Krogh et al.
Systematic Reviews (2015) 4:40
DOI 10.1186/s13643-015-0030-6
PROTOCOL Open Access
Exercise for patients with major depression:
a protocol for a systematic review with meta-
analysis and trial sequential analysis
Jesper Krogh1*, Helene Speyer1, Christian Gluud2 and Merete Nordentoft1
Abstract
Background: The lifetime prevalence of major depression is estimated to affect 17% of the population and is
considered the second largest health-care problem globally in terms of the number of years lived with disability.
The effects of most antidepressant treatments are poor; therefore, exercise has been assessed in a number of
randomized clinical trials. A number of reviews have previously analyzed these trials; however, none of these
reviews have addresses the effect of exercise for adults diagnosed with major depression.
Methods/design: The objective of this systematic review is to investigate the beneficial and harmful effects of
exercise, in terms of severity of depression, lack of remission, suicide, and so on, compared with treatment as usual
with or without co-interventions in randomized clinical trials involving adults with a clinical diagnosis of major
depression. A meta-analysis of the effect estimates of the individual trials, taking bias risk into consideration, will be
carried out. Any heterogeneity will be explored using meta-regression and subgroup analyses. Trial sequential
analysis will be carried out on the trials to control for risks of random errors. The results from the study will aid
health authorities and clinicians to understand whether exercise should be offered to patients with major
depression.
Keywords: Major depression, Depression, Exercise, Physical activity, Systematic review, Aerobic exercise, Strength
training
Background is poor; the dropout rate in clinical trials is reported to be
Depression is a common disease affecting up to 17% of between 12% and 40% within the initial 6 to 8 weeks of
the population during their lifetime [1]. Based on data treatment [3,8].
from the WHO, depression is thought to be the second The weakness of evidence for the beneficial effect of
largest health-care problem globally, in terms of years treatment, along with problems related to cost, harm,
lived with disability (YLD) [2]. Depression is also observed and low compliance, has resulted in an interest in using
as a co-morbidity in a number of somatic diseases, signifi- alternative or complementary therapies. The use of exer-
cantly contributing poorer outcomes in diseases such as cise as an intervention has attracted a lot of attention,
cancer, ischaemic heart disease, and diabetes. Depending and various forms of exercise varying in intensity have
on its severity, depression is often treated using psycho- been assessed in a number of randomized clinical trials
therapy, antidepressants, or a combination of both. How- to test their effectiveness as a treatment for patients with
ever, the clinical efficacy of antidepressants [3,4] and depression.
psychotherapy [5-7] has been challenged. Both treatments In 2011, the authors of this paper published a meta-
are costly in terms of time and money and may also have analysis of randomized clinical trials examining the ef-
adverse effects. Compliance with antidepressant treatment fect of exercise on depressive symptoms in patients with
clinical depression [9]. The results suggested that refer-
* Correspondence: [Link]@[Link] ring patients with clinical depression to exercise pro-
1
Mental Health Centre Copenhagen, Faculty of Health Sciences, University of grams was associated with a small to moderate effect on
Copenhagen, Bispebjerg Bakke 23, opg. 13a, DK-2400 Copenhagen, Denmark
Full list of author information is available at the end of the article
depressive symptoms. However, restricting the analysis
© 2015 Krogh et al.; licensee BioMed Central. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License ([Link] which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain
Dedication waiver ([Link] applies to the data made available in this article,
unless otherwise stated.
Krogh et al. Systematic Reviews (2015) 4:40 Page 2 of 7
to three trials with a low risk of bias, the effect estimate No restriction to type of publication (that is, we will
was non-significant. Since 2011, other reviews have been include abstracts and full text reports).
published on the effect of exercise on depressive symp-
toms [10], in older people [11] and in patients with Exclusion criteria
chronic illnesses [12]. However, none of these reviews
addressed the specific population of adults diagnosed Trials measuring depression immediately after a
with major depression according to valid diagnostic cri- single bout of exercise.
teria, such as the International Classification of Diseases Trials comparing one form of exercise versus
[13] or the Diagnostic and Statistical Manual of Mental another.
Disorders [14]. The reviews contained a number of trials Trials comparing different exercise intensities
that included volunteers who were defined as being de- without including a control group.
pressed on the basis of psychometric testing (for ex-
ample, Beck Depression Inventory [15]), as opposed to Interventions
individuals with a clinical diagnosis of major depression.
Furthermore, several randomized clinical trials investi- The trials had to allocate participants to an exercise
gating the effect of exercise in clinically depressed indi- intervention versus a control group (that is, exercise
viduals have been published since our 2011 review. versus a control group receiving no intervention or
treatment as usual or an attention control using
Objectives light exercise) or using exercise as an add-on-
The objective of this present systematic review is to in- treatment (that is, exercise plus medication in the
vestigate the beneficial and harmful effects of exercise, experimental group versus medication alone in the
in terms of severity of depression, lack of remission, control group).
quality of life, suicide, and so on, compared with or Exercise intervention is defined as a systematic
without co-interventions in adults with a clinical diagno- physical intervention with the intention to increase
sis of major depression. muscle strength and/or cardiovascular fitness. A
Apart from including new trials, the current systematic control group could include no treatment or only an
review differs from the previous study [9]. The current attention control using light exercise. However, it
review only considers trials including participants with a should specifically be mentioned by the authors that
diagnosis of major depression and does not include pa- the intervention is intended to be a control
tients referred with depressive symptoms. The harmful intervention. Light exercise would be equivalent to
effects of exercise interventions are also addressed, and stretching or light aerobic exercise.
bibliographical searches have been extended to include a
Chinese and a South-American database. Outcomes
The primary outcomes are 1) depressive symptoms mea-
Methods/design sured on a continuous scale assessed at the end of the
intervention; 2) lack of remission, that is, a binary out-
This systematic review will only include randomized come of the proportion of participants in each interven-
clinical trials. This protocol is not registered with tion group of the trial who did not obtain remission at
PROSPERO. the end of the intervention according to the authors’
own definition; and 3) serious adverse events defined ac-
Inclusion criteria cording to ICH-GCP as any untoward medical occur-
rence that was life threatening, resulted in death or
Participants should be diagnosed as having major persistent or significant disability (ICH-GCP 1997).
depression according to a valid and recognized Serious adverse events will accordingly include suicide
diagnostic system (that is, Research Diagnostic attempts as well as suicides. The secondary outcomes
Criteria (RDC) [16], International Classification of are non-serious adverse events, depressive symptoms,
Diseases (ICD) [13], or Diagnostic and Statistical and lack of remission assessed beyond the intervention.
Manual of Mental disorders (DSM) [14]).
Participants aged >17 years of both sexes. Search strategy
Randomized clinical trials. A trial is defined as a The search will include search CENTRAL, MEDLINE,
randomized clinical trial if the allocation of EMBASE, and Science Citation Index (Web of Science)
participants to intervention and comparison groups using medical subject headings (MESH or similar) when
is described as randomized (including terms such as possible and text word terms: depression, depressive dis-
‘randomly’, ‘random’, and ‘randomization’). order and exercise, aerobic, non-aerobic, physical activity,
Krogh et al. Systematic Reviews (2015) 4:40 Page 3 of 7
physical fitness, walking, jogging, running, bicycling, of outcome assessors, incomplete outcome data, selective
swimming, strength, and resistance. The search will also outcome reporting, for-profit bias, and other bias. Please
include LILAC (Latin American and Caribbean Health see Appendix for specifications on bias assessment.
Sciences Literature) and the Chinese Wanfang database Trials assessed as having ‘low risk of bias’ in all of the
using text word terms: depression, depressive disorder, above specified domains will be considered ‘trials with
and exercise or physical training. The flow of trial reports low risk of bias’. Trials assessed as having ‘uncertain risk
and reasons for exclusion will be presented in the of bias’ or ‘high risk of bias’ in one or more of the above
PRISMA flow chart and categorized: non-clinical popula- specified domains shall be considered trials with ‘high
tions (that is, not diagnosed according to a diagnostic sys- risk of bias’. In line with our previous systematic review
tem), review or commentary, not a randomized trial, acute [9] and the latest Cochrane review [10], trials with low
exercise (that is, studies/trials investigating the effect of a risk of bias in the allocation concealment domain,
single bout of exercise), and trials including patients with blinded outcome assessment domain, and the intention-
other psychiatric diagnoses (for example, bipolar). In to-treat analysis domain will also be characterized as tri-
addition, reference lists of relevant reviews will be als with ‘lower risk of bias.’ However, in case no or few
searched for additional trials. trials with low risk of bias will be included, we shall re-
member that the chance to know the ‘true’ intervention
Study selection effect in trials with ‘lower risk of bias’ is low or absent.
One investigator (JK) will examine titles and abstracts to
remove obviously irrelevant reports. Two investigators Data synthesis and analysis
(JK + HS) will examine the remaining full text reports In order to be able to include all of the studies in our
determining compliance with inclusion criteria. meta-analysis [25], estimates of standardized mean dif-
ference (SMD) for each individual study will be carried
Data extraction out. SMD is the mean difference in depression score be-
Two authors (JK, HS) will independently extract data tween the exercise and control groups dived by the
using a pre-piloted structured form. Any discrepancies pooled standard deviation. The result is a unit less effect
in the data extraction or inclusion/exclusion of trials will size measure, which is comparable to other studies using
be resolved by referring to the original papers. CG or other but similar measures of outcome. By convention,
MN will assist as adjudicator in cases of disagreements. SMD effect sizes of 0.2, 0.4, and 0.8 are considered
The authors will not be blinded to article results, au- small, medium, and large, respectively. For dichotomous
thors, or institutions. Data extraction will, in addition to variables, we will calculate the relative risks with a 95%
outcomes, include information regarding country of ori- confidence interval. It is expected that some trials have
gin, number of randomized participants, number of par- several intervention groups. Data from the experimental
ticipants included in efficacy analysis, mean age of groups will be pooled and compared with the data from
participants, diagnostic system, baseline assessment of the control group. In case of discrepancies between the
depression severity, type of intervention, frequency of random-effects model analysis and the fixed-effect
intervention, , duration of intervention, and recruitment model analysis, both results will be reported [26]; other-
setting (clinical vs. non-clinical). wise, only results from the random-effects analysis will
The authors JK, CG, and MN have previously pub- be reported.
lished trial reports assessing the effect of exercise in pa- The degree of heterogeneity will be quantified using
tients with depression [17,18]. To avoid academic bias, a the I-squared statistic [27], which can be interpreted as
third assessor (CH) will assist HS in bias assessment for the percentage of variation observed between the trials
these two trials. attributable to between-trial differences, rather than
sampling error (chance). Heterogeneity will be explored
Risk of bias assessment by analysis of sub-groups (see below).
Methodological studies show that trials with unclear or in- For the primary outcomes, trial sequential analysis will
adequate methodological quality regarding bias domains be attempted, based on mean differences or proportions
may be associated with bias (systematic error, the overesti- [28,29]. In order to calculate the required information size
mation of benefits, and the underestimation of harms) and the cumulative Z-curve’s eventual breach of relevant
when compared to trials using adequate methodology trial sequential monitoring boundaries, the required infor-
[19-24]. Definitions in the assessment of bias risk of a trial mation size for a primary continuous outcome will be
will be done according to the Cochrane Handbook for based on type I error of 5%, a beta of 10%, the standard
Systematic Reviews of Interventions [19] of the following error of the meta-analysis, and a minimal difference of
domains: allocation sequence generation, allocation con- three points on the HAM-D17. In order to calculate the
cealment, blinding of participants and personnel, blinding required information size and the cumulative Z-curve’s
Krogh et al. Systematic Reviews (2015) 4:40 Page 4 of 7
eventual breach of relevant trial sequential monitoring Subgroup analyses
boundaries, the required information size for the primary In subgroup analyses, the possible effects of a number of
dichotomous outcomes will be based on type I error of variables on outcomes and heterogeneity will be com-
5%, a beta of 10%, the proportion of patients in the control pared. It is expected that no, or very few, trials with low
group with the outcome, and a relative risk reduction of risk of bias will be found, and therefore, an assessment
15% or 30%. Most systematic reviews do not contain suffi- of the risk of bias by comparing trials with lower risk of
cient power [30], and if there is no significant effect of the bias is planned according to adequate allocation conceal-
intervention, it is also interesting to know whether this ment, blinded outcome assessment, and intention-to-
represents an absence of evidence (the cumulative Z- treat analysis to trials with high risk of bias according to
curve has not reached the futility area), or if it represents these domains. The effect of age will be assessed by
evidence of an absence of effect (the cumulative Z-curve comparing trials including older participants (mean age
has reached the futility area). If an absence of evidence >60 years) with trials including younger participants
persists, the likely number of participants still needed to (mean age <60 years). The effect of group versus individ-
answer the question raised can also be assessed. An inter- ual exercise will be assessed by comparing trials using
esting question is whether the trial sequential monitoring group exercises compared to trials using individual exer-
boundaries for benefit (or potentially for harms) are cises. The effect of the duration of intervention will be
crossed. This informs as to whether new trials should have assessed by comparing trials with short duration of
been stopped. Bayes factors will be calculated for all pri- intervention to trials with long duration of intervention.
mary values (the ratio between the P value probability di- The two groups formed will be based on the median
vided by the probability of the meta-analysis result, given duration of intervention employed. The effect of type of
that an anticipated intervention effect is the true effect) control group will be assessed by comparing trials with
[26]. trials, using attention control to trials with other forms
To assess the potential impact of missing data (incom- of control. Assessment of the effect of using exercise as
plete outcome data bias), a ‘best-worst’ case scenario will an add-on therapy by comparing trials using placebo/at-
be assessed, assuming that all participants lost to follow- tention control/TAU as control group to trials using
up in the intervention group had a beneficial outcome antidepressant as a control group will be carried out. In
(the group mean minus 1 standard deviation (SD)), and addition, a within-study comparison of low-dose exercise
all those with missing outcomes in the placebo group versus high-dose exercise in trials using different exercise
have had a harmful outcome (the group mean plus 1 SD intensities will be performed. The effect of co-morbid
and 2 SD). It is also planned to perform the reverse somatic disease will be assessed by comparing the effect
‘worst-best-case’ scenario analysis [26]. estimates from trials including patients with depression
Regarding the outcome of lack of remission, trials will compared to trials including patients with depression in
be included with incomplete or missing data. In case of addition to a somatic disease.
missing data for the ‘lack of remission’ outcome, missing Publication bias will be assessed by visual inspection
values will be imputed in sensitivity analysis according of a funnel plot and by Egger’s test. The meta-analyzed
to the following scenarios [31]: 1) poor outcome ana- results will be presented in a summary of findings table
lysis: assuming that none of the drop-outs/participants according to the GRADE system [32].
lost from both the experimental and the control arms
experienced the outcome, including all randomized par- Discussion
ticipants in the denominator; 2) good outcome analysis: In this systematic review, the assessment of the benefits
assuming that all of the drop-outs/participants lost from and harms of exercise interventions for adults with clin-
the experimental and the control arms experienced the ical diagnosis of major depression will be reviewed. It is
outcome, including all randomized participants in the intended to minimize selection bias by including biblio-
denominator; 3) extreme case analysis favoring the ex- graphical databases from South America (LILACS) and
perimental intervention (‘best-worse’ case scenario): China (Wanfang) in addition to standard search strat-
none of the drop-outs/participants lost from the experi- egies limited to western bibliographical databases (for
mental arm, but all of the drop-outs/participants lost example, CENTRAL, MEDLINE, EMBASE). In addition
from the control arm experienced the outcome, includ- to meta-analysis, trial sequential analysis to assess our
ing all randomized participants in the denominator; and risks of random error is planned. The final discussion
4) extreme case analysis favoring the control (‘worst- will include an analysis of the strength and limitations of
best’ case scenario): all drop-outs/participants lost from the evidence and of the current review.
the experimental arm, but none from the control arm Based on the authors’ previous review and intimate
experienced the outcome, including all randomized par- knowledge of the current subject, we expect to include
ticipants in the denominator. more than 1,000 patients diagnosed with depression
Krogh et al. Systematic Reviews (2015) 4:40 Page 5 of 7
included in randomized clinical trials. The current re- – Uncertain risk of bias: the method used to conceal
view will support health-care providers and decision the allocation was not described so that intervention
makers within the health-care system on the decision to allocations may have been foreseen in advance of, or
include exercise as a standard treatment for patients during, enrolment.
with depression. – High risk of bias: the allocation sequence was likely
to be known to the investigators who assigned the
Appendix: Search strategy participants.
We used MESH terms or similar and secondly text word
terms. Blinding of participants and personnel
We do not expect this to happen in many trials, but in
1. Depression case investigators have tried to blind these groups, we
2. Depressive disorder will use the following:
3. Major Depression
4. Exercise – Low risk of bias: blinding was performed
5. Aerobic adequately.
6. Non-aerobic – Uncertain risk of bias: there was insufficient
7. Physical activity information to assess whether blinding was likely to
8. Physical fitness induce bias on the results.
9. Walk* – High risk of bias: no blinding or incomplete
10. Jog* blinding.
11. Run*
12. Bicycling
13. Swim* Blinding of outcome assessors
14. Strength We do not expect this to happen in many trials, but in
15. Resistance case investigators have tried to blind the outcome asses-
16. 1 OR 2 OR 3 sors, we will use the following:
17. 4 OR 5 OR 6 OR 7 OR 8 OR 9 OR 10 OR 11 OR
12 OR 13 OR 14 OR 15 – Low risk of bias: the assessment of outcomes was
18. 16 AND 17 not likely to be influenced by lack of blinding.
– Uncertain risk of bias: there was insufficient
Appendix: Bias assessment information to assess whether blinding was likely to
Allocation sequence generation induce bias on the results.
– High risk of bias: the assessment of outcomes was
– Low risk of bias: sequence generation was achieved likely to be influenced by lack of blinding.
using computer random number generation or a
random number table. Drawing lots, tossing a coin, Incomplete outcome data
shuffling cards, and throwing dice are adequate if
performed by an independent person not otherwise – Low risk of bias: missing data were unlikely to make
involved in the trial. treatment effects depart from plausible values. Suf-
– Uncertain risk of bias: the method of sequence ficient methods, such as multiple imputation, have
generation was not specified. been employed to handle missing data.
– High risk of bias: the sequence generation method – Uncertain risk of bias: there was insufficient
was not random. information to assess whether missing data in com-
bination with the method used to handle missing
Allocation concealment data were likely to induce bias on the results.
– High risk of bias: the results were likely to be
– Low risk of bias: the participant allocations could biased due to missing data.
not have been foreseen in advance of, or during,
enrolment. Allocation was controlled by a central Analyses is considered intention to treat if missing
and independent randomization unit. The allocation data is handled by adequate methods (mixed models,
sequence was unknown to the investigators (for multiple imputations, or similar) or if no missing data
example, if the allocation sequence was hidden in was encountered (last observation carried forward was
sequentially numbered, opaque, and sealed envelopes). considered inadequate).
Krogh et al. Systematic Reviews (2015) 4:40 Page 6 of 7
Selective outcome reporting – High risk of bias: there are other factors in the trial
that could put it at risk of bias (for example,
– Low risk: all predefined, or clinically relevant and baseline imbalance or academic bias).
reasonably expected, outcomes are reported on. If
Abbreviations
the original trial protocol is available, the outcomes DSM: Diagnostic and Statistical Manual of Mental Disorders; HAM-D17: Hamilton
should be those called for in that protocol. (Note: If Depression Rating Scale - 17 Items; ICD: International Classification of Diseases;
the trial protocol is obtained from a trial registry ICH-GCP: International Conference on Harmonization of Technical Requirements
for Registration of Pharmaceuticals for Human Use - Guideline for Good Clinical
(for example, [Link]), the outcomes Practice; LILACS: Latin American and Caribbean Health Sciences Literature;
to be sought are those enumerated in the original RDC: Research Diagnostic Criteria; SMD: standardized mean difference;
protocol if the trial protocol was registered before WHO: World Health Organization; YLD: years lived with disability.
or at the time the trial has begun; if the trial
Competing interests
protocol was registered after the trial has begun, JK, CG, and MN have previously published trials assessing the effect of
those outcomes will not be considered to be exercise in patients with depression [17,18]. Otherwise, the authors declare
reliable in representing the outcomes initially being that they have no competing interests.
sought.) If the trial protocol is not available (or if
Authors’ contributions
the protocol was registered after the trial was JK conceived the idea for the systematic review and drafted the first
begun), the review authors will decide, when they manuscript. HS, CG, and MN participated in the design of the systematic
are writing the protocol for the systematic review, review and revised the manuscript for intellectual content. All authors read
and approved the final manuscript.
what clinically relevant and reasonably expected
outcomes would be and will explicitly state those Acknowledgements
outcomes in the pertinent methodology part of There are no funding sources to disclose.
their protocol for the systematic review. Author details
– Unclear risk: not all predefined, or clinically 1
Mental Health Centre Copenhagen, Faculty of Health Sciences, University of
relevant and reasonably expected, outcomes are Copenhagen, Bispebjerg Bakke 23, opg. 13a, DK-2400 Copenhagen, Denmark.
2
Copenhagen Trial Unit, Centre for Clinical Intervention Research,
reported fully, or it is unclear whether data on Rigshospitalet, Copenhagen University Hospital, Blegdamsvej 9, DK-2100
these outcomes were recorded or not. Copenhagen Ø, Denmark.
– High risk: one or more predefined or clinically
Received: 5 August 2014 Accepted: 11 March 2015
relevant and reasonably expected outcomes were
not reported, despite the fact that data on these
outcomes should have been likely to have been References
available and even recorded. 1. Lepine JP, Gastpar M, Mendelwicz J, Tylee A. Depression in the community:
the first pan-european study DEPRES (Depression Research in European
Society). Int Clin Psychopharmocol. 1997;12:19–29.
2. Ustun TB, Ayuso-Mateos JL, Chatterji S, Mathers C, Murray CJL. Global burden
For-profit bias of depressive disorders in the year 2000. Br J Psychiatry. 2004;184:386–92.
3. Kirsch I, Deacon B, Huedo-Medina T, Scoboria A, Moore T, Johnson B. Initial
severity and antidepressant benefits: a meta-analysis of data submitted to
– Low risk of bias: the trial appears to be free of the Food and Drug Administration. PLoS Med. 2008;5:e45. doi:10.1371/jour-
industry sponsorship or other type of for-profit [Link].0050045.
support that may manipulate the trial design, 4. Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R. Selective
publication of antidepressant trials and its influence on apparent efficacy. N
conductance, or results of the trial. Engl J Med. 2008;358:252–60.
– Uncertain risk of bias: the trial may or may not be 5. Jakobsen JC, Hansen JL, Simonsen E, Gluud C. The effect of interpersonal
free of for-profit bias as no information on clinical psychotherapy and other psychodynamic therapies versus ‘treatment as
usual’ in patients with major depressive disorder. PLoS One. 2011;6:e19044.
trial support or sponsorship was provided. 6. Jakobsen JC, Hansen JL, Storebo OJ, Simonsen E, Gluud C. The effects of
– High risk of bias: the trial was sponsored by industry cognitive therapy versus ‘no intervention’ for major depressive disorder.
or received other type of for-profit support. PLoS One. 2011;6:e28299.
7. Jakobsen JC, Lindschou HJ, Storebo OJ, Simonsen E, Gluud C. The effects of
cognitive therapy versus ‘treatment as usual’ in patients with major
depressive disorder. PLoS One. 2011;6:e22890.
8. Cipriani A, Furukawa TA, Salanti G, Geddes JR, Higgins JP, Churchill R, et al.
Other bias Comparative efficacy and acceptability of 12 new-generation antidepressants:
a multiple-treatments meta-analysis. Lancet. 2009;373:746–58.
– Low risk of bias: the trial appears to be free of 9. Krogh J, Nordentoft M, Sterne J, Lawlor D. The effect of exercise in clinically
depressed adults: systematic review and meta-analysis of randomized
other components (for example, academic bias) that controlled trials. J Clin Psychiatry. 2010;72:529–38.
could put it at risk of bias. 10. Cooney GM, Dwan K, Greig CA, Lawlor DA, Rimer J, Waugh FR, et al.
– Uncertain risk of bias: the trial may or may not be Exercise for depression. Cochrane Database Syst Rev. 2013;9:CD004366.
11. Bridle C, Spanjers K, Patel S, Atherton NM, Lamb SE. Effect of exercise on
free of other components that could put it at risk of depression severity in older people: systematic review and meta-analysis of
bias. randomised controlled trials. Br J Psychiatry. 2012;201:180–5.
Krogh et al. Systematic Reviews (2015) 4:40 Page 7 of 7
12. Herring MP, Puetz TW, O’Connor PJ, Dishman RK. Effect of exercise training
on depressive symptoms among patients with a chronic illness: a
systematic review and meta-analysis of randomized controlled trials. Arch
Intern Med. 2012;172:101–11.
13. World Health Organization. International Statistical Classification of Diseases,
10th Revision (ICD-10). Geneva, Switzerland: World Health Organization; 1992.
14. American Psychiatric Association. Diagnostic and Statistical Manual of Mental
Disorders. 4th ed. Washington DC: American Psychiatric Association; 1994.
15. Beck AT, Steer RA, Brown GK. BDI-II Manual. 2nd ed. New York: Psychological
Corporation; 1996.
16. Spitzer RL, Endicott J, Robins E. Research diagnostic criteria: rationale and
reliability. Arch Gen Psychiatry. 1978;35:773–82.
17. Krogh J, Saltin B, Gluud C, Nordentroft M. The DEMO trial: a randomized,
parallel-group, observer-blinded clinical trial of strength versus aerobic
versus relaxation training for patients with mild to moderate depression.
J Clin Psychiatry. 2009;70:790–800.
18. Krogh J, Videbech P, Thomsen C, Gluud C, Nordentoft M. DEMO-II trial.
Aerobic exercise versus stretching exercise in patients with major
depression - a randomised clinical trial. PLoS One. 2012;7(10):e48316.
19. Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of
Interventions. Chichester (UK): John Wiley & Sons; 2008.
20. Schulz K, Chalmers I, Hayes R, Altman DG. Empirical evidence of bias.
Dimensions of methodological quality associated with estimates of
treatment effects in controlled trials. JAMA. 1995;273:408–12.
21. Moher D, Pham B, Jones A, Cook D, Jada A, Moher M. Does quality of
reports of randomised trials affect estimates of intervention efficacy
reported in meta-analysis? Lancet. 1998;352:609–13.
22. Als-Nielsen B, Chen W, Gluud C, Kjaergard LL. Association of funding and
conclusions in randomized drug trials: a reflection of treatment effect or
edverse events? JAMA. 2003;290:921–8.
23. Wood L, Egger M, Gluud L, Schulz K, Jüni P, Altman D, et al. Empirical
evidence of bias in treatment effect estimates in controlled trials with
different interventions and outcomes: meta-epidemiological study. BMJ.
2008;336:601–5.
24. Kjaergard LL, Villumsen J, Gluud C. Reported methodologic quality and
discrepancies between large and small randomized trials in meta-analyses.
Ann Intern Med. 2001;135:982–9.
25. DerSimonian R, Laird N. Meta-regression analysis in clinical trials. Control Clin
Trials. 1986;7:177–88.
26. Jakobsen JC, Wetterslev J, Winkel P, Lange T, Gluud C. Thresholds for
statistical and clinical significance in systematic reviews with meta-analytic
methods. BMC Med Res Methodol. 2014;14:120.
27. Higgins JPT, Thompson SG, Deeks JJ, Altmann DG. Measuring inconsistency
in meta-analysis. BMJ. 2003;327:557–60.
28. Wetterslev J, Thorlund K, Brok J, Gluud C. Trial sequential analysis may
establish when firm evidence is reached in cumulative meta-analysis. J Clin
Epidemiol. 2008;61:64–75.
29. Brok J, Thorlund K, Gluud C, Wetterslev J. Trial sequential analysis reveals
insufficient information size and potentially false positive results in many
meta-analyses. J Clin Epidemiol. 2008;61:763–9.
30. Turner RM, Bird SM, Higgins J. The impact of study size on meta-analyses:
examination of underpowered studies in Cochrane reviews. PLoS One.
2013;8(3):e59202.
31. Hollis S, Campbell F. What is meant by intention to treat analysis? Survey of
published randomised controlled trials. BMJ. 1999;319:670–4.
32. Guyatt GH, Oxman AD, Schunemann HJ, Tugwell P, Knottnerus A. GRADE
guidelines: a new series of articles in the journal of clinical epidemiology.
J Clin Epidemiol. 2011;64:380–2.
Submit your next manuscript to BioMed Central
and take full advantage of:
• Convenient online submission
• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution
Submit your manuscript at
[Link]/submit