Author's Accepted Manuscript
Efficacy and tolerability of antidepressants for sub-threshold depression and for mild major depressive disorder Isobel M Cameron, Ian C Reid, Steve A MacGillivray
www.elsevier.com/locate/jad
PII: DOI: Reference:
S0165-0327(14)00272-9 http://dx.doi.org/10.1016/j.jad.2014.04.078 JAD6745
To appear in:
Journal of Affective Disorders
Received date: 11 November 2013 Revised date: 28 April 2014 Accepted date: 29 April 2014 Cite this article as: Isobel M Cameron, Ian C Reid, Steve A MacGillivray, Efficacy and tolerability of antidepressants for sub-threshold depression and for mild major depressive disorder, Journal of Affective Disorders, http://dx.doi.org/10.1016/j. jad.2014.04.078 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Title Efficacy and tolerability of antidepressants for sub-threshold depression and for mild major depressive disorder Authors Isobel M Cameron, Applied Medical Sciences (Psychiatry), University of Aberdeen, Royal Cornhill Hospital, Aberdeen, AB25 2ZH, Scotland.
Ian C Reid, Applied Medical Sciences (Psychiatry), University of Aberdeen, Royal Cornhill Hospital, Aberdeen, AB25 2ZH, Scotland. Steve A MacGillivray*, Social Dimensions of Health Institute, University of Dundee, DD1 4HJ, Scotland.
*Correspondence to: Steve MacGillivray, Social Dimensions of Health Institute, University of Dundee, DD1 4HJ. Tel: +44 (0)1382 381075; Fax: +44 (0)1382 388533; email: [email protected]
1
Abstract Background: It is commonly assumed that robust evidence exists for the lack of efficacy of antidepressants at the milder end of the depression severity spectrum. In light of specific limitations of existing reviews, we assess if antidepressants are efficacious and tolerable for sub-threshold or mild Major Depressive Disorder (MDD).
Method: Systematic review and meta-analysis of Randomised Controlled Trials of adults with sub-threshold depression or mild MDD (initial baseline symptom severity of HRSD20) comparing an antidepressant with placebo or treatment as usual (TAU) however defined. A pre-specified protocol was published (Prospero reference: CRD42013004505). Results: 8 trials were included: 5 trials (453 participants) of sub-threshold depression and 3 trials (502 participants) of mild MDD. Trials of sub-threshold depression exhibited low risk of bias whereas those of mild MDD exhibited high risk. Two trials of sub-threshold depression were pooled (n=102) to assess efficacy and favoured antidepressants over placebo statistically but the difference was small and unlikely to be clinically meaningful: mean difference -1.39 (-2.41, -0.36). Due to heterogeneity, no trials of mild MDD could be pooled for efficacy. There was no difference between antidepressant treatment and placebo for drop out due to adverse events. The maximum proportion in those receiving antidepressants dropping out due to adverse events was 17%, with indication of a dose effect.
Limitations: Not all data from identified trials could be included in the meta-analyses due to a lack of availability of relevant data.
Conclusion: There is insufficient evidence to support or contest the efficacy of antidepressant medication for sub-threshold depression or mild MDD. More trials, with adequate follow up, are required to address this question. Keywords: systematic review, meta-analysis, depressive disorder, antidepressants
2
Introduction
Establishing the efficacy of antidepressants for milder forms of depression is important because diagnoses such as minor, sub-threshold or mild depression have been found to be as prevalent as more severe forms (Ackermann, Williams 2002, Oxman, Sengupta 2002). Such diagnoses are associated with considerable functional impairment (Rapaport et al. 2002, Broadhead et al. 1990) and are linked with increased risk of developing a major depressive disorder (Kessler et al. 1997).
Robust evidence supports treatment of moderate to severe depression with antidepressant drugs (Bauer et al. 2007, NHS National Institute for Clinical Excellence 2009, Nutt 2011). It is commonly assumed that there is equally robust evidence for their lack of efficacy at the milder end of the severity spectrum (Nutt 2011, Baumeister 2012, Godlee 2013). Guidelines (Bauer et al. 2007, NHS National Institute for Clinical Excellence 2009, Nutt 2011) do not currently recommend the routine use of antidepressants for those with sub-threshold depression (also referred to as Minor Depression - MinD) or Mild Major Depressive Disorder (Mild MDD).
Four recent meta-analyses (Kirsch et al. 2008, Fournier et al. 2010, Barbui et al. 2011, Gibbons et al. 2012) have been taken to have settled the question as to whether antidepressants are effective or not for milder presentations. Three of these reviews (Kirsch et al. 2008, Fournier et al. 2010, Barbui et al. 2011) concluded that antidepressants were no better than placebo in treating either mild MDD (Kirsch et al. 2008, Fournier et al. 2010) or sub-threshold depression (Fournier et al. 2010, Barbui et al. 2011) and one in a patient level, industry sponsored, data analysis (Gibbons et al. 2012) found antidepressants to be favourable over placebo (regardless of baseline severity) . However, each of these meta-analyses had specific limitations in
3
this regard. Kirsch et al’s review was selective (including only industry sponsored trials) with only one small trial involving participants with mild depression. In a patient level meta-analysis, Fournier et al., were only able to include a relatively small proportion of the relevant data (26% of identified studies) due to limited availability. Barbui et al., pooled clinically heterogeneous studies (e.g. adult and older residential care adult populations, treatment with different classes of antidepressants). The study by Gibbons et al was limited to two antidepressants (fluoxetine and venlafaxine) and, in their patient level data analysis of adult populations, included only studies sponsored by the related drug companies (Wyeth, and Eli Lilly and Co) despite the existence of other studies in this area (Dunlop et al. 1990, Judd et al. 1998, Fabre et al. 1987). Given that none of the reviews highlighted above have focussed on milder forms of depressive disorder, we conducted a review that specifically attempts to fill this gap.
Assessing the efficacy of antidepressants for mild or minor depressive symptoms has been complicated by issues of definition, (Pincus, Davis & McQueen 1999) compounded by discrepancies between the two major diagnostic classification systems (American Psychiatric Association 1994), (World Health Organisation ). For example ICD-10 requires two or three symptoms to be present to meet criteria for mild depression whereas DSM-IV requires five. As such ICD-10 mild depression appears more akin to DSM-IV sub-threshold depression. The emergence of DSM-V does not allay these concerns as the core criterion symptoms for a major depressive disorder remain unaltered (American Psychiatric Association 2013). The issue is further confused in that sub-threshold depression, mild MDD and moderate or severe depressive disorders are frequently considered to be on a continuum (Judd et al. 1998) (in clinical trials, the severity of depressive symptoms has tended to be measured with the Hamilton Rating Scale for Depression [HRSD] (Hamilton 1960) or
4
with the Montgomery Åsberg Depression Rating Scale [MÅDRS])(Montgomery, Asberg 1979). Confusion exists because there are both categorical and dimensional aspects to the classification of these disorders.
Aims of the study Our study aimed to examine the evidence for the effectiveness of antidepressants in milder forms of depressive disorder and addressed the following questions: (1) Are antidepressants more efficacious than placebo / treatment as usual (TAU) in the treatment of sub-threshold depression and; (2) Are antidepressants more efficacious than placebo / TAU in the treatment of mild MDD or depression (however specified) with symptoms which have been assessed to be mild? (3) Are antidepressants tolerable for people with sub-threshold or mild MDD as compared with those receiving placebo / TAU?
5
Material and methods
The methods of this systematic review and meta-analysis were pre-specified and are registered with Prospero in a written protocol: http://www.crd.york.ac.uk/PROSPERO/. (Reference number: CRD42013004505).
Criteria for considering studies for this review
Type of studies: Randomised controlled trials
Type of participants: People aged 16 years with a diagnosis of depressive disorder / episode according to standardised diagnostic criteria (e.g. ICD-10, DSM-IV or other system). In order to be included in the review, study participants had either to be diagnosed with mild MDD or sub-threshold depression. For studies with patients with a diagnosis of depression (not specified as mild or minor) it was required for participants to have initial baseline symptom severity of HRSD20 (or equivalent on other validated severity scale).
Studies which have used the HRSD have used different ranges of severity measure scores for entry criteria. Therefore, before conducting this review, it was necessary to define our range of interest according to how mild severity of symptoms has been defined. Mild depressive symptoms on HRSD-17 have been defined as scores of 8 to 13 (Rush, First & Blacker 2008). However, trials interested in treatment efficacy in milder depression have tended to use higher entry criteria whilst excluding the upper most scores on the measure and NICE define mild depressive symptoms on HRSD17 to include scores of (14 to 18) (NHS National Institute for Clinical Excellence 2009). In the absence of definitive criteria, in the current review, we define milder severity as being indicated by a HRSD-1720.
6
Studies were excluded if study participants: had a baseline HRSD-17 >20 or equivalent on another validated scale; had received more than one psychotropic medication concomitantly; or data were not provided for mild or sub-threshold groups.
Type of interventions: Eligible studies had to compare an antidepressant drug with either placebo or TAU however defined. The following commonly used antidepressant drugs were included: (1) tricyclic and related antidepressant drugs (TCAs); (2) selective serotonin re-uptake inhibitors (SSRIs) and related antidepressants; (3) and the following other drugs used in the treatment of depression: venlafaxine; tryptophan; mianserin and mirtazepine. Studies investigating Monamine Oxidase Inhibitors (MAOIs) were not included as they are potentially lethal dietary and drug interactions they are reserved for later use in treatment algorithms (Taylor, Paton & Kapur 2009). Studies were also excluded where: antidepressant treatment interventions were for indications other than depressive disorder; if more than one psychotropic drug was administered concomitantly; they only had active comparators to antidepressants drugs; they were of antidepressant efficacy in depression with physical co-morbidity; there was no clear description of baseline severity / diagnosis; there was no clear description of population; and they included participants with baseline HRSD-17>20 (or equivalent on other validated scales).
Types of outcome:
The a priori primary outcomes of interest were mood change at end of trial as measured by: significant reduction in symptom severity using a validated symptom severity measurement tool (e.g. HRSD, MADRS); a score of 'very much improved' or 'much improved' on Clinical Global Impression (CGI) Scale; or no longer fulfilling
7
DSM-III-R / DSM-IV criteria for mild MDD or sub-threshold depression or ICD-10 mild MDD or HRSD score of 7 or less.
A secondary outcome of interest was tolerability as indexed by number of people dropping out during the trial due to adverse events.
Search methods for identification of studies
Eligibility criteria were not restricted to studies of a particular follow up length, year of publication, language or publication status.
Electronic searches: Cochrane Central, Cumulative Index of Nursing and Allied Health Literature (CINAHL), Embase, Medline, PsycINFO and Science Citation Index (SCI) were searched from inception to Feb/Mar 2013. Search strategies specific to the databases listed above were developed and included the following architecture: (Depression) AND (Antidepressant) AND (RCTs). For example of search used with OVID Medline and Embase see Appendix 1.
Searching other resources: We searched systematic reviews of antidepressant efficacy, reference lists of identified papers and trial (clinicaltrials.gov) and review registers (Cochrane Register and the NIHR international prospective register of systematic reviews (Prospero)) for completed or ongoing trials.
Selection of studies
Titles and abstracts were initially screened and studies discarded where ineligibility was certain e.g. from explicitly stated study design, population, intervention etc. Each study was considered by two members of the research team working independently. Where there was disagreement regarding retention, papers were retained to the next stage. Full publications were sought for all papers retained and assessed for eligibility according to the inclusion and exclusion criteria outlined above. These
8
studies were again considered by two member of the research team, working independently, after which agreement was sought.
Data extraction and management
The following data were extracted from included studies: sample size at baseline and endpoint; target population; treatment setting; target antidepressant, dose and treatment duration; follow up duration; comparison group; outcome measures, scores and standard deviations at baseline and endpoint and / or remission rates. Missing data were sought by writing to authors. Where the HRSD was used and the version not specified, it was assumed that the original HRSD-17 was used.
For each treatment arm, we extracted the number of randomised participants and the number of analysed participants for each outcome. Dichotomous outcomes e.g. remission, were recorded as the number of participants experiencing the event in each group of the trial. For continuous data we extracted group means and standard deviations (s.d.) at baseline and follow up or group mean changes scores in the absence of follow up group means.
Assessment of risk of bias in included studies
The Cochrane Collaboration’s tool (Higgins et al. 2011) was applied for assessing risk of bias as high, low or unclear risk in relation to: randomisation method, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, completeness of outcome data and selective reporting.
Measures of treatment effect
Data were entered into Review Manager 5.2 (The Nordic Cochrane Centre, The Cochrane Collaboration 2012). We calculated risk ratio (RR) for dichotomous data
9
and mean difference (MD) for continuous data, with respect to 95% confidence intervals (CI).
Dealing with missing data
We performed a complete case analysis, such that, patients for whom an outcome was measured were included in the analysis.
Assessment of heterogeneity We assessed statistical heterogeneity in the meta-analysis using the T2, I2 and 2 statistics. Heterogeneity was considered substantial where T2>0 and either I2>50% or p value50%) heterogeneity was detected, possible causes would be explored in subgroup and sensitivity analysis.
Data synthesis
The fixed effect model for meta-analysis was used. If significant heterogeneity was present, a random effects model was applied to calculate the pooled effects size.
Sub-group analysis and investigation of heterogeneity
10
In the event of significant heterogeneity being detected, we aimed to conduct subgroup analyses of diagnosis / severity and class of antidepressant in order to explore possible causes.
11
Results
Study selection
Figure 1 outlines the process of screening and selecting studies for inclusion. Eight studies met criteria for inclusion. In five studies severity status was defined as subthreshold by diagnostic criteria (Williams et al. 2000, Barrett et al. 2001, Burrows et al. 2002, Judd et al. 2004, Rapaport et al. 2011) and in three studies severity status was defined in terms of symptom severity entry criteria (HRSD80 years) resident in care settings (Burrows et al. 2002) and a further one on adults>59 years (mean age 71 years) (Williams et al. 2000). Two of the studies had treatment arms of problem-solving therapy and included patients with dysthymia (Williams et al. 2000, Barrett et al. 2001). In the present review, we focus on the findings in relation to sub-threshold depression and for antidepressant versus placebo. Two studies had low risk of bias across all domains assessed (Judd et al. 2004, Rapaport et al. 2011), two had low risk of bias in all domains except attrition bias (Williams et al. 2000, Barrett et al. 2001) and one study showed low risk of bias in all domains with the exception of selection bias whereby the risk was unclear in terms of random sequence generation and allocation concealment.
12
Studies of mild MDD made up 502 participants. All were placebo controlled trials which took place in the United States and had duration / follow up of 6 weeks. Two studies investigated an SSRI (Fabre, Putman 1987, Dunlop et al. 1990) and one a Noradrenergic and specific serotonergic antidepressant (NaSSA) (McGrath et al. 1985). In all three studies, severity of depression was established with HRSD. It is of note that each study used slightly different thresholds for mild severity. Potential for high risk of bias was present in all three studies in terms of random sequence generation, allocation concealment and attrition bias. There was potential for high risk of bias in two studies (McGrath et al. 1985, Fabre, Putman 1987) in terms of detection bias and reporting bias.
Efficacy in sub-threshold depression
Five trials (all USA based) involving participants with a diagnosis of sub-threshold depression compared an SSRI with placebo. Two of these trials (Williams et al. 2000, Burrows et al. 2002) focused on older adults; the remaining 3 studies (Barrett et al. 2001, Judd et al. 2004, Rapaport et al. 2011) focused on adult participants.
Older adults
It was not possible to pool the data from the two studies of older adults since they were conducted in different treatment settings (Burrows et al., 2002 in a residential care setting and Williams et al., 2000 in a primary care setting). Burrows et al., 2002 reported that there was no difference in remission rates between older adults residential care receiving paroxetine or placebo (2=0.18, p=0.67). They did not report further outcome data e.g. means or remission rates within treatment groups.
13
Williams et al., 2000 also found no difference between paroxetine and placebo in older adults with sub-threshold depression attending four or more treatment sessions, however data were not presented in such a way as to allow reporting of remission rates for sub-threshold depression in this study.
Adults
Barrett et al., 2001, in a trial at high risk of attrition bias, found no difference between paroxetine and placebo in remission at eleven weeks for those participants attending four or more treatment sessions (proportion known to have remitted: paroxetine group 45% versus placebo group 54%) (Barrett et al. 2001). Rapaport et al., 2011 found no difference between fluoxetine and placebo in end point mean scores on the HRSD-17 at 12 weeks: (fluoxetine group mean difference -7.36 [sd 3.87] versus placebo group -7.09 [sd 3.79])(Rapaport et al. 2011). However Judd et al., 2004 in a similarly well designed trial, but with a larger sample did find a significant difference in favour of those treated with fluoxetine than those receiving placebo at 12 weeks: (fluoxetine group mean 7.1 [sd 5.0] versus placebo group 8.1 [sd 5.0]) (Judd et al. 2004).
It was possible to pool two of the three adult studies (Judd et al. 2004, Rapaport et al. 2011), the third study (Barrett et al. 2001) could not be included as relevant data were not available from the publication (where data were aggregated with data pertaining to patients with dysthymia) or after contact with the study authors (personal communication with JE Barrett). These are presented in Figure 2 and significantly favour antidepressant treatment over placebo (Mean Difference = -1.39 [-2.41, -0.36]; Z=2.66, P = 0.008).
14
Results of studies of efficacy in mild MDD
Three trials (all USA based) included participants with mild MDD. Two studies compared fluoxetine with placebo (Fabre, Putman 1987, Dunlop et al. 1990) and one Mianserin with placebo (McGrath et al. 1985). In the mild MDD studies, it was not possible to pool any of the three studies as only two of them investigated the same class of drug (Fabre, Putman 1987, Dunlop et al. 1990) and only one of those provided outcome data (Dunlop et al. 1990).
No difference in efficacy between groups were found in the studies of Dunlop et al., 1990 and Fabre et al., 1987 in patients with mild MDD receiving either fluoxetine or placebo at a range of doses (following up to six weeks of treatment) (Fabre, Putman 1987, Dunlop et al. 1990). Dunlop et al., reported an endpoint mean difference of: 6.2 (s.d. 7) (fluoxetine 20 mg); -6 (6.6) (fluoxetine 40 mg); -5.4 (5.6) (fluoxetine 60 mg) and -5.8 (5.9) placebo (see figure 3). Although Fabre et al., state there was no difference, no outcome data were reported to support this. McGrath et al., 1985 reported a significant difference in CGI remission in favour of mianserin at six weeks (57% versus 30%) (McGrath et al. 1985) (see figure 4).
Tolerability of antidepressants in sub-threshold depression and mild MDD
Data on dropouts due to adverse events were available for three studies of adults and one study of older adults with sub-threshold depression (Barrett et al. 2001), (Judd et al. 2004, Rapaport et al. 2011). Rates were only fully reported in one of the studies of mild MDD (Dunlop et al. 1990). We deemed it inappropriate to pool the five studies due to three issues: 1) four studies were of sub-threshold and one of mild
15
MDD; 2) of the former, three were of adults and one was of older adults; 3) furthermore studies varied with regard to the upper and lower limit of the dose prescribed (range 10-60 mg). We have reported the relative risk of each of the studies individually (figure 5). There was no difference in dropout due to adverse events between treatment and placebo in any of the studies. However, although the data are limited there is indication of a dose effect.
16
Discussion
Summary of evidence
To the best of our knowledge, our study is the first to systematically review the research literature on the efficacy and tolerability of antidepressants in the treatment of sub-threshold depression and mild MDD, without the limitations found in previous reviews. There is a scarcity of studies with low risk of bias which can address our research questions. None of the studies which assessed efficacy in mild MDD could be pooled due to clinical heterogeneity and insufficient reporting. These three studies of mild MDD reported conflicting results with one favouring the NaSSA mianserin over placebo (McGrath et al. 1985) and two finding no difference between an SSRI and placebo (Dunlop et al. 1990, Fieve et al. 1986). It should be noted that mirtazapine has pharmacological actions and side effects which are different to SSRIs and is now relatively little used, having been largely replaced by mirtazepine. All three studies of mild MDD exhibited either high risk of bias or were left with uncertainty. Only two studies which assessed the efficacy of antidepressants in subthreshold depression could be pooled for meta-analysis (Judd et al. 2004, Rapaport et al. 2011). This was due to clinical heterogeneity and lack of availability of compatible outcome data. The pooled mean difference between baseline and endpoint HRSD-17 showed antidepressant treatment to be more effective than placebo. Whilst this finding was statistically significant, the difference was small and unlikely to represent a meaningful clinical difference. Antidepressants were significantly less well tolerated than placebo. This was true generally but also specifically in relation to number dropping out due to adverse events. This held for studies of sub-threshold depression and mild MDD alike. However, these findings should also be considered in the context of considerable risk of bias. For example, in the study by Judd et al., 2004, individuals who had previously experienced an
17
adverse event with fluoxetine were excluded (Judd et al. 2004). As such, the findings may be a conservative estimate of what is a greater difference in tolerability.
Limitations
We did not search for data from clinical trials submitted for licensing to organisation such as the US Food and Drug Administration. However we did search for references cited in reviews which focused on such sources (Kirsch et al. 2008, Turner et al. 2008, Ani et al. 2008, Khan et al. 2007). Potential biases may arise in a review where authors deviate from the plans set out in the study protocol.
The HRSD is the common gold standard for assessing efficacy of antidepressants and was used in several of the studies we identified.
As the focus of our review was
on mild MDD and sub-threshold depression, a risk of bias may emerge through using this measure. Isacsson and Adler, 2012 demonstrated that the HRSD lacks precision and sensitivity to change when used to assess milder depression (Isacsson, Adler 2011). This must be taken into consideration when interpreting our findings.
Several of the studies in this review included a placebo washout period prior to randomisation (Burrows et al. 2002, Judd et al. 2004, McGrath et al. 1985, Fabre, Putman 1987, Dunlop et al. 1990, Stewart et al. 1983). Whilst previous reviewers have excluded trials with placebo washout (Fournier et al. 2010), we viewed these studies as relevant and have included them. This method is designed to remove participants with a propensity to respond to placebo and in so doing improve the capacity of trials to assess the pharmacological effects of a drug. This means that
18
such trials are likely to underestimate the true placebo effect. It may also be one aspect that explains the difference in findings between the studies of Judd et al. 2004 (who found a significant difference favouring antidepressants using a placebo washout period) and Rapaport et al., 2011 (who did not find a difference without placebo washout).
Our review focused on symptom severity. Additionally, other aspects such as symptom duration and functional ability should also be considered in relation to likely treatment response.
Context of existing literature
Our review yielded different studies to that of the review of antidepressant efficacy in sub-threshold depression of Barbui et al., 2011. Whilst four studies were common to both reviews (Williams et al. 2000, Barrett et al. 2001, Burrows et al. 2002, Judd et al. 2004) we included a further study published in 2011 (Rapaport et al. 2011)and excluded two studies. One study was excluded as it recruited participants with anxious depression being treated in an in-patient setting (Davidson et al. 1988) and a further study excluded because a post hoc analysis was conducted (Paykel et al. 1988) of a primary care trial (Hollyman et al. 1988) which was not powered to conduct sub-group analyses. The original study by Hollyman et al., included individuals with HRSD-1720 at baseline. In comparing our review with that of Kirsch et al., 2008 (Kirsch et al. 2008) one study was common to both (Dunlop et al. 1990). This was the only study in the review of Kirsch et al., that focused on individuals meeting mild MDD criteria. Indeed it was the only study in the 35 included where the baseline HRSD was mild) with undifferentiat ed analysis
Study participants received >1 psychotropic medication or more than one drug was being tested without randomisation X
Duplicate or secondary analysis
Inability to determine mild or minor sub-group
X X X X X X X X X X X X X X X X X
X
X X X X X X X X X X X X X X X X
22
References Ackermann, R.T. & Williams, J.W.,Jr 2002, "Rational treatment choices for non-major depressions in primary care: an evidence-based review", Journal of General Internal Medicine, vol. 17, no. 4, pp. 293-301. American Psychiatric Association 2013, Diagnostic and statistical manual of mental disorders: DSM5, 5th edn, American Psychiatric Association. American Psychiatric Association 1994, Diagnostic and Statistical Manual of Mental Disorders, fourth edn, American Psychiatric Association, Washington DC. Ani, C., Bazargan, M., Hindman, D., Bell, D., Farooq, M.A., Akhanjee, L., Yemofio, F., Baker, R. & Rodriguez, M. 2008, "Depression symptomatology and diagnosis: Discordance between patients and physicians in primary care settings.", BMC Family Practice, vol. 9. Arroll, B., Macgillivray, S., Ogston, S., Reid, I., Sullivan, F., Williams, B. & Crombie, I. 2005, "Efficacy and tolerability of tricyclic antidepressants and SSRIs compared with placebo for treatment of depression in primary care: a meta-analysis", Annals of Family Medicine, vol. 3, pp. 449-456. Barbui, C., Cipriani, A., Patel, V., Ayuso-Mateos, J.L. & van Ommeren, M. 2011, "Efficacy of antidepressants and benzodiazepines in minor depression: systematic review and meta-analysis", British Journal of Psychiatry, vol. 198, no. 1, pp. 11-16. Barrett, J.E., Williams, J.W.,Jr, 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, vol. 50, no. 5, pp. 405-412. Bauer, M., Bschor, T., Pfennig, A., Whybrow, P.C., Angst, J., Versiani, M., Moller, H.J. & WFSBP Task Force on Unipolar Depressive,Disorders 2007, "World Federation of Societies of Biological Psychiatry (WFSBP) Guidelines for Biological Treatment of Unipolar Depressive Disorders in Primary Care", World Journal of Biological Psychiatry, vol. 8, no. 2, pp. 67-104. Baumeister, H. 2012, "Inappropriate prescriptions of antidepressant drugs in patients with subthreshold to mild depression: time for the evidence to become practice", Journal of affective disorders, vol. 139, no. 3, pp. 240-243. Broadhead, W.E., Blazer, D.G., George, L.K. & Tse, C.K. 1990, "Depression, disability days, and days lost from work in a prospective epidemiologic survey.", JAMA, vol. 264, no. 19, pp. 2524-2528. Burrows, A.B., Salzman, C., Satlin, A., Noble, K., Pollock, B.G. & Gersh, T. 2002, "A randomized, placebo-controlled trial of paroxetine in nursing home residents with non-major depression.", Depression & Anxiety, vol. 15, no. 3, pp. 102-110. Davidson, J.R., Giller, E.L., Zisook, S. & Overall, J.E. 1988, "An efficacy study of isocarboxazid and placebo in depression, and its relationship to depressive nosology.", Archives of General Psychiatry, vol. 45, no. 2, pp. 120-127.
23
Dimidjian, S., Hollon, S.D., Dobson, K.S., Schmaling, K.B., Kohlenberg, R.J., Addis, M.E., Gallop, R., McGlinchey, J.B., Markley, D.K., Gollan, J.K., Atkins, D.C., Dunner, D.L. & Jacobson, N.S. 2006, "Randomized trial of behavioral activation, cognitive therapy, and antidepressant medication in the acute treatment of adults with major depression.", Journal of consulting and clinical psychology, vol. 74, no. 4, pp. 658-670. Dunlop, S.R., Dornseif, B.E., Wernicke, J.F. & Potvin, J.H. 1990, "Pattern analysis shows beneficial effect of fluoxetine treatment in mild depression.", Psychopharmacology bulletin, vol. 26, no. 2, pp. 173-180. Fabre, L.F. & Putman, H.P.,3rd 1987, "A fixed-dose clinical trial of fluoxetine in outpatients with major depression.", Journal of Clinical Psychiatry, vol. 48, no. 10, pp. 406-408. Fieve, R.R., Goodnick, P.J., Peselow, E.D., Barouche, F. & Schlegel, A. 1986, "Pattern analysis of antidepressant response to fluoxetine.", Journal of Clinical Psychiatry, vol. 47, no. 11, pp. 560-562. Fournier, J.C., DeRubeis, R.J., Hollon, S.D., Dimidjian, S., Amsterdam, J.D., Shelton, R.C. & Fawcett, J. 2010, "Antidepressant drug effects and depression severity: a patient-level meta-analysis", JAMA, vol. 303, no. 1, pp. 47-53. Gibbons, R.D., Hur, K., Brown, C.H., Davis, J.M. & Mann, J.J. 2012, "Benefits from antidepressants: synthesis of 6-week patient-level outcomes from double-blind placebo-controlled randomized trials of fluoxetine and venlafaxine.", Archives of General Psychiatry, vol. 69, no. 6, pp. 572-579. Godlee, F. 2013, "Don't keep taking the tablets", BMJ, vol. 347, pp. f7438. Hamilton, M. 1960, "A rating scale for depression", Journal of Neurology, Neurosurgery and Psychiatry, vol. 23, pp. 56-62. Higgins, J.P., Altman, D.G., Gotzsche, P.C., Juni, P., Moher, D., Oxman, A.D., Savovic, J., Schulz, K.F., Weeks, L., Sterne, J.A., Cochrane Bias Methods, G. & Cochrane Statistical Methods, G. 2011, "The Cochrane Collaboration's tool for assessing risk of bias in randomised trials.", BMJ, vol. 343, pp. 5928. Hollyman, J.A., Freeling, P., Paykel, E.S., Bhat, A. & Sedgwick, P. 1988, "Doubleblind placebo-controlled trial of amitriptyline among depressed patients in general practice.", Journal of the Royal College of General Practitioners, vol. 38, no. 314, pp. 393-397. Isacsson, G. & Adler, M. 2011, "Randomized clinical trials underestimate the efficacy of antidepressants in less severe depression", Acta Psychiatrica Scandinavica, , pp. no-no. Judd, L.L., Akiskal, H.S.:.M., J.D., Zeller, P.J., Endicott, J., Coryell, W., Paulus, M.P., Kunovac, J.L., Leon, A.C., Mueller, T.I., Rice, J.A. & Keller, M.B. 1998, "A prospective 12-year study of subsyndromal and syndromal depressive symptoms in unipolar major depressive disorders", Archives of General Psychiatry, vol. 55, pp. 694-700.
24
Judd, L.L., Rapaport, M.H., Yonkers, K.A., Rush, A.J., Frank, E., Thase, M.E., Kupfer, D.J., Plewes, J.M., Schettler, P.J. & Tollefson, G. 2004, "Randomized, placebo-controlled trial of fluoxetine for acute treatment of minor depressive disorder", American Journal of Psychiatry, vol. 161, no. 10, pp. 1864-1871. Kendrick, T., Chatwin, J., Dowrick, C., Tylee, A., Morriss, R., Peveler, R., Leese, M., McCrone, P., Harris, T., Moore, M., Byng, R., Brown, G., Barthel, S., Mander, H., Ring, A., Kelly, V., Wallace, V., Gabbay, M., Craig, T. & Mann, A. 2009, Randomised controlled trial to determine the cost-effectiveness of fluoxetine for mild to moderate depression with somatic symptoms in primary care THREshold for AntiDepressant treatment (THREAD) Study, Health Technology Assessment. Kessler, R.C., Zhao, S., Blazer, D.G. & Swartz, M. 1997, "Prevalence, correlates, and course of minor depression and major depression in the national comorbidity survey", Journal of affective disorders, vol. 45, no. 1-2, pp. 19-30. Khan, A., Schwartz, K., Kolts, R.L., Ridgway, D. & Lineberry, C. 2007, "Relationship between depression severity entry criteria and antidepressant clinical trial outcomes.", Biological psychiatry, vol. 62, no. 1, pp. 65-71. Kirsch, I., Deacon, B.J., Huedo-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 / Public Library of Science, vol. 5, no. 2, pp. e45. MacGillivray, S., Arroll, B., Hatcher, S., Ogston, S., Reid, I., Sullivan, F., Williams, B. & Crombie, I. 2003, "Efficacy and tolerability of selective serotonin reuptake inhibitors compared with tricyclic antidepressants in depression treated in primary care: systematic review and meta-analysis", BMJ, vol. 326, no. 7397, pp. 1014. McGrath, P.J., Rabkin, J.G., Stewart, J.W., Harrison, W., Quitkin, F.M. & Markowitz, J. 1985, "Placebo-controlled study of mianserin in depressed outpatients.", Neuropsychobiology, vol. 14, no. 3, pp. 128-132. Montgomery, S.A. & Asberg, M. 1979, "A new depression scale designed to be sensitive to change", British Journal of Psychiatry, vol. 134, pp. 382-389. NHS National Institute for Clinical Excellence 2009, Depression: the treatment and management in adults (update). Nutt, D.J. 2011, "Highlights of the international consensus statement on major depressive disorder.", Journal of Clinical Psychiatry, vol. 72, no. 6, pp. e21. Oxman, T.E. & Sengupta, A. 2002, "Treatment of minor depression", American Journal of Geriatric Psychiatry, vol. 10, no. 3, pp. 256-264. Paykel, E.S., Hollyman, J.A., Freeling, P. & Sedgwick, P. 1988, "Predictors of therapeutic benefit from amitriptyline in mild depression: a general practice placebo-controlled trial", Journal of Affective Disorders, vol. 14, no. 1, pp. 83-95.
25
Pincus, H.A., Davis, W.W. & McQueen, L.E. 1999, "'Subthreshold' mental disorders. A review and synthesis of studies on minor depression and other 'brand names'", British Journal of Psychiatry, vol. 174, pp. 288-296. Rapaport, M.H., Judd, L.L., Schettler, P.J., Yonkers, K.A., Thase, M.E., Kupfer, D.J., Frank, E., Plewes, J.M., Tollefson, G.D. & Rush, A.J. 2002, "A descriptive analysis of minor depression.", American Journal of Psychiatry, vol. 159, no. 4, pp. 637-643. Rapaport, M.H., Nierenberg, A.A., Howland, R., Dording, C., Schettler, P.J. & Mischoulon, D. 2011, "The treatment of minor depression with St. John's Wort or citalopram: failure to show benefit over placebo.", Journal of psychiatric research, vol. 45, no. 7, pp. 931-941. Rush, A.J., First, M.B. & Blacker, D. 2008, Handbook of Psychiatric Measures, Second edn, American Psychiatric Association, USA. Song, F., Freemantle, N., Sheldon, T.A., House, A., Watson, P., Long, A. & Mason, J. 1993, "Selective serotonin reuptake inhibitors: meta-analysis of efficacy and acceptability", BMJ, vol. 306, pp. 683-687. Stewart, J.W., Quitkin, F.M., Liebowitz, M.R., McGrath, P.J., Harrison, W.M. & Klein, D.F. 1983, "Efficacy of desipramine in depressed outpatients. Response according to research diagnosis criteria diagnoses and severity of illness.", Archives of General Psychiatry, vol. 40, no. 2, pp. 202-207. Taylor, D., Paton, C. & Kapur, S. 2009, Maudsley Prescribing Guidelines, 10th edition, 10th edn, Informa Healthcare, London. The Nordic Cochrane Centre, The Cochrane Collaboration 2012, . 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, vol. 358, no. 3, pp. 252-260. Williams, J.W.,Jr, 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, vol. 284, no. 12, pp. 1519-1526.
Screening
Identification
World Health Organisation , International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Version for 2007. Available: http://www.who.int/classifications/apps/icd/icd10online/ [2008, 05/08].
Recordsidentifiedthrough databasesearching (n=5807)
Additionalrecordsidentified throughothersources (n=3)
Recordsscreened (n=5810)
Recordsexcluded Duplicates,notRCT, 26 sampleswithcomorbidity, secondaryanalysisof includedstudy
Figure 1 Process of identifying and including studies
27
102
78 -2.4 3.7 24 -7.09 3.79 78.2% -1.70 [-2.86, -0.54] 21.8% -0.27 [-2.46, 1.92]
102 100.0% -1.39 [-2.41, -0.36]
79 23
-10
-5 0 5 10 Favours SSRI Favours Placebo
Mean Difference IV, Fixed, 95% CI
Figure 4: Risk ratio of response (much or very much improved on CGI) for one study of mianserin for Mild MDD
Figure 3: Mean difference in efficacy (change score of HDRS from baseline) for 3 dosing schedules of flusoxetine in one study of Mild MDD
Figure 2 Mean difference in efficacy (change score of HDRS from baseline) for two pooled Sub-threshold depression studies
Heterogeneity: Chi² = 1.28, df = 1 (P = 0.26); I² = 22% Test for overall effect: Z = 2.66 (P = 0.008)
Total (95% CI)
-4.1 3.7 -7.36 3.87
Judd 2004 Fluoxetine Rapaport 2011 Citalopram
SSRI Placebo Mean Difference SD Total Mean SD Total Weight IV, Fixed, 95% CI
Mean
Study or Subgroup
28
Figure 5: Risk ratio of study drop-out due to adverse events for 4 studies of sub-threshold depression and 1 at 3 dosage levels of Mild MDD
29
McGrath 1985
Fabre 1987
Dunlop 1990
Williams 2000
Rapaport 2011
Judd 2004
Burrows 2002
Barrett 2001
Paroxetine 20(10-40mg) Placebo Paroxetine 10 - 30 mg Placebo Fluoxetine 20(10–20mg) Placebo Citalopram 20 mg Placebo Paroxetine 20(10-40mg) Placebo Fluoxetine 20 mg Fluoxetine 40 mg Fluoxetine 60 mg Placebo Fluoxetine 20 mg Fluoxetine 40 mg Fluoxetine 60 mg Placebo Mianserin 30 – 150 mg Placebo 50
6 42
11
11
56 10
104
105
70 107
25 69
80 27
12 81
39 12
38
92
38
372
138
52
162
24
77
Table 1 Characteristics of included studies Study Intervention N Total N
RDC DD HRSD-21 (
Efficacy and tolerability of antidepressants for sub-threshold depression and for mild major depressive disorder Isobel M Cameron, Ian C Reid, Steve A MacGillivray
www.elsevier.com/locate/jad
PII: DOI: Reference:
S0165-0327(14)00272-9 http://dx.doi.org/10.1016/j.jad.2014.04.078 JAD6745
To appear in:
Journal of Affective Disorders
Received date: 11 November 2013 Revised date: 28 April 2014 Accepted date: 29 April 2014 Cite this article as: Isobel M Cameron, Ian C Reid, Steve A MacGillivray, Efficacy and tolerability of antidepressants for sub-threshold depression and for mild major depressive disorder, Journal of Affective Disorders, http://dx.doi.org/10.1016/j. jad.2014.04.078 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Title Efficacy and tolerability of antidepressants for sub-threshold depression and for mild major depressive disorder Authors Isobel M Cameron, Applied Medical Sciences (Psychiatry), University of Aberdeen, Royal Cornhill Hospital, Aberdeen, AB25 2ZH, Scotland.
Ian C Reid, Applied Medical Sciences (Psychiatry), University of Aberdeen, Royal Cornhill Hospital, Aberdeen, AB25 2ZH, Scotland. Steve A MacGillivray*, Social Dimensions of Health Institute, University of Dundee, DD1 4HJ, Scotland.
*Correspondence to: Steve MacGillivray, Social Dimensions of Health Institute, University of Dundee, DD1 4HJ. Tel: +44 (0)1382 381075; Fax: +44 (0)1382 388533; email: [email protected]
1
Abstract Background: It is commonly assumed that robust evidence exists for the lack of efficacy of antidepressants at the milder end of the depression severity spectrum. In light of specific limitations of existing reviews, we assess if antidepressants are efficacious and tolerable for sub-threshold or mild Major Depressive Disorder (MDD).
Method: Systematic review and meta-analysis of Randomised Controlled Trials of adults with sub-threshold depression or mild MDD (initial baseline symptom severity of HRSD20) comparing an antidepressant with placebo or treatment as usual (TAU) however defined. A pre-specified protocol was published (Prospero reference: CRD42013004505). Results: 8 trials were included: 5 trials (453 participants) of sub-threshold depression and 3 trials (502 participants) of mild MDD. Trials of sub-threshold depression exhibited low risk of bias whereas those of mild MDD exhibited high risk. Two trials of sub-threshold depression were pooled (n=102) to assess efficacy and favoured antidepressants over placebo statistically but the difference was small and unlikely to be clinically meaningful: mean difference -1.39 (-2.41, -0.36). Due to heterogeneity, no trials of mild MDD could be pooled for efficacy. There was no difference between antidepressant treatment and placebo for drop out due to adverse events. The maximum proportion in those receiving antidepressants dropping out due to adverse events was 17%, with indication of a dose effect.
Limitations: Not all data from identified trials could be included in the meta-analyses due to a lack of availability of relevant data.
Conclusion: There is insufficient evidence to support or contest the efficacy of antidepressant medication for sub-threshold depression or mild MDD. More trials, with adequate follow up, are required to address this question. Keywords: systematic review, meta-analysis, depressive disorder, antidepressants
2
Introduction
Establishing the efficacy of antidepressants for milder forms of depression is important because diagnoses such as minor, sub-threshold or mild depression have been found to be as prevalent as more severe forms (Ackermann, Williams 2002, Oxman, Sengupta 2002). Such diagnoses are associated with considerable functional impairment (Rapaport et al. 2002, Broadhead et al. 1990) and are linked with increased risk of developing a major depressive disorder (Kessler et al. 1997).
Robust evidence supports treatment of moderate to severe depression with antidepressant drugs (Bauer et al. 2007, NHS National Institute for Clinical Excellence 2009, Nutt 2011). It is commonly assumed that there is equally robust evidence for their lack of efficacy at the milder end of the severity spectrum (Nutt 2011, Baumeister 2012, Godlee 2013). Guidelines (Bauer et al. 2007, NHS National Institute for Clinical Excellence 2009, Nutt 2011) do not currently recommend the routine use of antidepressants for those with sub-threshold depression (also referred to as Minor Depression - MinD) or Mild Major Depressive Disorder (Mild MDD).
Four recent meta-analyses (Kirsch et al. 2008, Fournier et al. 2010, Barbui et al. 2011, Gibbons et al. 2012) have been taken to have settled the question as to whether antidepressants are effective or not for milder presentations. Three of these reviews (Kirsch et al. 2008, Fournier et al. 2010, Barbui et al. 2011) concluded that antidepressants were no better than placebo in treating either mild MDD (Kirsch et al. 2008, Fournier et al. 2010) or sub-threshold depression (Fournier et al. 2010, Barbui et al. 2011) and one in a patient level, industry sponsored, data analysis (Gibbons et al. 2012) found antidepressants to be favourable over placebo (regardless of baseline severity) . However, each of these meta-analyses had specific limitations in
3
this regard. Kirsch et al’s review was selective (including only industry sponsored trials) with only one small trial involving participants with mild depression. In a patient level meta-analysis, Fournier et al., were only able to include a relatively small proportion of the relevant data (26% of identified studies) due to limited availability. Barbui et al., pooled clinically heterogeneous studies (e.g. adult and older residential care adult populations, treatment with different classes of antidepressants). The study by Gibbons et al was limited to two antidepressants (fluoxetine and venlafaxine) and, in their patient level data analysis of adult populations, included only studies sponsored by the related drug companies (Wyeth, and Eli Lilly and Co) despite the existence of other studies in this area (Dunlop et al. 1990, Judd et al. 1998, Fabre et al. 1987). Given that none of the reviews highlighted above have focussed on milder forms of depressive disorder, we conducted a review that specifically attempts to fill this gap.
Assessing the efficacy of antidepressants for mild or minor depressive symptoms has been complicated by issues of definition, (Pincus, Davis & McQueen 1999) compounded by discrepancies between the two major diagnostic classification systems (American Psychiatric Association 1994), (World Health Organisation ). For example ICD-10 requires two or three symptoms to be present to meet criteria for mild depression whereas DSM-IV requires five. As such ICD-10 mild depression appears more akin to DSM-IV sub-threshold depression. The emergence of DSM-V does not allay these concerns as the core criterion symptoms for a major depressive disorder remain unaltered (American Psychiatric Association 2013). The issue is further confused in that sub-threshold depression, mild MDD and moderate or severe depressive disorders are frequently considered to be on a continuum (Judd et al. 1998) (in clinical trials, the severity of depressive symptoms has tended to be measured with the Hamilton Rating Scale for Depression [HRSD] (Hamilton 1960) or
4
with the Montgomery Åsberg Depression Rating Scale [MÅDRS])(Montgomery, Asberg 1979). Confusion exists because there are both categorical and dimensional aspects to the classification of these disorders.
Aims of the study Our study aimed to examine the evidence for the effectiveness of antidepressants in milder forms of depressive disorder and addressed the following questions: (1) Are antidepressants more efficacious than placebo / treatment as usual (TAU) in the treatment of sub-threshold depression and; (2) Are antidepressants more efficacious than placebo / TAU in the treatment of mild MDD or depression (however specified) with symptoms which have been assessed to be mild? (3) Are antidepressants tolerable for people with sub-threshold or mild MDD as compared with those receiving placebo / TAU?
5
Material and methods
The methods of this systematic review and meta-analysis were pre-specified and are registered with Prospero in a written protocol: http://www.crd.york.ac.uk/PROSPERO/. (Reference number: CRD42013004505).
Criteria for considering studies for this review
Type of studies: Randomised controlled trials
Type of participants: People aged 16 years with a diagnosis of depressive disorder / episode according to standardised diagnostic criteria (e.g. ICD-10, DSM-IV or other system). In order to be included in the review, study participants had either to be diagnosed with mild MDD or sub-threshold depression. For studies with patients with a diagnosis of depression (not specified as mild or minor) it was required for participants to have initial baseline symptom severity of HRSD20 (or equivalent on other validated severity scale).
Studies which have used the HRSD have used different ranges of severity measure scores for entry criteria. Therefore, before conducting this review, it was necessary to define our range of interest according to how mild severity of symptoms has been defined. Mild depressive symptoms on HRSD-17 have been defined as scores of 8 to 13 (Rush, First & Blacker 2008). However, trials interested in treatment efficacy in milder depression have tended to use higher entry criteria whilst excluding the upper most scores on the measure and NICE define mild depressive symptoms on HRSD17 to include scores of (14 to 18) (NHS National Institute for Clinical Excellence 2009). In the absence of definitive criteria, in the current review, we define milder severity as being indicated by a HRSD-1720.
6
Studies were excluded if study participants: had a baseline HRSD-17 >20 or equivalent on another validated scale; had received more than one psychotropic medication concomitantly; or data were not provided for mild or sub-threshold groups.
Type of interventions: Eligible studies had to compare an antidepressant drug with either placebo or TAU however defined. The following commonly used antidepressant drugs were included: (1) tricyclic and related antidepressant drugs (TCAs); (2) selective serotonin re-uptake inhibitors (SSRIs) and related antidepressants; (3) and the following other drugs used in the treatment of depression: venlafaxine; tryptophan; mianserin and mirtazepine. Studies investigating Monamine Oxidase Inhibitors (MAOIs) were not included as they are potentially lethal dietary and drug interactions they are reserved for later use in treatment algorithms (Taylor, Paton & Kapur 2009). Studies were also excluded where: antidepressant treatment interventions were for indications other than depressive disorder; if more than one psychotropic drug was administered concomitantly; they only had active comparators to antidepressants drugs; they were of antidepressant efficacy in depression with physical co-morbidity; there was no clear description of baseline severity / diagnosis; there was no clear description of population; and they included participants with baseline HRSD-17>20 (or equivalent on other validated scales).
Types of outcome:
The a priori primary outcomes of interest were mood change at end of trial as measured by: significant reduction in symptom severity using a validated symptom severity measurement tool (e.g. HRSD, MADRS); a score of 'very much improved' or 'much improved' on Clinical Global Impression (CGI) Scale; or no longer fulfilling
7
DSM-III-R / DSM-IV criteria for mild MDD or sub-threshold depression or ICD-10 mild MDD or HRSD score of 7 or less.
A secondary outcome of interest was tolerability as indexed by number of people dropping out during the trial due to adverse events.
Search methods for identification of studies
Eligibility criteria were not restricted to studies of a particular follow up length, year of publication, language or publication status.
Electronic searches: Cochrane Central, Cumulative Index of Nursing and Allied Health Literature (CINAHL), Embase, Medline, PsycINFO and Science Citation Index (SCI) were searched from inception to Feb/Mar 2013. Search strategies specific to the databases listed above were developed and included the following architecture: (Depression) AND (Antidepressant) AND (RCTs). For example of search used with OVID Medline and Embase see Appendix 1.
Searching other resources: We searched systematic reviews of antidepressant efficacy, reference lists of identified papers and trial (clinicaltrials.gov) and review registers (Cochrane Register and the NIHR international prospective register of systematic reviews (Prospero)) for completed or ongoing trials.
Selection of studies
Titles and abstracts were initially screened and studies discarded where ineligibility was certain e.g. from explicitly stated study design, population, intervention etc. Each study was considered by two members of the research team working independently. Where there was disagreement regarding retention, papers were retained to the next stage. Full publications were sought for all papers retained and assessed for eligibility according to the inclusion and exclusion criteria outlined above. These
8
studies were again considered by two member of the research team, working independently, after which agreement was sought.
Data extraction and management
The following data were extracted from included studies: sample size at baseline and endpoint; target population; treatment setting; target antidepressant, dose and treatment duration; follow up duration; comparison group; outcome measures, scores and standard deviations at baseline and endpoint and / or remission rates. Missing data were sought by writing to authors. Where the HRSD was used and the version not specified, it was assumed that the original HRSD-17 was used.
For each treatment arm, we extracted the number of randomised participants and the number of analysed participants for each outcome. Dichotomous outcomes e.g. remission, were recorded as the number of participants experiencing the event in each group of the trial. For continuous data we extracted group means and standard deviations (s.d.) at baseline and follow up or group mean changes scores in the absence of follow up group means.
Assessment of risk of bias in included studies
The Cochrane Collaboration’s tool (Higgins et al. 2011) was applied for assessing risk of bias as high, low or unclear risk in relation to: randomisation method, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, completeness of outcome data and selective reporting.
Measures of treatment effect
Data were entered into Review Manager 5.2 (The Nordic Cochrane Centre, The Cochrane Collaboration 2012). We calculated risk ratio (RR) for dichotomous data
9
and mean difference (MD) for continuous data, with respect to 95% confidence intervals (CI).
Dealing with missing data
We performed a complete case analysis, such that, patients for whom an outcome was measured were included in the analysis.
Assessment of heterogeneity We assessed statistical heterogeneity in the meta-analysis using the T2, I2 and 2 statistics. Heterogeneity was considered substantial where T2>0 and either I2>50% or p value50%) heterogeneity was detected, possible causes would be explored in subgroup and sensitivity analysis.
Data synthesis
The fixed effect model for meta-analysis was used. If significant heterogeneity was present, a random effects model was applied to calculate the pooled effects size.
Sub-group analysis and investigation of heterogeneity
10
In the event of significant heterogeneity being detected, we aimed to conduct subgroup analyses of diagnosis / severity and class of antidepressant in order to explore possible causes.
11
Results
Study selection
Figure 1 outlines the process of screening and selecting studies for inclusion. Eight studies met criteria for inclusion. In five studies severity status was defined as subthreshold by diagnostic criteria (Williams et al. 2000, Barrett et al. 2001, Burrows et al. 2002, Judd et al. 2004, Rapaport et al. 2011) and in three studies severity status was defined in terms of symptom severity entry criteria (HRSD80 years) resident in care settings (Burrows et al. 2002) and a further one on adults>59 years (mean age 71 years) (Williams et al. 2000). Two of the studies had treatment arms of problem-solving therapy and included patients with dysthymia (Williams et al. 2000, Barrett et al. 2001). In the present review, we focus on the findings in relation to sub-threshold depression and for antidepressant versus placebo. Two studies had low risk of bias across all domains assessed (Judd et al. 2004, Rapaport et al. 2011), two had low risk of bias in all domains except attrition bias (Williams et al. 2000, Barrett et al. 2001) and one study showed low risk of bias in all domains with the exception of selection bias whereby the risk was unclear in terms of random sequence generation and allocation concealment.
12
Studies of mild MDD made up 502 participants. All were placebo controlled trials which took place in the United States and had duration / follow up of 6 weeks. Two studies investigated an SSRI (Fabre, Putman 1987, Dunlop et al. 1990) and one a Noradrenergic and specific serotonergic antidepressant (NaSSA) (McGrath et al. 1985). In all three studies, severity of depression was established with HRSD. It is of note that each study used slightly different thresholds for mild severity. Potential for high risk of bias was present in all three studies in terms of random sequence generation, allocation concealment and attrition bias. There was potential for high risk of bias in two studies (McGrath et al. 1985, Fabre, Putman 1987) in terms of detection bias and reporting bias.
Efficacy in sub-threshold depression
Five trials (all USA based) involving participants with a diagnosis of sub-threshold depression compared an SSRI with placebo. Two of these trials (Williams et al. 2000, Burrows et al. 2002) focused on older adults; the remaining 3 studies (Barrett et al. 2001, Judd et al. 2004, Rapaport et al. 2011) focused on adult participants.
Older adults
It was not possible to pool the data from the two studies of older adults since they were conducted in different treatment settings (Burrows et al., 2002 in a residential care setting and Williams et al., 2000 in a primary care setting). Burrows et al., 2002 reported that there was no difference in remission rates between older adults residential care receiving paroxetine or placebo (2=0.18, p=0.67). They did not report further outcome data e.g. means or remission rates within treatment groups.
13
Williams et al., 2000 also found no difference between paroxetine and placebo in older adults with sub-threshold depression attending four or more treatment sessions, however data were not presented in such a way as to allow reporting of remission rates for sub-threshold depression in this study.
Adults
Barrett et al., 2001, in a trial at high risk of attrition bias, found no difference between paroxetine and placebo in remission at eleven weeks for those participants attending four or more treatment sessions (proportion known to have remitted: paroxetine group 45% versus placebo group 54%) (Barrett et al. 2001). Rapaport et al., 2011 found no difference between fluoxetine and placebo in end point mean scores on the HRSD-17 at 12 weeks: (fluoxetine group mean difference -7.36 [sd 3.87] versus placebo group -7.09 [sd 3.79])(Rapaport et al. 2011). However Judd et al., 2004 in a similarly well designed trial, but with a larger sample did find a significant difference in favour of those treated with fluoxetine than those receiving placebo at 12 weeks: (fluoxetine group mean 7.1 [sd 5.0] versus placebo group 8.1 [sd 5.0]) (Judd et al. 2004).
It was possible to pool two of the three adult studies (Judd et al. 2004, Rapaport et al. 2011), the third study (Barrett et al. 2001) could not be included as relevant data were not available from the publication (where data were aggregated with data pertaining to patients with dysthymia) or after contact with the study authors (personal communication with JE Barrett). These are presented in Figure 2 and significantly favour antidepressant treatment over placebo (Mean Difference = -1.39 [-2.41, -0.36]; Z=2.66, P = 0.008).
14
Results of studies of efficacy in mild MDD
Three trials (all USA based) included participants with mild MDD. Two studies compared fluoxetine with placebo (Fabre, Putman 1987, Dunlop et al. 1990) and one Mianserin with placebo (McGrath et al. 1985). In the mild MDD studies, it was not possible to pool any of the three studies as only two of them investigated the same class of drug (Fabre, Putman 1987, Dunlop et al. 1990) and only one of those provided outcome data (Dunlop et al. 1990).
No difference in efficacy between groups were found in the studies of Dunlop et al., 1990 and Fabre et al., 1987 in patients with mild MDD receiving either fluoxetine or placebo at a range of doses (following up to six weeks of treatment) (Fabre, Putman 1987, Dunlop et al. 1990). Dunlop et al., reported an endpoint mean difference of: 6.2 (s.d. 7) (fluoxetine 20 mg); -6 (6.6) (fluoxetine 40 mg); -5.4 (5.6) (fluoxetine 60 mg) and -5.8 (5.9) placebo (see figure 3). Although Fabre et al., state there was no difference, no outcome data were reported to support this. McGrath et al., 1985 reported a significant difference in CGI remission in favour of mianserin at six weeks (57% versus 30%) (McGrath et al. 1985) (see figure 4).
Tolerability of antidepressants in sub-threshold depression and mild MDD
Data on dropouts due to adverse events were available for three studies of adults and one study of older adults with sub-threshold depression (Barrett et al. 2001), (Judd et al. 2004, Rapaport et al. 2011). Rates were only fully reported in one of the studies of mild MDD (Dunlop et al. 1990). We deemed it inappropriate to pool the five studies due to three issues: 1) four studies were of sub-threshold and one of mild
15
MDD; 2) of the former, three were of adults and one was of older adults; 3) furthermore studies varied with regard to the upper and lower limit of the dose prescribed (range 10-60 mg). We have reported the relative risk of each of the studies individually (figure 5). There was no difference in dropout due to adverse events between treatment and placebo in any of the studies. However, although the data are limited there is indication of a dose effect.
16
Discussion
Summary of evidence
To the best of our knowledge, our study is the first to systematically review the research literature on the efficacy and tolerability of antidepressants in the treatment of sub-threshold depression and mild MDD, without the limitations found in previous reviews. There is a scarcity of studies with low risk of bias which can address our research questions. None of the studies which assessed efficacy in mild MDD could be pooled due to clinical heterogeneity and insufficient reporting. These three studies of mild MDD reported conflicting results with one favouring the NaSSA mianserin over placebo (McGrath et al. 1985) and two finding no difference between an SSRI and placebo (Dunlop et al. 1990, Fieve et al. 1986). It should be noted that mirtazapine has pharmacological actions and side effects which are different to SSRIs and is now relatively little used, having been largely replaced by mirtazepine. All three studies of mild MDD exhibited either high risk of bias or were left with uncertainty. Only two studies which assessed the efficacy of antidepressants in subthreshold depression could be pooled for meta-analysis (Judd et al. 2004, Rapaport et al. 2011). This was due to clinical heterogeneity and lack of availability of compatible outcome data. The pooled mean difference between baseline and endpoint HRSD-17 showed antidepressant treatment to be more effective than placebo. Whilst this finding was statistically significant, the difference was small and unlikely to represent a meaningful clinical difference. Antidepressants were significantly less well tolerated than placebo. This was true generally but also specifically in relation to number dropping out due to adverse events. This held for studies of sub-threshold depression and mild MDD alike. However, these findings should also be considered in the context of considerable risk of bias. For example, in the study by Judd et al., 2004, individuals who had previously experienced an
17
adverse event with fluoxetine were excluded (Judd et al. 2004). As such, the findings may be a conservative estimate of what is a greater difference in tolerability.
Limitations
We did not search for data from clinical trials submitted for licensing to organisation such as the US Food and Drug Administration. However we did search for references cited in reviews which focused on such sources (Kirsch et al. 2008, Turner et al. 2008, Ani et al. 2008, Khan et al. 2007). Potential biases may arise in a review where authors deviate from the plans set out in the study protocol.
The HRSD is the common gold standard for assessing efficacy of antidepressants and was used in several of the studies we identified.
As the focus of our review was
on mild MDD and sub-threshold depression, a risk of bias may emerge through using this measure. Isacsson and Adler, 2012 demonstrated that the HRSD lacks precision and sensitivity to change when used to assess milder depression (Isacsson, Adler 2011). This must be taken into consideration when interpreting our findings.
Several of the studies in this review included a placebo washout period prior to randomisation (Burrows et al. 2002, Judd et al. 2004, McGrath et al. 1985, Fabre, Putman 1987, Dunlop et al. 1990, Stewart et al. 1983). Whilst previous reviewers have excluded trials with placebo washout (Fournier et al. 2010), we viewed these studies as relevant and have included them. This method is designed to remove participants with a propensity to respond to placebo and in so doing improve the capacity of trials to assess the pharmacological effects of a drug. This means that
18
such trials are likely to underestimate the true placebo effect. It may also be one aspect that explains the difference in findings between the studies of Judd et al. 2004 (who found a significant difference favouring antidepressants using a placebo washout period) and Rapaport et al., 2011 (who did not find a difference without placebo washout).
Our review focused on symptom severity. Additionally, other aspects such as symptom duration and functional ability should also be considered in relation to likely treatment response.
Context of existing literature
Our review yielded different studies to that of the review of antidepressant efficacy in sub-threshold depression of Barbui et al., 2011. Whilst four studies were common to both reviews (Williams et al. 2000, Barrett et al. 2001, Burrows et al. 2002, Judd et al. 2004) we included a further study published in 2011 (Rapaport et al. 2011)and excluded two studies. One study was excluded as it recruited participants with anxious depression being treated in an in-patient setting (Davidson et al. 1988) and a further study excluded because a post hoc analysis was conducted (Paykel et al. 1988) of a primary care trial (Hollyman et al. 1988) which was not powered to conduct sub-group analyses. The original study by Hollyman et al., included individuals with HRSD-1720 at baseline. In comparing our review with that of Kirsch et al., 2008 (Kirsch et al. 2008) one study was common to both (Dunlop et al. 1990). This was the only study in the review of Kirsch et al., that focused on individuals meeting mild MDD criteria. Indeed it was the only study in the 35 included where the baseline HRSD was mild) with undifferentiat ed analysis
Study participants received >1 psychotropic medication or more than one drug was being tested without randomisation X
Duplicate or secondary analysis
Inability to determine mild or minor sub-group
X X X X X X X X X X X X X X X X X
X
X X X X X X X X X X X X X X X X
22
References Ackermann, R.T. & Williams, J.W.,Jr 2002, "Rational treatment choices for non-major depressions in primary care: an evidence-based review", Journal of General Internal Medicine, vol. 17, no. 4, pp. 293-301. American Psychiatric Association 2013, Diagnostic and statistical manual of mental disorders: DSM5, 5th edn, American Psychiatric Association. American Psychiatric Association 1994, Diagnostic and Statistical Manual of Mental Disorders, fourth edn, American Psychiatric Association, Washington DC. Ani, C., Bazargan, M., Hindman, D., Bell, D., Farooq, M.A., Akhanjee, L., Yemofio, F., Baker, R. & Rodriguez, M. 2008, "Depression symptomatology and diagnosis: Discordance between patients and physicians in primary care settings.", BMC Family Practice, vol. 9. Arroll, B., Macgillivray, S., Ogston, S., Reid, I., Sullivan, F., Williams, B. & Crombie, I. 2005, "Efficacy and tolerability of tricyclic antidepressants and SSRIs compared with placebo for treatment of depression in primary care: a meta-analysis", Annals of Family Medicine, vol. 3, pp. 449-456. Barbui, C., Cipriani, A., Patel, V., Ayuso-Mateos, J.L. & van Ommeren, M. 2011, "Efficacy of antidepressants and benzodiazepines in minor depression: systematic review and meta-analysis", British Journal of Psychiatry, vol. 198, no. 1, pp. 11-16. Barrett, J.E., Williams, J.W.,Jr, 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, vol. 50, no. 5, pp. 405-412. Bauer, M., Bschor, T., Pfennig, A., Whybrow, P.C., Angst, J., Versiani, M., Moller, H.J. & WFSBP Task Force on Unipolar Depressive,Disorders 2007, "World Federation of Societies of Biological Psychiatry (WFSBP) Guidelines for Biological Treatment of Unipolar Depressive Disorders in Primary Care", World Journal of Biological Psychiatry, vol. 8, no. 2, pp. 67-104. Baumeister, H. 2012, "Inappropriate prescriptions of antidepressant drugs in patients with subthreshold to mild depression: time for the evidence to become practice", Journal of affective disorders, vol. 139, no. 3, pp. 240-243. Broadhead, W.E., Blazer, D.G., George, L.K. & Tse, C.K. 1990, "Depression, disability days, and days lost from work in a prospective epidemiologic survey.", JAMA, vol. 264, no. 19, pp. 2524-2528. Burrows, A.B., Salzman, C., Satlin, A., Noble, K., Pollock, B.G. & Gersh, T. 2002, "A randomized, placebo-controlled trial of paroxetine in nursing home residents with non-major depression.", Depression & Anxiety, vol. 15, no. 3, pp. 102-110. Davidson, J.R., Giller, E.L., Zisook, S. & Overall, J.E. 1988, "An efficacy study of isocarboxazid and placebo in depression, and its relationship to depressive nosology.", Archives of General Psychiatry, vol. 45, no. 2, pp. 120-127.
23
Dimidjian, S., Hollon, S.D., Dobson, K.S., Schmaling, K.B., Kohlenberg, R.J., Addis, M.E., Gallop, R., McGlinchey, J.B., Markley, D.K., Gollan, J.K., Atkins, D.C., Dunner, D.L. & Jacobson, N.S. 2006, "Randomized trial of behavioral activation, cognitive therapy, and antidepressant medication in the acute treatment of adults with major depression.", Journal of consulting and clinical psychology, vol. 74, no. 4, pp. 658-670. Dunlop, S.R., Dornseif, B.E., Wernicke, J.F. & Potvin, J.H. 1990, "Pattern analysis shows beneficial effect of fluoxetine treatment in mild depression.", Psychopharmacology bulletin, vol. 26, no. 2, pp. 173-180. Fabre, L.F. & Putman, H.P.,3rd 1987, "A fixed-dose clinical trial of fluoxetine in outpatients with major depression.", Journal of Clinical Psychiatry, vol. 48, no. 10, pp. 406-408. Fieve, R.R., Goodnick, P.J., Peselow, E.D., Barouche, F. & Schlegel, A. 1986, "Pattern analysis of antidepressant response to fluoxetine.", Journal of Clinical Psychiatry, vol. 47, no. 11, pp. 560-562. Fournier, J.C., DeRubeis, R.J., Hollon, S.D., Dimidjian, S., Amsterdam, J.D., Shelton, R.C. & Fawcett, J. 2010, "Antidepressant drug effects and depression severity: a patient-level meta-analysis", JAMA, vol. 303, no. 1, pp. 47-53. Gibbons, R.D., Hur, K., Brown, C.H., Davis, J.M. & Mann, J.J. 2012, "Benefits from antidepressants: synthesis of 6-week patient-level outcomes from double-blind placebo-controlled randomized trials of fluoxetine and venlafaxine.", Archives of General Psychiatry, vol. 69, no. 6, pp. 572-579. Godlee, F. 2013, "Don't keep taking the tablets", BMJ, vol. 347, pp. f7438. Hamilton, M. 1960, "A rating scale for depression", Journal of Neurology, Neurosurgery and Psychiatry, vol. 23, pp. 56-62. Higgins, J.P., Altman, D.G., Gotzsche, P.C., Juni, P., Moher, D., Oxman, A.D., Savovic, J., Schulz, K.F., Weeks, L., Sterne, J.A., Cochrane Bias Methods, G. & Cochrane Statistical Methods, G. 2011, "The Cochrane Collaboration's tool for assessing risk of bias in randomised trials.", BMJ, vol. 343, pp. 5928. Hollyman, J.A., Freeling, P., Paykel, E.S., Bhat, A. & Sedgwick, P. 1988, "Doubleblind placebo-controlled trial of amitriptyline among depressed patients in general practice.", Journal of the Royal College of General Practitioners, vol. 38, no. 314, pp. 393-397. Isacsson, G. & Adler, M. 2011, "Randomized clinical trials underestimate the efficacy of antidepressants in less severe depression", Acta Psychiatrica Scandinavica, , pp. no-no. Judd, L.L., Akiskal, H.S.:.M., J.D., Zeller, P.J., Endicott, J., Coryell, W., Paulus, M.P., Kunovac, J.L., Leon, A.C., Mueller, T.I., Rice, J.A. & Keller, M.B. 1998, "A prospective 12-year study of subsyndromal and syndromal depressive symptoms in unipolar major depressive disorders", Archives of General Psychiatry, vol. 55, pp. 694-700.
24
Judd, L.L., Rapaport, M.H., Yonkers, K.A., Rush, A.J., Frank, E., Thase, M.E., Kupfer, D.J., Plewes, J.M., Schettler, P.J. & Tollefson, G. 2004, "Randomized, placebo-controlled trial of fluoxetine for acute treatment of minor depressive disorder", American Journal of Psychiatry, vol. 161, no. 10, pp. 1864-1871. Kendrick, T., Chatwin, J., Dowrick, C., Tylee, A., Morriss, R., Peveler, R., Leese, M., McCrone, P., Harris, T., Moore, M., Byng, R., Brown, G., Barthel, S., Mander, H., Ring, A., Kelly, V., Wallace, V., Gabbay, M., Craig, T. & Mann, A. 2009, Randomised controlled trial to determine the cost-effectiveness of fluoxetine for mild to moderate depression with somatic symptoms in primary care THREshold for AntiDepressant treatment (THREAD) Study, Health Technology Assessment. Kessler, R.C., Zhao, S., Blazer, D.G. & Swartz, M. 1997, "Prevalence, correlates, and course of minor depression and major depression in the national comorbidity survey", Journal of affective disorders, vol. 45, no. 1-2, pp. 19-30. Khan, A., Schwartz, K., Kolts, R.L., Ridgway, D. & Lineberry, C. 2007, "Relationship between depression severity entry criteria and antidepressant clinical trial outcomes.", Biological psychiatry, vol. 62, no. 1, pp. 65-71. Kirsch, I., Deacon, B.J., Huedo-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 / Public Library of Science, vol. 5, no. 2, pp. e45. MacGillivray, S., Arroll, B., Hatcher, S., Ogston, S., Reid, I., Sullivan, F., Williams, B. & Crombie, I. 2003, "Efficacy and tolerability of selective serotonin reuptake inhibitors compared with tricyclic antidepressants in depression treated in primary care: systematic review and meta-analysis", BMJ, vol. 326, no. 7397, pp. 1014. McGrath, P.J., Rabkin, J.G., Stewart, J.W., Harrison, W., Quitkin, F.M. & Markowitz, J. 1985, "Placebo-controlled study of mianserin in depressed outpatients.", Neuropsychobiology, vol. 14, no. 3, pp. 128-132. Montgomery, S.A. & Asberg, M. 1979, "A new depression scale designed to be sensitive to change", British Journal of Psychiatry, vol. 134, pp. 382-389. NHS National Institute for Clinical Excellence 2009, Depression: the treatment and management in adults (update). Nutt, D.J. 2011, "Highlights of the international consensus statement on major depressive disorder.", Journal of Clinical Psychiatry, vol. 72, no. 6, pp. e21. Oxman, T.E. & Sengupta, A. 2002, "Treatment of minor depression", American Journal of Geriatric Psychiatry, vol. 10, no. 3, pp. 256-264. Paykel, E.S., Hollyman, J.A., Freeling, P. & Sedgwick, P. 1988, "Predictors of therapeutic benefit from amitriptyline in mild depression: a general practice placebo-controlled trial", Journal of Affective Disorders, vol. 14, no. 1, pp. 83-95.
25
Pincus, H.A., Davis, W.W. & McQueen, L.E. 1999, "'Subthreshold' mental disorders. A review and synthesis of studies on minor depression and other 'brand names'", British Journal of Psychiatry, vol. 174, pp. 288-296. Rapaport, M.H., Judd, L.L., Schettler, P.J., Yonkers, K.A., Thase, M.E., Kupfer, D.J., Frank, E., Plewes, J.M., Tollefson, G.D. & Rush, A.J. 2002, "A descriptive analysis of minor depression.", American Journal of Psychiatry, vol. 159, no. 4, pp. 637-643. Rapaport, M.H., Nierenberg, A.A., Howland, R., Dording, C., Schettler, P.J. & Mischoulon, D. 2011, "The treatment of minor depression with St. John's Wort or citalopram: failure to show benefit over placebo.", Journal of psychiatric research, vol. 45, no. 7, pp. 931-941. Rush, A.J., First, M.B. & Blacker, D. 2008, Handbook of Psychiatric Measures, Second edn, American Psychiatric Association, USA. Song, F., Freemantle, N., Sheldon, T.A., House, A., Watson, P., Long, A. & Mason, J. 1993, "Selective serotonin reuptake inhibitors: meta-analysis of efficacy and acceptability", BMJ, vol. 306, pp. 683-687. Stewart, J.W., Quitkin, F.M., Liebowitz, M.R., McGrath, P.J., Harrison, W.M. & Klein, D.F. 1983, "Efficacy of desipramine in depressed outpatients. Response according to research diagnosis criteria diagnoses and severity of illness.", Archives of General Psychiatry, vol. 40, no. 2, pp. 202-207. Taylor, D., Paton, C. & Kapur, S. 2009, Maudsley Prescribing Guidelines, 10th edition, 10th edn, Informa Healthcare, London. The Nordic Cochrane Centre, The Cochrane Collaboration 2012, . 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, vol. 358, no. 3, pp. 252-260. Williams, J.W.,Jr, 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, vol. 284, no. 12, pp. 1519-1526.
Screening
Identification
World Health Organisation , International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Version for 2007. Available: http://www.who.int/classifications/apps/icd/icd10online/ [2008, 05/08].
Recordsidentifiedthrough databasesearching (n=5807)
Additionalrecordsidentified throughothersources (n=3)
Recordsscreened (n=5810)
Recordsexcluded Duplicates,notRCT, 26 sampleswithcomorbidity, secondaryanalysisof includedstudy
Figure 1 Process of identifying and including studies
27
102
78 -2.4 3.7 24 -7.09 3.79 78.2% -1.70 [-2.86, -0.54] 21.8% -0.27 [-2.46, 1.92]
102 100.0% -1.39 [-2.41, -0.36]
79 23
-10
-5 0 5 10 Favours SSRI Favours Placebo
Mean Difference IV, Fixed, 95% CI
Figure 4: Risk ratio of response (much or very much improved on CGI) for one study of mianserin for Mild MDD
Figure 3: Mean difference in efficacy (change score of HDRS from baseline) for 3 dosing schedules of flusoxetine in one study of Mild MDD
Figure 2 Mean difference in efficacy (change score of HDRS from baseline) for two pooled Sub-threshold depression studies
Heterogeneity: Chi² = 1.28, df = 1 (P = 0.26); I² = 22% Test for overall effect: Z = 2.66 (P = 0.008)
Total (95% CI)
-4.1 3.7 -7.36 3.87
Judd 2004 Fluoxetine Rapaport 2011 Citalopram
SSRI Placebo Mean Difference SD Total Mean SD Total Weight IV, Fixed, 95% CI
Mean
Study or Subgroup
28
Figure 5: Risk ratio of study drop-out due to adverse events for 4 studies of sub-threshold depression and 1 at 3 dosage levels of Mild MDD
29
McGrath 1985
Fabre 1987
Dunlop 1990
Williams 2000
Rapaport 2011
Judd 2004
Burrows 2002
Barrett 2001
Paroxetine 20(10-40mg) Placebo Paroxetine 10 - 30 mg Placebo Fluoxetine 20(10–20mg) Placebo Citalopram 20 mg Placebo Paroxetine 20(10-40mg) Placebo Fluoxetine 20 mg Fluoxetine 40 mg Fluoxetine 60 mg Placebo Fluoxetine 20 mg Fluoxetine 40 mg Fluoxetine 60 mg Placebo Mianserin 30 – 150 mg Placebo 50
6 42
11
11
56 10
104
105
70 107
25 69
80 27
12 81
39 12
38
92
38
372
138
52
162
24
77
Table 1 Characteristics of included studies Study Intervention N Total N
RDC DD HRSD-21 (
