI

HAMD

HAMD

HAMD, SIP

HAMD, CGI

MADRS, CGI, CAS*

HAMD, CGI

HAMD, MADRS, CGI

MADRS, CGI

Reported Outcomes

* Quality of life outcome. CAS ⫽ Clinical Anxiety Scale; CGI ⫽ Clinical Global Improvement Scale; DSMIII ⫽ Diagnostic Statistical Manual 3rd edition; DSMIIIR ⫽ Diagnostic Statistical Manual, 3rd revised edition; DSMIV ⫽ Diagnostic Statistical Manual, 4th edition; ELDRS ⫽ Evans Liverpool Depression Rating Scale; GMS AGECAT ⫽ Geriatric Mental State and a New Computerized Diagnostic System; HAMA ⫽ Hamilton Anxiety Scale; HAMD ⫽ Hamilton Rating Depression Scale; HSCL ⫽ Hopkins Symptom Checklist; ICD ⫽ International Classification of Diseases; MADRS ⫽ Montgomery-Asberg Depression Rating Scale; RASKIN ⫽ Raskin Depression Rating Scale; RDC ⫽ Research Diagnostic Criteria; SF-36 ⫽ Medical Outcomes Study SF36; SIP ⫽ Sickness Impact Profile; Well-being ⫽ Visual Analogue Scale; ZUNG ⫽ Zung Depression Scale.

Simon et al (41)

Kivela and Lehtomaki (40)

Evans et al (38)

Barge Schaapveld et al (36) Dahl et al (37)

Aylward et al (35)

Laws et al (34)

Moon et al (53)

Fluoxetine (20 mg) vs Desipramine (100 mg) vs Imipramine (100 mg)

Placebo vs Amisulpride (50 mg) vs Amineptine (200 mg) Paroxetine (20–40 mg) vs Clomipramine (75–150 mg) Sertraline (50–150 mg) vs Clomipramine (50–150 mg) Fluvoxamine (100–300 mg) vs Lorazepam (2–6 mg) Sulpiride (400–1000 mg) vs Amitriptyline (75–200 mg) Fluvoxamine (100 mg) vs Amitriptyline (150 mg) Femoxetine (600–800 mg) vs Desipramine (150–225 mg) Placebo vs Fluoxetine (20 mg)

Boyer and Lecrubier (31)

Ravindran et al (32)

Intervention (daily dose)

Study (reference)

Table 2. Continued

Newer Medications for Treating Depression/Mulrow et al

THE AMERICAN JOURNAL OF MEDICINE威

Volume 108 59

Newer Medications for Treating Depression/Mulrow et al

Figure 1. Comparison of different antidepressants in trials of patients with major depression in primary care settings. Results are presented as ratios of response rates on a log scale. Barge (36) and Dahl (37) are not included because their outcome data were not provided as response rates.

three trials were 12 weeks long and three were 20 to 24 weeks long. Dropout rates for the 27 trials giving such information were less than 20% in 12 trials, 20% to 30% in 12 trials, and greater than 30% in three trials. All trials except for the US trial (41), which was designed as an effectiveness trial in a managed care setting, and the trial by Barge Schaapveld et al (36) were double-blind. All trials were randomized, but few reported the methods of randomization or the allocation concealment. Only four trials measured functional status or quality of life (29,30,34,41). All trials had recruitment and inclusion criteria that specified participants as patients of primary care providers or general practitioners, although few described the actual treatment environment or characteristics of the providers who delivered treatment. Decisions regarding antidepressant management (initial dose, dosage changes, treatment discontinuation or switches) generally followed set protocols, except in the US trial in which patients and their physicians made these decisions. Follow-up was generally conducted with outpatient visits, although the US trial relied primarily on telephone follow-up assessments. Follow-up assessments primarily measured depressive symptoms using standardized rating scales; most studies did not specify who made these 60

January 2000

THE AMERICAN JOURNAL OF MEDICINE威 Volume 108

assessments. Whether general supportive measures and counseling occurred during follow-up visits was not specified.

Efficacy of Therapy Comparisons of the response rates for the different trials are shown in Figures 1 and 2. Sixty-three percent of the patients who received newer antidepressants had at least a 50% improvement in symptoms, as measured by a depressive symptoms rating scale (eg, the Hamilton Depression Rating Scale), or were much to very much improved, as measured by a global assessment method (eg, the Clinical Global Impression Scale). Response rates did not vary significantly among the newer agents except for minaprine, for which two studies showed lower response rates of approximately 34%. The average response rate for patients given placebo was 35%, whereas the average response rate for patients given tricyclic agents was 60%. Modified intention-to-treat analyses showed that newer antidepressant agents were significantly more effective than placebo [risk ratio ⫽ 1.6; 95% confidence interval (CI), 1.2 to 2.1) but similar to tricyclic agents (risk ratio ⫽ 1.0; 95% CI, 0.9 to 1.1) in efficacy. Endpoint analyses gave similar risk ratios (newer vs placebo risk ratio ⫽ 1.5; 95% CI, 1.1 to 2.0; newer vs tricyclic agent

Newer Medications for Treating Depression/Mulrow et al

Figure 3. Publication bias funnel plot for trials of newer antidepressants compared with placebo. Begg’s funnel plot with pseudo 95% confidence limits, P value ⫽ 0.33. Numbers in parentheses indicate reference number of study.

Adverse Effects of Treatment

Figure 2. Comparison of different antidepressants in trials of depressed patients in primary care settings by diagnosis other than major depression. Results are presented as ratios of response rates on a log scale. Aylward (35), Geisler (39), and Simon (41) are not included because their outcome data were not provided as response rates.

Of the patients treated with placebo, 2% dropped out because of adverse effects. Dropouts from adverse effects with the newer agents (8%) were significantly less common than with tricyclic agents (13%). The pooled rate difference for the dropout rate was ⫺4% (95% CI, ⫺7% to 0%). There was some variability among newer agents, with two venlafaxine studies having greater (19%) dropout rates from adverse effects and one minaprine study having a lower dropout rate of 2%.

Publication Bias risk ratio ⫽ 1.0; 95% CI, 0.9 to 1.1). Tests for heterogeneity were significant for both analyses (newer vs placebo, P ⫽ 0.03; newer vs tricyclic agents, P ⫽ 0.003). When either the two placebo-controlled trials in physically frail elders (38,40) or the trial in patients with dysthymia (31) were excluded from analyses, tests for heterogeneity were no longer significant (P ⫽ 0.33; P ⫽ 0.29). Risk ratios excluding the trials in physically frail elders and the trial in patients with dysthymia were 1.8 (95% CI, 1.5 to 2.2) for the comparison of newer agents with placebo and 1.5 (95% CI, 1.2 to 1.9) for the comparison of newer agents with tricyclic agents. Excluding the trial (23) responsible for the statistical heterogeneity in the comparison of newer with tricyclic agents did not change the risk ratio. Response rates appeared similar among the different depressive disorders (Figures 1 and 2), although there were too few studies and too many types of disorders to evaluate this comparison formally with meta-regression techniques. Although newer antidepressants were more effective than placebo in analyses that included all trials, neither of the two small trials that were conducted in medically ill frail older patients showed significant differences between newer agents (sulpiride or fluoxetine) and placebo.

There was no statistically significant evidence of publication bias (Figures 3 and 4). The horizontal axis of the funnel plot evaluating the few placebo-controlled trials was shifted above 0, compatible with publication bias for positive studies that may not have been detected statistically because of low power.

Figure 4. Publication bias funnel plot for trials of newer antidepressants compared with tricyclic agents. Begg’s funnel plot with pseudo 95% confidence limits, P value ⫽ 0.88. Numbers in parentheses indicate reference number of study. January 2000

THE AMERICAN JOURNAL OF MEDICINE威

Volume 108 61

Newer Medications for Treating Depression/Mulrow et al

DISCUSSION This analysis focused on primary care trials that evaluated newer antidepressant agents, including SSRIs, norepinephrine reuptake inhibitors, reversible inhibitors of monoamine oxidase, and dopamine antagonists. Only five of the 28 trials were considered in either the US (4) or the North of England (5) guideline projects. Our results confirmed that although antidepressant therapy is more efficacious than placebo in treating depression in primary care patients, there were no major differences in efficacy between the newer agents and tricyclic agents. There were also not any differences among the newer agents, although only a few studies directly compared them. The similar efficacy of various antidepressants is consonant with findings from several recent large meta-analyses that were not limited to primary care settings (5,42– 44). The magnitude of benefit from treatment appears at least as good as, and perhaps better than, that seen in mental health outpatient settings (44), perhaps because participants in mental health and primary care trials have generally had similar, moderately severe depression. Previous meta-analyses have also examined dropout rates. The average dropout rates from all causes among patients in antidepressant trials of newer agents have ranged from 16% to 33%, whereas dropout rates among those treated with older agents (primarily tricyclic agents) have ranged from 24% to 38% (42– 47). The differences in dropout rates are typically in the range of 1% to 5% in favor of newer agents, although the differences are usually not statistically significant. The average dropout rates as a result of adverse effects of the medications range from 5% to 15% with the newer agents compared with 11% to 19% with tricyclic agents. Pooled rate differences have generally been 4% to 5%, except for one analysis that found nonsignificant differences of 2% to 3%. These analyses have also shown that rates of specific adverse effects, such as headache, nausea, dry mouth, and constipation, vary significantly among agents (42,44). This report highlights major gaps in the evidence concerning the optimal treatment of depression in primary care settings. Trials of the newer medications are limited in scope and have not addressed the benefits, risks, and costs of alternative psychosocial and herbal treatments. Even though primary care trials have studied various diagnoses and forms of depression, many types of depressed patients who are seen in the primary care setting, such as those with minor forms of depression, serious medical or physical illness, and other psychiatric illness, generally have not been offered the opportunity to participate in trials. Clinicians may harbor beliefs that antidepressant therapy is less effective in such patients, even though they have substantial impairment. Most available trials, except for the large US trial that 62

January 2000

THE AMERICAN JOURNAL OF MEDICINE威 Volume 108

evaluated the effectiveness of treatment within a health maintenance organization (41), were fairly similar, industry-funded, efficacy studies. Little information was provided about aspects of care that may be relevant to providers and their patients, such as whether routine educational messages that may improve adherence—to take medicine daily, not to expect results for 2 to 4 weeks, and continue to take medications even when feeling better— were used (48). The optimal content and frequency of visits, and the utility of specialty consults, was never evaluated. This is particularly important, because a US trial has shown that intensive collaborative management between primary care physicians and consulting psychiatrists compared with “usual care” improved adherence, satisfaction, and depressive symptoms in patients with major, but not minor, depression (49). None of the reviewed trials compared newer agents with psychosocial interventions. A recently published trial based in Norwegian general practices showed that the combination of sertraline and simple psychological treatment was more effective than psychological treatment alone, particularly for patients with recurrent depression (50). Psychological treatment delivered by general practitioners included counseling and emotional support but no specific psychotherapy. Primary care trials suggest that approximately 60% of participating patients have a 50% or greater response to medication. Whether these results are applicable to or overly optimistic compared with patients who do not participate in trials is not known. Regardless, they suggest the continued need for better and innovative methods of treating depression for the remaining 40% of patients. More evaluations of yet another “me-too” antidepressant are not likely to improve the care of depressed patients. Rather, effectiveness studies conducted in broader groups of primary care patients with less severe depression and greater medical, as well as psychiatric, comorbidity are needed. Such trials should study longer durations of treatment, employ and describe delivery characteristics that are feasible in general practice settings, evaluate alternative therapies that are of interest to patients, and measure adherence, as well as functional, general well-being, and cost outcomes.

ACKNOWLEDGMENT We would like to thank Kelly Montgomery and Dave Mullins for their technical support in the development of this manuscript.

REFERENCES 1. Greenberg PE, Stiglin LE, Finkelstein SN, Berndt ER. Depression: a neglected major illness. J Clin Psychiatry. 1993;54:419 – 424. 2. Hays RD, Wells KB, Sherbourne CD, Rogers W, Spritzer K. Func-

Newer Medications for Treating Depression/Mulrow et al

3.

4.

5.

6.

7.

8.

9.

10. 11. 12. 13. 14.

15.

16.

17.

18.

19.

20.

21.

22.

tioning and well-being outcomes of patients with depression compared with chronic general medical illnesses. Arch Gen Psychiatry. 1995;52:11–19. Simon G, Ormel J, VonKorff M, Barlow W. Health care costs associated with depressive and anxiety disorders in primary care. Am J Psychiatry. 1995;152:352–357. Depression Guideline Panel. Depression in primary care; vols. 1 and 2. Detection and diagnosis. Treatment of major depression. US Department of Health and Human Services, 1993. North of England Evidence Based Guideline Development Project. The choice of antidepressants for depression in primary care. Centre for Health Services Research, 1998. Mun˜oz RF, Hollon SD, McGrath E, Rehm LP, VandenBos GR. On the AHCPR depression in primary care guidelines: further consideration for practitioners. Am Psychol. 1994;49:42– 61. Cooper-Patrick L, Crum RM, Ford DE. Characteristics of patients with major depression who received care in general medical and specialty mental health settings. Med Care. 1994;32:15–24. Cerrahoglu L, Cogalgil S, Senel K, Turkmen MB. The effects of antidepressive therapy in fibromyalgia syndrome. Fiz Tedavy Rehabil Derg. 1995;19:212–216. Thase ME, Trevidi MH, Rush AJ. MAOIs in the contemporary treatment of depression. Neuropsychopharmacology 1995;12:185– 219. Berkey CS, Hoaglin DC, Mosteller F, Colditz GA. A random-effects regression model for meta-analysis. Stat Med. 1995;14:395– 411. Sharp S. sbe23: meta-regression. Stata Tech Bull. 1998;42:16 –22. Sharp S, Sterne J. sbe16: meta-analysis. Stata Tech Bull. 1997;38:9 – 14. Galbraith R. A note on graphical presentation of estimated odds ratios from several clinical trials. Stat Med. 1988;7:889 – 894. Patris M, Bouchard JM, Bougerol T, et al. Citalopram versus fluoxetine: a double-blind, controlled, multicentre, phase III trial in patients with unipolar major depression treated in general practice. Int Clin Psychopharmacol. 1996;11:129 –136. Ekselius L, Von Knorring L, Eberhard G. A double-blind multicenter trial comparing sertraline and citalopram in patients with major depression treated in general practice. Int Clin Psychopharmacol. 1997;12:323–331. Tylee A, Beaumont G, Bowden MW, Reynolds A. A double-blind, randomized, 12-week comparison study of the safety and efficacy of venlafaxine and fluoxetine in moderate to severe major depression in general practice. Prim Care Psychiatr. 1997;3:51–58. Corne SJ, Hall JR. A double-blind comparative study of fluoxetine and dothiepin in the treatment of depression in general practice. Int Clin Psychopharmacol. 1989;4:245–254. Hutchinson DR, Tong S, Moon CA, Vince M, Clarke A. Paroxetine in the treatment of elderly depressed patients in general practice: a double-blind comparison with amitriptyline. Int Clin Psychopharmacol. 1992;6(suppl 4):43–51. Moon CA, Vince M. Treatment of major depression in general practice: a double-blind comparison of paroxetine and lofepramine. Br J Clin Pract. 1996;50:240 –244. Doogan DP, Langdon CJ. A double-blind, placebo-controlled comparison of sertraline and dothiepin in the treatment of major depression in general practice. Int Clin Psychopharmacol. 1994;9:95– 100. Moon CA, Jesinger DK. The effects of psychomotor performance of fluvoxamine versus mianserin in depressed patients in general practice. Br J Clin Pract. 1991;45:259 –262. Kragh Sorensen P, Muller B, Andersen JV, Buch D, Stage KB. Moclobemide versus clomipramine in depressed patients in general practice. A randomized, double-blind, parallel, multicenter study. J Clin Psychopharmacol. 1995;15(4 suppl 2):24S–30S.

23. Beaumont G, Gringras M, Hobbs FD, et al. A randomised, doubleblind, multi-centre, parallel-group study comparing the tolerability and efficacy of moclobemide and dothiepin hydrochloride in depressed patients in general practice. Int Clin Psychopharmacol. 1993;7:159 –165. 24. Wheatley D. Minaprine: an anticholinergic-free antidepressant? Results of a controlled trial of mianserin. Br J Psychiatry. 1989;155: 106 –107. 25. Wheatley D. Minaprine in depression. A controlled trial with amitriptyline. Br J Psychiatry. 1992;161:113–115. 26. Rosenberg C, Damsbo N, Fuglum E, Jacobsen LV, Horsgard S. Citalopram and imipramine in the treatment of depressive patients in general practice. A Nordic multicentre clinical study. Int Clin Psychopharmacol. 1994;9(suppl 1):41– 48. 27. Skrumsager BK, Jeppesen K. Femoxetine and amitriptyline in general practice: a randomized double-blind group comparison. Pharmacopsychiatry. 1986;19:368 –377. 28. Lingjaerde O, Jorgensen J, Storen R, et al. A double-blind comparison of moclobemide and doxepin in depressed general practice patients. Acta Psychiatr Scand. 1995;92:125–131. 29. Christiansen PE, Behnke K, Black CH, Ohrstrom JK, Bork Rasmussen H, Nilsson J. Paroxetine and amitriptyline in the treatment of depression in general practice. Acta Psychiatr Scand. 1996;93:158 – 163. 30. Lecrubier Y, Bourin M, Moon CA, et al. Efficacy of venlafaxine in depressive illness in general practice. Acta Psychiatr Scand. 1997;95: 485–93. 31. Boyer P, Lecrubier Y. Atypical antipsychotic drugs in dysthymia: placebo controlled studies of amisulpride versus imipramine versus amineptine. Eur Psychiatry. 1996;11(suppl 3):135S–140S. 32. Ravindran AV, Jude R, Hunter BN, Bray J, Morton NH. A doubleblind, multicenter study in primary care comparing paroxetine and clomipramine in patients with depression and associated anxiety. Paroxetine study group. J Clin Psychiatry. 1997;58:112–118. 33. Moon CA, Jago W, Wood K, Doogan DP. A double-blind comparison of sertraline and clomipramine in the treamtent of major depressive disorder and associated anxiety in general practice. J Psychopharmacology. 1994;8:171–176. 34. Laws D, Ashford JJ, Anstee JA. A multicentre double-blind comparative trial of fluvoxamine versus lorazepam in mixed anxiety and depression treated in general practice. Acta Psychiatr Scand. 1990;81:185–189. 35. Aylward M, Maddock J, Dewland PM, Lewis PA. Sulpride in depressive illness. Adv Biol Psychiatr. 1981;7:154 –165. 36. Barge Schaapveld DQCM, Nicolson NA, van der Hoop RG, DeVries MW. Changes in daily life experience associated with clinical improvement in depression. J Affect Disord. 1995;34:139 –154. 37. Dahl LE, Lundin L, le Fevre Honore P, Dencker SJ. Antidepressant effect of femoxetine and desipramine and relationship to the concentration of amine metabolites in cerebrospinal fluid. A doubleblind evaluation. Acta Psychiatr Scand. 1982;66:9 –17. 38. Evans M, Hammond M, Wilson K, Lye M, Copeland J. Placebocontrolled treatment trial of depression in elderly physically ill patients. Int J Geriatr Psychiatry. 1997;12:817– 824. 39. Geisler A, Mygind S, Knudsen OR, Sloth-Nielsen M. Ritanserin and flupenthixol in dysthymic disorder: a controlled double blind study in general practice. Nord J Psychiatry. 1992;46:237–243. 40. Kivela SL, Lehtomaki E. Sulpiride and placebo in depressed elderly outpatients: a double-blind study. Int J Geriatr Psychiatry. 1987;2: 255–260. 41. Simon GE, VonKorff M, Heiligenstein JH, Revicki DA, Grothaus L, Katon W, Wagner EH. Initial antidepressant choice in primary care: effectiveness and cost of fluoxetine vs tricyclic antidepressants. JAMA. 1996;275:1897–1902. January 2000

THE AMERICAN JOURNAL OF MEDICINE威

Volume 108 63

Newer Medications for Treating Depression/Mulrow et al 42. Canadian Coordinating Office for Health Technology Assessment. Selective serotonin reuptake inhibitors (SSRIs) for major depression. Part I: evaluation of the clinical literature. CCOHTA Report 1997 3E, 1997. 43. Hotopf M, Hardy R, Lewis G. Discontinuation rates of SSRIs and tricyclic antidepressants: a meta-analysis and investigation of heterogeneity. Br J Psychiatry. 1997;170:120 –127. 44. Mulrow C, Williams J, Chiquette E, et al. Newer pharmacotherapies for depression. Psycho Pharm Bull. 1999;34:409 –795. 45. Anderson IM, Tomenson BM, Treatment discontinuation with selective serotonin reuptake inhibitors compared with tricyclic antidepressants: a meta-analysis. BMJ. 1995;310:1433–1438. 46. Montgomery SA, Henry J, McDonald G, et al. Selective serotonin

64

January 2000

THE AMERICAN JOURNAL OF MEDICINE威 Volume 108

47.

48.

49. 50.

reuptake inhibitors: meta-analysis of discontinuation rates. Int Clin Psychopharmacol. 1994;9:47–53. Song F, Freemantle N, Sheldon TA, et al. Selective serotonin reuptake inhibitors: meta-analysis of efficacy and acceptability. BMJ. 1993;306:683– 687. Lin E, Von Korff M, Katon W, et al. The role of the primary care physician in patients’ adherence to antidepressant therapy. Med Care. 1995;1:67–74. Katon W, Von Korff M, Lin E, et al. Collaborative management to achieve treatment guidelines. JAMA. 1995;273:1026 –1031. Malt UF, Robak OH, Madsbu H-P, Bakke O, Loeb M. The Norwegian naturalistic treatment study of depression in general practice (NORDEP)-I: randomised double blind study. BMJ. 1999;318: 1180 –1184.