dysthymic research

See d is c u ssio n s, s ta ts , a n d a u th or p ro fil e s f o r t h is p ub lic a tio n a t: http s:/ /w ww.r e se a rc h ga te .n et/ p ub lic a tio n /2 30828533 Sele ctiv e s e ro to n in r e u p ta k e i n h ib it o rs a n d tr ic y clic a n tid ep re ssa n ts i n t h e a cu te t r e atm en t of c h ro n ic d ep re ssio n a n d...

Artic le in Jo u rn al o f A ffe ctiv e D is o rd ers · S ep te m ber 2 012 DO I: 1 0.1 016/j. ja d .2 012.0 6.0 07 · So u rc e : P ub M ed CIT A TIO NS 39 REA D S 335 5 a u th ors , i n clu d in g:

So m e o f t h e a u th ors o f t h is p ub lic a tio n a re a ls o w ork in g o n t h ese r e la te d p ro je cts :

Patie n t P re fe re n ce s a n d S h are d D ecis io n M akin g i n t h e T re a tm en t o f S u b sta n ce U se D is o rd ers : A Syste m atic R evie w o f t h e L it e ra tu re Vie w p ro je ct Effic a cy o f a b rie f m an ualiz e d i n te rv e n tio n M an agin g C an ce r a n d L iv in g M ea n in gfu lly ( C ALM ) a d ap te d to G erm an c a n ce r c a re s e ttin gs Vie w p ro je ct Ale ssa J a n se n Bun d esp sy ch oth era p eu te n ka m mer 44 PU BLIC ATIO NS 225 CIT A TIO NS SEE P R O FIL E La rs P H ölz e l Park klin ik W ie sb ad en S ch la n gen b ad 74 PU BLIC ATIO NS 535 CIT A TIO NS SEE P R O FIL E Martin H ärte r Univ e rs it y M ed ic a l C en te r H am burg - E p pen d … 551 PU BLIC ATIO NS 6,0 51 CIT A TIO NS SEE P R O FIL E Le ve n te K ris to n Univ e rs it y M ed ic a l C en te r H am burg - E p pen d … 238 PU BLIC ATIO NS 2,2 38 CIT A TIO NS SEE P R O FIL E All c o n te n t f o llo w in g t h is p age w as u p lo ad ed b y Le ve n te K ris to n o n 2 8 S ep te m ber 2 014.

Th e u se r h as r e q ueste d e n han ce m en t o f t h e d ow nlo ad ed f il e . All i n -t e xt r e fe re n ce s un d erlin ed i n b lu e a re a d ded t o t h e o rig in al d ocu m en t an d a re l in ked t o p ub lic a tio n s o n R ese a rc h G ate , l e ttin g y o u a cce ss a n d r e a d t h em i m med ia te ly . Review Selective serotonin reuptake inhibitors and tricyclic antidepressants in the acute treatment of chronic depression and dysthymia: A systematic review and meta-analysis A. von Wolff a,n, L.P. H¨ olzel b, A. Westphal a,M.H¨ arter a, L. Kriston a aDepartment of Medical Psychology, University Medical Center Hamburg-Eppendorf, Hamburg, GermanybDepartment of Psychiatry and Psychotherapy, University Medical Center Freiburg, Freiburg, Germany article info Article history:

Received 2 February 2012 Received in revised form 4 June 2012 Accepted 5 June 2012 Available online 7 September 2012 Keywords:

Antidepressants Chronic depression Systematic review Meta-analysis Efﬁcacy abstract Introduction:Chronic depression represents a substantial portion of depressive disorders and is associated with severe consequences. This review examined the efﬁcacy and acceptability of selective serotonin reuptake inhibitors (SSRIs) and tricyclic antidepressants (TCAs) in the treatment of chronic depression. Additionally, the comparative effectiveness of the two types of antidepressants has been examined.

Methods:A systematic search was conducted in the following databases: CENTRAL, MEDLINE, EMBASE, ISI Web of Science, BIOSIS, PsycINFO, and CINAHL. Primary efﬁcacy outcome was a response to treatment; primary acceptance outcome was dropping out of the study. Only randomized controlled trials were considered.

Results:We identiﬁed 20 studies with 22 relevant comparisons. 19 studies focused on samples with a majority of dysthymic patients. Both SSRIs and TCAs are efﬁcacious in terms of response rates when compared to placebo (Beneﬁt Ratio [BR]¼1.49;po0.001 for SSRIs and BR¼1.74;po0.001 for TCAs) and no statistically signiﬁcant differences between the active drugs and placebo in terms of dropout rates could be found. No differences in effectiveness were found between SSRIs and TCAs in terms of response rates (BR¼1.01;p¼0.91), yet, SSRIs showed statistically better acceptability in terms of dropout rates than TCAs (Odds Ratio [OR]¼0.41;p¼0.02).

Limitations:The methodological quality of the primary studies was evaluated as unclear in many cases and more evidence is needed to assess the efﬁcacy of SSRIs and TCAs in patients suffering from chronic forms of depression other than dysthymia.

Conclusions:This systematic review provides evidence for the efﬁcacy of both SSRIs and TCAs in the treatment of chronic depression and showed a better acceptability of SSRIs.

Contents 1. Introduction........................................................................................................8 2. Methods ...........................................................................................................8 2.1. Eligibility criteria..............................................................................................8 2.2. Search strategy ................................................................................................8 2.3. Study selection ................................................................................................9 2.4. Data collection and assessment of methodological quality . ............................................................9 2.5. Data synthesis ................................................................................................9 3. Results . ...........................................................................................................9 3.1. Study selection ................................................................................................9 3.2. Study characteristics...........................................................................................9 3.3. Risk of bias within studies ......................................................................................11 Contents lists available atSciVerse ScienceDirect journal homepage:www.elsevier.com/locate/jad Journal of Affective Disorders 0165-0327/$ - see front matter&2012 Elsevier B.V. All rights reserved.

http://dx.doi.org/10.1016/j.jad.2012.06.007 nCorrespondence to: Department of Medical Psychology, University Medical Center Hamburg-Eppendorf, Martinistrasse 52, 20246 Hamburg, Germany. Tel:þ49 40 7410 58649; fax:þ49 40 7410 54965.

E-mail address:[email protected] (A. von Wolff). Journal of Affective Disorders 144 (2013) 7–15 3.4. Synthesis of results...........................................................................................11 3.5. Risk of bias across studies......................................................................................11 3.6. Subgroup analysis............................................................................................11 4. Discussion........................................................................................................12 4.1. Summary...................................................................................................12 4.2. Limitations . . ................................................................................................12 5. Conclusions.......................................................................................................13 Role of funding source...............................................................................................14 Conﬂict of interest ..................................................................................................14 Acknowledgments ..................................................................................................14 References........................................................................................................14 1. Introduction Approximately 20% of all patients who experience a major depressive episode develop a chronic course (Gilmer et al., 2005) and approximately 47% of patients who are treated in mental health care facility suffer from some form of chronic depression (Torpey and Klein, 2008). Four subtypes of chronic depression are usually distinguished: (1) dysthymia, (2) chronic major depres- sion, (3) recurrent major depression with incomplete remission between episodes, and (4) double depression (Klein, 2010).

Dysthymic disorder is deﬁned as a mild condition that is chronic and persistent for at least 2 years. A major depressive episode, chronic type, refers to a more severe condition that meets all criteria for major depression continuously for a minimum of 2 years. Patients who have recovered to the point where they no longer meet all criteria for a major depressive episode but continue to experience signiﬁcant symptoms for a total duration of illness greater than 2 years are referred to as having recurrent major depression with incomplete remission during episodes. The superimposition of a major depressive episode on antecedent dysthymia is referred to as double depression (Klein, 2010).

Chronic depression is associated with increased functional impairment (Satyanarayana et al., 2009), increased health care utilization, and higher rates of hospitalization compared with non- chronic forms of depression (Berndt et al., 2000;Gilmer et al., 2005).

Selective serotonin reuptake inhibitors (SSRIs) and tricyclic anti- depressants (TCAs) are two groups of antidepressants often used as pharmacological interventions in the treatment of chronic depression.

An increasing number of studies have assessed the efﬁcacy and effectiveness of these two groups of antidepressants in the last several decades.

These studies have been partly summarized in current guide- lines and systematic reviews. Meta-analytic ﬁndings regarding the treatment of dysthymic patients showed that both types of antidepressants are effective but highlighted the differential acceptability of the two drug groups (Lima et al., 2005). Accord- ingly, the practice guideline for the treatment of patients with major depressive disorder of the American Psychiatric Association (American Psychiatric Association, 2010) comes to the conclusion, that both types of antidepressants are effective in the treatment of dysthymia and may also be used for other forms of chronic depression. The national clinical practice guideline for the treat- ment and management of depression in adults of the National Institute for Health and Clinical Excellence (National Institute for Health and Clinical Excellence (NICE), 2009)further highlights that SSRIs are better tolerated than TCAs by patients suffering from subthreshold depressive symptoms (including dysthymia).

In contrast to these ﬁndings, the superiority of antidepressants to placebo was questioned for mild to moderately depressed patients in a controversially discussed meta-analysis (Kirsch et al., 2008). Thus, it remains unclear how chronicity and severity of depression inﬂuence the effectiveness of antidepressants. A systematic review considering the whole spectrum (i.e., all sub- types) of chronic depressions is still missing.The objectives of this systematic review are ﬁrst to examine the efﬁcacy and acceptability of SSRIs and TCAs in the treatment of chronic depression, and second, to assess the comparative effectiveness of the two types of antidepressants.

2. Methods The methods and results are reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta- Analyses (PRISMA) statement (Moher et al., 2009). Methods were speciﬁed a priori in a freely accessible review protocol, which included a detailed description of the methods, which are sum- marized brieﬂy here (Kriston et al., 2010).

2.1. Eligibility criteria Studies that were conducted in adults with a diagnosis of chronic major depression, dysthymia, double depression, or recurrent depression without complete remission between epi- sodes were included. The diagnosis of depression had to rely on a formal classiﬁcation system. Studies focusing on preselected samples (e.g., predeﬁned comorbidities) of chronically depressed patients were excluded. Pharmacological interventions that included the administration of either an SSRI or a TCA and that focused primarily on the treatment of depressive symptoms were considered. Only acute treatments (no maintenance or continuation treatments) were included. Somatic, psychotherapeutic, non-pharmacological, and organizational inter- ventions were not considered.

The comparator treatment required either the administration of a pharmacological placebo or another antidepressant pharma- cological intervention (SSRI or TCA).

Only studies that reported at least one outcome regarding the efﬁcacy of the interventions and only randomized controlled trials (RCTs) were included.

2.2. Search strategy An electronic database search was conducted in the following databases on January 18, 2010: Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, ISI Web of Science, BIOSIS, PsycINFO, and CINAHL. A disease component was combined (AND) with a design component for all searches:

(((chron$ adj3 depress$) or dysthym$ or (double adj1 depress$) or (treatment adj1 resist$ adj1 depress$) or (non adj1 respon$ adj3 depress$) or (recurrent adj3 depress$)).ab,ti,sh.) AND ((random$ or rct).ab,ti. or random$.sh.)(e.g., MEDLINE). No language restrictions were applied and all publications from 1970 forward were considered.

Additionally, all volumes of the Archives of General Psychiatry, the Journal of Consulting and Clinical Psychology, and the Journal of Affective Disorders were searched by hand beginning with the A. von Wolff et al. / Journal of Affective Disorders 144 (2013) 7–15 8 year 1970. The reference lists from all the studies retrieved and all relevant reviews were searched, followed by a cited reference search in the Social Sciences and Science Citation Index. The ﬁrst author of all included studies was contacted for further informa- tion regarding published and unpublished trials. Clinicaltrials.gov was searched for ongoing and unpublished trials.

2.3. Study selection Two reviewers independently screened the title and abstract of each identiﬁed article. In the next step, two reviewers inde- pendently examined the full texts of all potentially relevant studies using the inclusion criteria that were deﬁned a priori.

Disagreements were resolved through discussion.

2.4. Data collection and assessment of methodological quality A data extraction form for study characteristics, treatment characteristics, sample characteristics, and outcomes was developed.

The risk of bias was assessed in accordance with the Cochrane Collaboration’s Risk of Bias Tool (Higgins and Green, 2009). The global risk of bias was judged as ‘‘low’’ when at least four of the six Cochrane items were met and no indication of a high risk of bias could be detected. The global risk of bias was judged to be ‘‘high’’ when three or more of the Cochrane items were not met and there was and at least one of the unmet criteria suggested serious risk of bias.

Two researchers independently conducted the data extraction of outcome data and the assessment of risk of bias. Disagreements were resolved through discussion. The absolute proportion of corresponding judgments and Cohen’s kappa coefﬁcient were calculated to assess inter-rater agreement.

2.5. Data synthesis The primary efﬁcacy outcome was the response to treatment, which was deﬁned as an at least 50% decrease on a depression scale from baseline to the end of treatment. The secondary efﬁcacy outcome was the dichotomous outcome remission (e.g., reporting a severity score on a depression scale below a threshold of clinical signiﬁcance at the end of treatment). The primary acceptance outcome was dropping out of the study due to any reason. The secondary acceptance outcome was experiencing any adverse event. The primary endpoint was twelve weeks after the beginning of treatment. If no data were available for this end- point, any outcome data between 6 and 18 weeks after the beginning of treatment were accepted.

The statistical analysis followed actual guidelines (Egger et al., 2003;Higgins and Green, 2009;Khan et al., 2001). Effectiveness measures included beneﬁt ratios (BR; a ‘risk ratio’ for positive endpoints) for dichotomous effectiveness outcomes and odds ratios for dropout rates and rates of adverse events with corre- sponding 95% conﬁdence intervals (CI). Odds ratios were chosen because we expected dropout rates and rates of adverse events to be a rare outcome with highly varying baseline rates (Higgins and Green, 2009).

Effect sizes were calculated using the intention-to-treat prin- ciple for all studies. All randomized patients were included in the analyses of primary outcomes irrespective of how the authors of the primary studies deﬁned their intention-to-treat sample. For the primary efﬁcacy analysis (response) patients dropping out of the study were classiﬁed as non-responders. For secondary out- comes, the deﬁnition of the intention-to-treat sample provided by the authors of the primary studies was followed. In all analyses, a random effects model with inverse variance weights was applied (DerSimonian and Laird, 1986).Statistical heterogeneity between study results was tested for signiﬁcance using Cochran’s Q-test and quantiﬁed using theI 2 statistic (Higgins et al., 2003).

Possible publication bias was investigated using visual exam- ination of funnel plots and applying Egger’s test (Sterne et al., 2003).

An a priori deﬁned subgroup analysis was performed accord- ing to the subtype of chronic depression (studies that included only pure dysthymic patients vs. studies that included patients with other or mixed diagnoses of chronic depression). Addition- ally we performed a subgroup analysis according to the metho- dological quality of the studies (studies with a low risk of bias vs.

studies with an unclear or high risk of bias). Differences between subgroups were tested formally (Higgins and Green, 2009).

Analyses were performed using Review Manager 5 (The Nordic Cochrane Center, The Cochrane Collaboration, Copenhagen) and PASW Statistics 18 (SPSS, Inc., Chicago, IL) with an additional macro by Wilson (Wilson, 2011).

3. Results 3.1. Study selection After the removal of 1710 duplicates 2417 potentially relevant publications were identiﬁed through the electronic database search (Fig. 1). The number of publications was reduced to 275 after screening the title and abstract. Through the additional search strategies another 83 possibly relevant studies were identiﬁed. Of these 358 studies, seven studies were not available in full text. The full text of the available studies was screened and 20 primary studies, reported in 51 publications, with a total of 2918 participants, were included in the systematic review (Aguglia et al., 1995;Anisman et al., 1999;Baca et al., 2003; Bakish et al., 1993;Barrett et al., 2001;Boyer and Lecrubier, 1996; Devanand et al., 2005;Hellerstein et al., 2010;Hellerstein et al., 1994;Hellerstein et al., 1993;Keller et al., 1998;Kocsis et al., 1988;Ravindran et al., 1999;Ravindran et al., 2000;Rosenthal et al., 1992;Serrano-Blanco et al., 2006;Stewart et al., 1985; Thase et al., 1996;Tyrer et al., 1988;Vanelle et al., 1997;Versiani et al., 1997;Williams et al., 2000). The included studies comprise a total of 22 comparisons: nine comparing SSRIs to a placebo, another seven comparing TCAs to a placebo and six comparing SSRIs directly to TCAs.

3.2. Study characteristics The sample sizes ranged from 15 to 635. Nine of the 20 studies were conducted in the United States. Most studies focused on samples with a majority of dysthymic patients. One study included only patients with chronic major depression, one study included patients with chronic major depression and double depression and four studies included dysthymic patients as well as patients suffering from double depression. The average age of the study participants ranged from 34.0 to 69.9 years. Nearly all studies included a majority of women. All studies included out- patients and in two studies inpatients were additionally included.

Two antidepressants (amineptine and dothiepin) used in the included studies are currently not approved by the US Food and Drug Administration. The reported dosages of antidepressants were mainly in accordance with actual guidelines (American Psychiatric Association, 2010), the initial starting dosage was below the recommendations in four studies and no overdosages were reported. The detailed characteristics of each study are presented inTable 1. A. von Wolff et al. / Journal of Affective Disorders 144 (2013) 7–159 Fig. 1.Study ﬂow diagram. Table 1 Study characteristics.

Relevant study arms Diagnosis Mean age of sample% female in sampleDosage(mg)Treatment duration (weeks)N randCountry Aguglia et al. (1995)Clomipramine, Fluoxetine cMD 45.0 n.r. 75–150/ 20–404 48 Italy Anisman et al. (1999)Sertraline, Placebo Dys 40.5 52 50–200 12 68 Canada Baca et al. (2003)Imipramine, Sertraline Dys, DD 40.0 79 75–225/ 50–2008 121 Spain Bakish et al. (1993)Imipramine, Placebo Dys 37.6 52 50–200 7 33 Canada Barrett et al. (2001)/ Williams et al. (2000)Paroxetine, Placebo Dys 61.0 47 10–40 11 223 USA Boyer and Lecrubier (1996)Amineptine, Placebo Dys 48.0 75 200 12 219 France Devanand et al. (2005)Fluoxetine, Placebo Dys 69.9 37 10–60 12 90 USA Hellerstein et al. (1993)Fluoxetine, Placebo Dys 36.2 50 10–60 8 35 USA Hellerstein et al. (1994)/ Rosenthal et al. (1992)Fluoxetine, Trazodone Dys, DD 40.2 74 20–60/ 50–35012 38 USA Hellerstein et al. (2010)Escitalopram, Placebo Dys 44.7 50 10–20 12 34 USA Kocsis et al. (1988)Imipramine, Placebo Dys, DD 39.0 70 100–300 6 54 USA Ravindran et al. (1999)Sertraline, Placebo Dys 38.0 58 50–200 12 48 Canada Ravindran et al. (2000)Sertraline, Placebo Dys 45.1 67 50–200 12 310 Canada, France, Italy, Spain, Sweden, UK Rush et al. (1998)/ Keller et al. (1998)Imipramine, Sertraline cMD, DD 41.1 63 50–300/ 50–20012 635 USA Serrano-Blanco et al. (2006)Fluoxetine, Imipramine Dys 43.5 73 no dosage rec. 12 (24) 15 Spain Stewart et al. (1985)Desipramine, Placebo Dys n.r. n.r. 50–300 6 25 USA Thase et al. (1996)Imipramine, Sertraline, PlaceboDys 41.7 65 50–300/ 50–20012 410 USA Tyrer et al. (1988)Dothiepin, Placebo Dys 34.0 73 25–150 6,10 19 UK Vanelle et al. (1997)Fluoxetine, Placebo Dys 43.0 76 20 12 140 France Versiani et al. (1997)Imipramine, Placebo Dys, DD 41.5 70 25–250 8 207 Argentina, Brazil, Chile Nrand¼number of patients randomized to relevant study arms; Dys¼dysthymia; DD¼double depression; cMD¼chronic major depression; n.r.¼not reported; no dosage rec¼no dosage recommendation.A. von Wolff et al. / Journal of Affective Disorders 144 (2013) 7–15 10 3.3. Risk of bias within studies The overall risk of bias was evaluated as ‘‘low’’ for four studies, as ‘‘unclear’’ for ten studies and six studies were judged to have a ‘‘high’’ risk of bias. Knowledge of allocation was adequately prevented in nearly all studies (17/20), whereas an adequate concealment of the allocation was reported in only four of the studies. The results of the methodological quality assessment are presented for individual studies inTable 2. Inter-rater agreement for single quality assessment items ranged from 15/20 to 20/20 (corresponding kappa coefﬁcients ranged from 0.17 (other sources of bias) to 1.0 (allocation sequence adequately generated and allocation adequately concealed)).

3.4. Synthesis of results SSRIs yielded statistically signiﬁcant better response and remission rates than placebo (response: BR¼1.49; 95% CI: 1.29– 1.72;po0.001; remission: BR¼1.53; 95% CI: 1.29–1.80; po0.001). This result corresponds with a number needed to treat (NNT) of six for response and seven for remission. These results were not heterogeneous across studies (I 2¼0%).

In terms of the negative outcomes - dropout rates and rates of adverse events - no differences between SSRIs and placebo could be detected (dropout: OR¼0.81; 95% CI: 0.49–1.33;p¼0.41; adverse event: OR¼1.21; 95% CI: 0.77–2.17;p¼0.52), yet increased statistical heterogeneity among the results ofthe primary studies was observed for these outcomes (I 2¼52% for dropout;I 2¼43% for adverse event).

TCAs proved to be superior to placebo regarding response and remission rates (response: BR¼1.74; 95% CI: 1.50–2.02; po0.001; remission: BR¼1.77; 95% CI: 1.24–2.53;p¼0.002).

This result corresponds with a NNT of four for response and seven for remission. The results of the primary studies were not heterogeneous across studies for response rates (I 2¼0%) but some heterogeneity was observed for rates of remission (I 2¼51%). No statistically signiﬁcant difference between TCA and placebo was found in relation to dropout rates (OR¼1.14; 95% CI: 0.74–1.78; p¼0.55) but TCAs showed signiﬁcantly higher rates of patients experiencing adverse events than placebo (OR¼2.87; 95% CI: 1.83– 4.50;po0.001).No differences in efﬁcacy outcomes could be shown between SSRIs and TCAs (response: BR¼1.01; 95% CI: 0.84–1.21;p¼0.91; remission: BR¼1.02; 95% CI: 0.81–1.28;p¼0.87). However, SSRIs showed statistically signiﬁcant lower dropout rates and rates of adverse events than TCAs (dropout: OR¼0.41; 95% CI: 0.19–0.86; p¼0.02; adverse event: OR¼0.48; 95% CI: 0.32–0.70;po0.001).

This result corresponds with a NNT of 9 for dropouts and 11 for adverse events. Heterogeneity among the results of the primary studies was low for rates of adverse events (I 2¼1%) but substantial for dropout rates (I 2¼71%).

Results for each individual study describing each outcome as well as overall effects are displayed inTable 3. Forest plots are displayed inFig. 2.

3.5. Risk of bias across studies No indication of publication bias for the primary efﬁcacy out- come response to treatment by visual examination of the funnel plots was observed. Accordingly, no indication for publication bias was found when applying Egger’s test (SSRI vs. placebo: beta¼0.58; p¼0.11 TCA vs. placebo: beta¼0.26;p¼0.62; SSRI vs. TCA: beta¼- 0.07;p¼0.89).

3.6. Subgroup analysis As all studies comparing SSRIs to placebo included only pure dysthymic patients no further subgroup analyses according to diagnosis could be performed for this comparison. For the two comparisons - TCA vs. placebo and SSRI vs. TCA - subgroup analyses were performed regarding all outcomes (response, remission, dropout and adverse event). No statistically signiﬁcant differences between studies including pure dysthymic patients and studies that included patients with other or mixed diagnoses of chronic depression could be found for any outcome (see Table 4).

Subgroup analyses for the primary outcomes response and dropout comparing studies with a low risk of bias with studies with an unclear or high risk of bias did not reveal any systematic bias (seeTable 4). In one subgroup analysis (SSRI vs. placebo) studies with an unclear or high risk of bias showed statistically Table 2 Risk of bias in individual studies.

1 2 3 4 5 6 global judgment Aguglia et al. (1995)unclear unclear yes unclear no no high Anisman et al. (1999)yes unclear yes yes yes yes low Baca et al., 2003unclear unclear no yes yes yes high Bakish et al. (1993)unclear unclear yes unclear unclear unclear high Barrett et al. (2001)/Williams et al. (2000)yes yes yes yes yes yes low Boyer and Lecrubier (1996)unclear unclear yes yes yes yes unclear Devanand et al. (2005)yes yes yes yes yes yes low Hellerstein et al. (1993)unclear unclear yes yes yes yes unclear Hellerstein et al. (1994)/Rosenthal et al. (1992)unclear unclear no no yes yes high Hellerstein et al. (2010)unclear unclear yes unclear yes yes unclear Kocsis et al. (1988)unclear unclear yes no unclear unclear high Ravindran et al. (1999)yes unclear yes yes no yes unclear Ravindran et al., 2000unclear unclear yes yes yes yes unclear Rush et al. (1998)/Keller et al. (1998)unclear unclear yes yes unclear yes unclear Serrano-Blanco et al. (2006)yes yes no yes yes yes unclear Stewart et al. (1985)unclear unclear yes unclear yes unclear high Thase et al. (1996)yes unclear yes yes yes unclear low Tyrer et al. (1988)unclear yes yes yes no yes unclear Vanelle et al. (1997)unclear unclear yes yes yes yes unclear Versiani et al. (1997)unclear unclear yes yes yes yes unclear Summary ‘‘yes’’ 6/20 4/20 17/20 14/20 14/20 15/20 4/20 1: Allocation sequence adequately generated; 2: allocation adequately concealed; 3: knowledge of allocation adequately prevented; 4: incompleteoutcome adequately addressed; 5: free of selective outcome reporting; 6: free of other problems.A. von Wolff et al. / Journal of Affective Disorders 144 (2013) 7–1511 signiﬁcant larger effects, yet in this comparison only one study was judged to have a low risk of bias (Anisman et al., 1999).

4. Discussion 4.1. Summary Similar results were obtained in terms of efﬁcacy for both SSRIs and TCAs. Both classes of drugs showed statistically sig- niﬁcant higher rates of response and remission than their placebo control treatments. The low NNTs assessed for both SSRIs and TCAs further highlight the clinical relevance of our ﬁndings, more precisely between four and seven patients have to be treated to cause one additional clinical improvement. No differences in terms of efﬁcacy could be detected in studies directly comparing SSRIs and TCAs. In terms of acceptability, results demonstrated the superiority of SSRIs, as TCAs yielded signiﬁcantly higher rates of dropouts and adverse events.

Heterogeneity among the results of the primary studies was lower for efﬁcacy outcomes (response and remission) than for outcomes of acceptability (dropouts and adverse events).

The ﬁnding that SSRIs and TCAs do not differ in terms of efﬁcacy, but SSRIs tend to be superior in terms of acceptability, is consistent with previous ﬁndings for patients suffering from dysthymia (Lima et al., 2005;National Institute for Health and Clinical Excellence (NICE), 2009). Our ﬁndings further highlight the relevance of antidepressants in the treatment of patients with mild but chronic forms of depression in contrast to previous ﬁndings questioning the superiority of antidepressant to placebo for mild to moderately depressed patients (Kirsch et al., 2008).Subgroup analyses comparing studies with pure dysthymic patients to studies with other forms of chronically depressed patients did not indicate any differences for efﬁcacy or acceptability according to diagnostic subgroups.

4.2. Limitations Several limitations of the current literature emerged upon this review. The sample sizes of the primary studies varied greatly and the methodological quality of the primary studies was evaluated as unclear in many cases. Even though we could not detect any systematic bias, statistical testing when applying Egger’s test and conducting subgroup analyses may have been underpowered due to the limited number of studies. Additionally, not all studies that were suitable for the systematic review reported all relevant outcomes.

Therefore, our results are partly based on a small number of studies (e.g., only four studies that compared SSRIs to TCAs reported dropout rates) and consequently we could not further explore heterogeneity for these outcomes. Another shortcoming of this systematic review is the limited search for unpublished trials (gray literature). The inclusion of unpublished trials in systematic reviews is a controversially discussed issue: on one hand it may reduce bias as unpublished studies can have smaller effect sizes than published studies (Turner et al., 2008). On the other hand the inclusion of data from unpublished studies can itself introduce bias as unpublished studies may be of lower methodological quality than published studies (Higgins and Green, 2009). Due to these shortcomings, some uncertainty about the reliability of our results persists.

Our systematic literature review also highlighted a lack of relevant studies providing evidence for the efﬁcacy of SSRIs and TCAs in patients suffering from chronic forms of depression other Table 3 Results of all outcomes (individual studies and overall effect).

Response Remission Dropout Adverse Events BR (95% CI) BR (95% CI) OR (95% CI) OR (95% CI) SSRI vs. Placebo Anisman et al. (1999)2.23 (1.27, 3.94) 2.59 (1.16, 5.77) 0.23 (0.06, 0.94) 0.20 (0.04, 1.04) Barrett et al. (2001)/Williams et al. (2000)1.31 (0.86, 1.99) 1.33 (0.97, 1.81) 1.66 (0.83, 3.32) not reported Devanand et al. (2005)1.39 (0.65, 2.98) 1.24 (0.62, 2.46) 2.09 (0.74, 5.93) 2.22 (0.82, 5.98) Hellerstein et al. (1993)2.81 (0.93, 8.48) 2.50 (0.57, 11.05) 7.00 (0.33, 146.45) 4.00 (0.40, 40.11) Hellerstein et al. (2010)1.40 (0.55, 3.55) 4.00 (0.50, 32.20) 0.47 (0.04, 5.72) 0.54 (0.04, 6.58) Ravindran et al. (1999)1.77 (0.89, 3.54) 2.95 (1.08, 8.12) 0.22 (0.01, 4.79) not reported Ravindran et al. (2000)1.54 (1.17, 2.03) 1.56 (1.07, 2.28) 0.92 (0.55, 1.54) 1.79 (0.85, 3.78) Thase et al. (1996)1.33 (1.05, 1.68) 1.46 (1.08, 1.98) 0.58 (0.32, 1.06) 1.71 (0.55, 5.38) Vanelle et al. (1997)1.62 (0.98, 2.65) 1.72 (0.93, 3.20) 0.42 (0.17, 1.01) 0.77 (0.38, 1.55) overall effect (p) 1.49 (1.29, 1.71) (o0.001) 1.53 (0.29, 1.80) (o0.001) 0.81 (0.49, 1.33) (0.41) 1.21 (0.77, 2.17) (0.52) I 2(p) 0% (0.77) 0% (0.66) 52% (0.03) 43% (0.11) TCA vs. Placebo Bakish et al. (1993)1.99 (1.18, 3.35) 2.28 (1.27, 4.07) 0.65 (0.16, 2.72) not reported Boyer and Lecrubier (1996)1.92 (1.41, 2.61) 2.26 (1.50, 3.40) 0.86 (0.44, 1.68) 2.05 (1.20, 3.52) Kocsis et al. (1988)2.16 (0.99, 4.71) 3.74 (1.20, 11.63) 7.64 (0.87, 67.13) 4.64 (0.21, 101.26) Stewart et al. (1985)1.00 (0.18, 5.63) not reported not reported not reported Thase et al. (1996)1.44 (1.15, 1.81) 1.21 (0.88, 1.67) 1.54 (0.91, 2.61) 6.08 (2.25, 16.41) Tyrer et al. (1988)3.75 (0.20, 68.89) 3.75 (0.20, 68.89) 0.22 (0.01, 6.07) not reported Versiani et al. (1997)2.15 (1.55, 2.99) 1.23 (0.67, 2.24) 1.01 (0.47, 2.19) 2.93 (1.57, 5.47) overall effect (p) 1.74 (1.50, 2.02) (o0.001) 1.77 (1.24, 2.53) (0.002) 1.14 (0.74, 1.78) (0.55) 2.87 (1.83, 4.50) (o0.001) I 2(p) 0% (0.44) 51% (0.07) 23% (0.26) 20% (0.29) SSRI vs. TCA Aguglia et al. (1995)0.73 (0.43, 1.25) 0.60 (0.16, 2.23) not estimable 0.67 (0.19, 2.34) Baca et al. (2003)1.32 (0.93, 1.88) 1.31 (0.72, 2.38) not reported not reported Hellerstein et al. (1994)/Rosenthal et al. (1992)1.80 (0.76, 4.27) not reported 0.11 (0.02, 0.53) 0.53 (0.34, 0.84) Rush et al. (1998)/Keller et al. (1998)1.02 (0.87, 1.20) 0.89 (0.73, 1.10) 0.69 (0.46, 1.03) 0.28 (0.12, 0.65) Serrano-Blanco et al. (2006)0.25 (0.04, 1.71) 0.33 (0.04, 2.48) not reported not reported Thase et al. (1996)0.92 (0.76, 1.11) 1.21 (0.91, 1.59) 0.38 (0.21, 0.68) not reported overall effect (p) 1.01 (0.84, 1.21) (0.91) 1.02 (0.81, 1.28) (0.87) 0.41 (0.19, 0.86) (0.02) 0.48 (0.32, 0.70) (o0.001) I 2(p) 39% (0.15) 26% (0.25) 71% (0.03) 1% (0.36) SSRI¼selective serotonin reuptake inhibitor; TCA¼tricyclic antidepressant; BR¼Beneﬁt Ratio; CI¼conﬁdence interval; OR¼odds ratio;p¼level of signiﬁcance; I 2¼measure of heterogeneity.A. von Wolff et al. / Journal of Affective Disorders 144 (2013) 7–15 12 than dysthymia. We could not identify any study including patients with recurrent depression without complete remission between episodes and only two of the studies included patients suffering from chronic major depression. Further, all studies comparing SSRIs to a placebo included only pure dysthymic patients. Though our results - based on the present evidence - indicated no differences between the diagnostic subgroupsin treatment response, we cannot rule out the possibility that patients with forms of chronic depres- sion other than dysthymia might respond differently to SSRIs and TCAs. As patients suffering from forms of chronic depression other than dysthymia may differ remarkably from patients with dysthy- mia in terms of the severity of depressive symptoms further evidence on the efﬁcacy and acceptability of the two antidepres- sants for these subgroups is needed. This evidence is required todevelop treatment guidelines considering both the chronicity and the severity of the disorder.

Another shortcoming of our review is that the evidence presented is based on clinical trials, which focus on high internal validity, thus evidence from routine mental health care is needed to evaluate the effectiveness conclusively for both groups of drugs. In addition, more research on the long-term effects of the medications is required because of the high risk of relapse in chronic depression.

5. Conclusions This systematic review provides evidence for the efﬁcacy of SSRIs and TCAs in the treatment of chronic depression. As both Fig. 2.Forest plots of beneﬁts ratios (response) and odds ratios (dropout) for the three comparisons.A. von Wolff et al. / Journal of Affective Disorders 144 (2013) 7–1513 groups of drugs lead to promising results, clinicians may make treatment decisions about the best option based upon the patients’ personal needs and drug tolerance. As TCAs are more likely to produce dropouts and adverse events than SSRIs, SSRIs seem to be the ﬁrst choice for the treatment of chronic depres- sion. These ﬁndings may serve as supplementary information for present clinical guidelines (American Psychiatric Association, 2010;National Institute for Health and Clinical Excellence (NICE), 2009).

Furthermore, the efﬁcacy of both SSRIs and TCAs for the treatment of forms of chronic depression other than dysthymia needs to be conﬁrmed.

Further research on the cost-effectiveness, long-term outcomes and patient related outcomes (such as quality of life) in treating chronically depressed patients is required to conclusively assess the effectiveness of SSRIs and TCAs.

Role of funding source This study was funded by a grant from the German Ministry of Education and Research (project 01KG0923).

Conﬂict of interest All other authors declare that they have no conﬂicts of interest.

Acknowledgments None. References Aguglia, E., Devanna, M., Digraza, M.F., 1995. Role of selective serotonin reuptake inhibitors on the treatment of chronic and/or resistant major depression.

Giornale Di Neuropsicofarmacologia 17, 73–77.

American Psychiatric Association, 2010. Practice Guideline for the Treatment of Patients with Major Depressive Disorder, third ed. American Psychiatric Association, Arlington (VA).

Anisman, H., Ravindran, A.V., Grifﬁths, J., Merali, Z., 1999. Interleukin-1 beta production in dysthymia before and after pharmacotherapy. Biological Psy- chiatry 46, 1649–1655.

Baca, E., de Chavez, M.G., Garcia-Toro, M., Perez-Arnau, F., Porras-Chavarino, A., 2003. Sertraline is more effective than imipramine in the treatment of non- melancholic depression: Results from a multicentre, randomized study. Pro- gress in Neuro-Psychopharmacology and Biological Psychiatry 27, 493–500.

Bakish, D., Lapierre, Y.D., Weinstein, R., Klein, J., Wiens, A., Jones, B., Horn, E., Browne, M., Bourget, D., Blanchard, A., Thibaudeau, C., Waddell, C., Raine, D., 1993. Ritanserin, imipramine and placebo in the treatment of dysthymic disorder. Journal of Clinical Psychopharmacology 13, 409–414.

Barrett, J.E., Williams, J.W., Oxman, T.E., Frank, E., Katon, W., Sullivan, M., Hegel, M.T., Cornell, J.E., Sengupta, A.S., 2001. Treatment of dysthymia and minor depression in primary care: a randomized trial in patients aged 18 to 59 years.

Journal of Family Practice 50, 405–412.

Berndt, E.T., Koran, L.M., Finkelstein, S.N., Gelenberg, A.J., Kornstein, S.G., Miller, I.W., Thase, M.E., Trapp, G.A., Keller, M.B., 2000. Lost human capital from early- onset chronic depression. American Journal of Psychiatry 157, 940–947.

Boyer, P., Lecrubier, Y., 1996. Atypical antipsychotic drugs in dysthymia: placebo controlled studies of amisulpride versus imipramine, versus amineptine.

European Psychiatry 11, S135–S140.

DerSimonian, R., Laird, N., 1986. Meta-analysis in clinical trials. Controlled Clinical Trials 7, 177–188.

Devanand, D.P., Nobler, M.S., Cheng, J., Turret, N., Pelton, G.H., Roose, S.P., Sackeim, H.A., 2005. Randomized, double-blind, placebo-controlled trial of ﬂuoxetine treatment for elderly patients with dysthymic disorder. American Journal of Geriatric Psychiatry 13, 59–68.

Egger, M., Smith, G.D., Altman, D.G., 2003. Systematic reviews in health care, Meta- Analysis in Context. BMJ Publishing Group, London. Table 4 Subgroup analyses.

BR/OR (95% CI)I 2within subgroup w2(df)p SSRI vs. Placebo Response0.81 (1) 0.37 Low risk of bias (4) n 1.41 (1.17, 1.69) 0% Unclear/ high risk of bias (5) 1.60 (1.29, 1.99) 0% Dropout0.20 (1) 0.66 Low risk of bias (4) 0.89 (0.39, 2.06) 73% Unclear/ high risk of bias (5) 0.71 (0.38, 1.31) 21% TCA vs. Placebo Response2.45 (1) 0.12 Pure dysthymic (5) 1.63 (1.37, 1.93) 0% Mixed diagnosis (2) 2.15 (1.59, 2.91) 0% Dropout0.43 (1) 0.51 Pure dysthymic (4) 1.10 (0.69, 1.75) 14% Mixed diagnosis (2) 2.13 (0.32, 14.41) 66% Response4.74 (1) 0.03 Low risk of bias (1) 1.44 (1.15, 1.81) Not applicable Unclear/ high risk of bias (6) 2.01 (1.65, 2.45) 0% Dropout1.48 (1) 0.22 Low risk of bias (1) 1.54 (0.91, 2.61) Not applicable Unclear/ high risk of bias (5) 0.96 (0.56, 1.66) 14% SSRI vs. TCA Response0.64 (1) 0.42 Pure dysthymic (2) 0.69 (0.24, 1.99) 43% Mixed diagnosis (4) 1.08 (0.85, 1.38) 40% Dropout0.02 0.88 Pure dysthymic (1) 0.38 (0.21, 0.68) Not applicable Mixed diagnosis (3) 0.33 (0.06, 1.89) 80% Response0.63 (1) 0.43 Low risk of bias (1) 0.92 (0.76, 1.11) Not applicable Unclear/ high risk of bias (5) 1.05 (0.80, 1.39) 44% Dropout0.02 (1) 0.88 Low risk of bias (1) 0.38 (0.21, 0.68) Not applicable Unclear/ high risk of bias (3) 0.33 (0.06, 1.89) 80% SSRI¼selective serotonin reuptake inhibitor; TCA¼tricyclic antidepressant; n¼numbers in brackets indicated the number of studies within the subgroup; BR¼Beneﬁt Ratio [for response]; CI¼conﬁdence interval; OR¼odds ratio [for dropout];w2¼chi-square statistic for difference testing between subgroups, df¼degrees of freedom for difference testing between subgroups;p¼level of signiﬁcance of difference testing between subgroups;I 2¼measure of heterogeneity. A. von Wolff et al. / Journal of Affective Disorders 144 (2013) 7–15 14 Gilmer, W.S., Trivedi, M.H., Rush, A.J., Wisniewski, S.R., Luther, J., Howland, R.H., Yohanna, D., Khan, A., Alpert, J., 2005. Factors associated with chronic depressive episodes: a preliminary report from the STAR-D project. Acta Psychiatrica Scandinavica 112, 425–433.

Hellerstein, D.J., Batchelder, S.T., Hyler, S., Arnaout, B., Toba, C., Benga, I., Gangure, D., 2010. Ecitalopram versus placebo in the treatment of dysthymic disorder.

International Clinical Psychopharmacology 25, 143–148.

Hellerstein, D.J., Yanowitch, P., Rosenthal, J., Hemlock, C., Kasch, K., Samstag, L.W., Winston, A., 1994. Lomg-term treatment of double depression - a preliminary- study with serotonergic antidepressants. Progress in Neuro-Psychopharma- cology and Biological Psychiatry 18, 139–147.

Hellerstein, D.J., Yanowitch, P., Rosenthal, J., Samstag, L.W., Maurer, M., Kasch, K., Burrows, L., Poster, M., Cantillon, M., Winston, A., 1993. A randomized double- blind study of ﬂuoxetine versus placebo in the treatment of Dysthymia.

American Journal of Psychiatry 150, 1169–1175.

Higgins, J.P.T., Green, S. (Eds.), 2009. Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.2 [updated September 2009]. The Cochrane Collaboration.

Higgins, J.P.T., Thompson, S.G., Deeks, J.J., Altman, D.G., 2003. Measuring incon- sistency in meta-analyses. BMJ—British Medical Journal 327, 557–560.

Keller, M.B., Gelenberg, A.J., Hirschfeld, R.M.A., Rush, A.J., Thase, M.E., Kocsis, J.H., Markowitz, J.C., Fawcett, J.A., Koran, L.M., Klein, D.N., Russell, J.M., Kornstein, S.G., McCullough, J.P., Davis, S.M., Harrison, W.M., 1998. The treatment of chronic depression, part 2: a double-blind, randomized trial of sertraline and imipramine. Journal of Clinical Psychiatry 59, 598–607.

Khan, K.S., Ter Riet, G., Glanville, J., Sowden, A.J., Kleijnen, J., 2001. Undertaking Systematic Reviews of Research on Effectiveness, CRD’s Guidance for those Carrying Out or Commissioning Reviews. CRD Report Number 4, 2nd Edition.

NHS Centre for Reviews and Dissemination, University of York, Heslington, York.

Kirsch, I., Deacon, B.J., Heudo-Medina, T.B., Scoboria, A., Moore, T.J., Johnson, B.T., 2008. Initial severity and antidepressant beneﬁts: a meta-analysis of data submitted to the food and drug administration. PLoS Medicine 5, e45, http://dx.

doi.org/10.1371/journal.pmed.0050045.

Klein, D.N., 2010. Chronic depression: diagnosis and classiﬁcation. Current Direc- tions in Psychological Science 19, 96–100.

Kocsis, J.H., Frances, A.J., Voss, C., Mason, B.J., Mann, J.J., Sweeney, J., 1988.

Imipramine and social-vocational adjustment in chronic depression. American Journal of Psychiatry 145, 997–999.

Kriston, L., von Wolff, A., H¨ olzel, L.P., 2010. Effectiveness of psychotherapeutic, pharmacological, and combined treatments for chronic depression: a systematic review (METACHRON). BMC Psychiatry 23 (10), 95.

Lima, M.S., Moncrieff, J., Soares, B.G.O., 2005. Drugs versus placebo for dysthymia.

Cochrane Database System Review, art. no.: CD001130. http://doi:10.1002/ 14651858.CD001130.

Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., 2009. the PRISMA Group, 2009.

Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Annals of Internal Medicine 151, 264–269.

National Institute for Health and Clinical Excellence (NICE), 2009. Depression: The Treatment and Management of Depression in Adults. National Clinical Practice Guideline 90. National Institute for Health and Clinical Excellence (NICE), London.

Ravindran, A.V., Anisman, H., Merali, Z., Charbonneau, Y., Telner, J., Bialik, R.J., Wiens, A., Ellis, J., Grifﬁths, J., 1999. Treatment of primary dysthymia with group cognitive therapy and pharmacotherapy: clinical symptoms and func- tional impairments. American Journal of Psychiatry 156, 1608–1617.Ravindran, A.V., Guelﬁ, J.D., Lane, R.M., Cassano, G.B., 2000. Treatment of dysthy- mia with sertraline: a double-blind, placebo-controlled trial in dysthymic patients without major depression. Journal of Clinical Psychiatry 61, 821–827.

Rosenthal, J., Hemlock, C., Hellerstein, D.J., Yanowitch, P., Kasch, K., Schupak, C., Samstag, L., Winston, A., 1992. A preliminary study of serotonergic antide- pressants in treatment of Dysthymia. Progress in Neuro-Psychopharmacology and Biological Psychiatry 16, 933–941.

Rush, A.J., Koran, L.M., Keller, M.B., Markowitz, J.C., Harrison, W.M., Miceli, .R.J, Fawcett, J.A., Gelenberg, A.J., Hirschfeld, R.M., Klein, D.N., Kocsis, J.H., McCul- lough, J.P., Schatzberg, A.F., Thase, M.E., 1998. The treatment of chronic depression, part 1: study design and rationale for evaluating the comparative efﬁcacy of sertraline and imipramine as acute, crossover, continuation, and maintenance phase therapies. Journal of Clinical Psychiatry 59, 589–597.

Satyanarayana, S., Enns, M.W., Cox, B.J., Sareen, J., 2009. Prevalence and correlates of chronic depression in the Canadian Community Health Survey: Mental health and well-being. Canadian Journal of Psychiatry 54, 389–398.

Serrano-Blanco, A., Gabarron, E., Garcia-Bayo, I., Soler-Vila, M., Carames, E., Penarru- bia-Maria, M.T., Pinto-Meza, A., Haro, J.M., grp, D., 2006. Effectiveness and cost- effectiveness of antidepressant treatment in primary health care: a six-month randomised study comparing Fluoxetine to Imipramine. Journal of Affective Disorders 91, 153–163.

Sterne, J.A.C., Egger, M., Smith, G.D., 2003. Investigating and dealing with publica- tion and other biases. In: Egger, M., Smith, G.D., Altman, D.G. (Eds.), Systematic reviews in health care: meta-analysis in context. BMJ Publishing Group, London, pp. 248–282.

Stewart, J.W., McGrath, P.J., Liebowitz, M.R., Harrison, W., Quitkin, F., Rabkin, J.G., 1985. Treatment outcome validation of DSM-III depressive subtypes: clinical usefulness in outpatients with mild to moderate depression. Archives of General Psychiatry 42, 1148–1153.

Thase, M.E., Fava, M., Halbreich, U., Kocsis, J.H., Koran, L., Davidson, J., Rosenbaum, J., Harrison, W., 1996. A placebo-controlled, randomized clinical trial compar- ing Sertraline and Imipramine for the treatment of dysthymia. Archives of General Psychiatry 53, 777–784.

Torpey, D.C., Klein, D.N., 2008. Chronic depression: update on classiﬁcation and treatment. Current Psychiatry Reports 10, 458–464.

Turner, E.H., Matthews, A.M., Linardatos, E., Tell, R.A., Rosenthal, R., 2008. Selective publication of antidepressant trials and its inﬂuence on apparent efﬁcacy. New England Journal of Medicine 358, 252–260.

Tyrer, P., Murphy, S., Kingdon, D., Brothwell, J., Gregory, S., Seivewright, N., Ferguson, B., Barczak, P., Darling, C., Johnson, A.L., 1988. The Nottingham study of neurotic disorder: comparison of drug and psychological treatments.

Lancet 2, 235–240.

Vanelle, J.M., AttarLevy, D., Poirier, M.F., Bouhassira, M., Blin, P., Olie, J.P., 1997.

Controlled efﬁcacy study of Fluoxetine in dysthymia. British Journal of Psychiatry 170, 345–350.

Versiani, M., Amrein, R., Stabl, M., Magistris, Udabe, R.U., DaCosta, D.A., Nardi, A.E., Capponi, R., 1997. Moclobemide and imipramine in chronic depression (dysthymia): an international double-blind, placebo-controlled trial. Interna- tional Clinical Psychopharmacology 12, 183–193.

Williams, J.W., Barrett, J., Oxman, T., Frank, E., Katon, W., Sullivan, M., Cornell, J., Sengupta, A., 2000. Treatment of dysthymia and minor depression in primary care: a randomized controlled trial in older adults. Jama—Journal of the American Medical Association 284, 1519–1526.

Wilson, D.B., 2010. SPSS, Stata, and SAS macros for performing meta-analytic analyses.

/http://mason.gmu.edu/ dwilsonb/ma.htmlS, (accessed 31.10.11). A. von Wolff et al. / Journal of Affective Disorders 144 (2013) 7–1515 View publication statsView publication stats