Journal of Affective Disorders 185 (2015) 4759

Contents lists available at ScienceDirect

Journal of Affective Disorders

journal homepage: www.elsevier.com/locate/jad

Review

Depression in patients with alcohol use disorders: Systematic review

and meta-analysis of outcomes for independent and substance-in-
duced disorders
James A. Foulds a,n, Simon J. Adamson a, Joseph M. Boden b, Jonathan A. Williman c,
Roger T. Mulder a
a
Department of Psychological Medicine, University of Otago, Christchurch, New Zealand
b
Christchurch Health and Development Study, University of Otago, Christchurch, New Zealand
c
Biostatistics and Computational Biology, Department of Public Health and General Practise, University of Otago, Christchurch, New Zealand

art ic l e i nf o a b s t r a c t

Article history: Background: In patients with an alcohol use disorder, depression is commonly categorised as in-
Received 20 February 2015 dependent (ID) or substance-induced (SID). It is not established whether these conditions respond dif-
Received in revised form ferently to treatment.
11 June 2015
Methods: MEDLINE, Embase and Cochrane databases from 1980 to 2014 were searched for studies on
Accepted 14 June 2015
Available online 23 June 2015
alcohol use disorders with coexisting depressive symptoms. Meta-analyses were conducted using ran-
dom effects models, to derive pooled effect estimates of the change in depression during treatment and
Keywords: the effect of antidepressant therapy.
Alcohol drinking Results: Twenty-two studies met inclusion criteria for the review, of which 11/22 were included in the
Alcohol-induced disorder
meta-analysis. All studies reported a large improvement in depression symptom score, most of which
Alcohol-related disorders
occurred within the rst 36 weeks of treatment. The amount of improvement during follow up was
Depressive disorder
Depression similar in studies on ID in comparison to those in undifferentiated depression. Evidence on the outcome
Antidepressive agents for SID was limited.
The effect size of antidepressant therapy compared to placebo was 0.25 (0.06, 0.44) for ID and 0.08
(
0.31, 0.47) for SID or undifferentiated depression.
Limitations: Few studies examined the natural history and treatment response of SID. There was het-
erogeneity between studies, which was partly explained by baseline depression severity.
Conclusions: Treatment for depression co-occurring with an alcohol use disorder is associated with a
large early improvement in depression, even if depression is believed to be independent of drinking. The
effect of antidepressant therapy on depression in patients with alcohol use disorders is modest, with
stronger evidence in ID.
& 2015 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.1. Subtyping depression in patients with alcohol use disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.2. Implications of subtypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.3. Aims of the review and meta-analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.1. Selection criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.2. Information sources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.3. Study selection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.4. Data items and summary measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.5. Assessment of bias and quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

n
Corresponding author. Tel.: 64 3 3720 400; fax: 64 3 3720 407.
E-mail address: james.foulds@otago.ac.nz (J.A. Foulds).

http://dx.doi.org/10.1016/j.jad.2015.06.024
0165-0327/& 2015 Elsevier B.V. All rights reserved.
48 J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759

2.6. Synthesis of results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

2.7. Antidepressant efcacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.8. Psychological treatments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.9. Predictors of treatment outcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.1. Search results and quality of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.2. Dening and measuring depression in subjects with an alcohol use disorder. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.3. Change in depression during treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.4. Moderators of depression outcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.5. Antidepressant efcacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.6. Efcacy of psychological treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.7. Predictors of depression outcome: narrative review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.1. Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.2. Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.3. Comparison with existing literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.4. Typology of depression in alcohol use disorders: a false dichotomy? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.5. Clinical implications and directions for future research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Role of funding source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Conict of interest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

1. Introduction ID requires specic depression treatment (Pettinati, 2013; Schuck-

it, 1985; Schuckit et al., 1997). In particular the view has been put
Patients entering treatment for an alcohol use disorder often forward that ID, but not SID, responds to antidepressant medica-
have high levels of depressive symptoms (Davidson, tion (Schuckit, 2006) but the evidence for this has not been sys-
1995; Schuckit, 1995). These symptoms typically improve rapidly tematically evaluated.
with treatment (Baker et al., 2013; Brown, 1988; Davidson, 1995;
Kiefer and Barocka, 1999; Schuckit, 1985, 1995) but in spite of this, 1.3. Aims of the review and meta-analysis
comorbid depression predicts worse outcomes in alcohol treat-
ment (Lejoyeux and Lehert, 2011; Pettinati, 2013) and heavy This review aimed to answer the following questions:
drinkers have an increased risk of future depression even if they
cut down (Hasin and Grant, 2002). Knowing which groups of pa- i. In depressed patients with an alcohol use disorder, how much
tients are more or less likely to improve during treatment would do depressive symptoms improve during treatment for
allow scarce treatment resources to be allocated more effectively, depression?
but to date the ability to predict patient outcomes accurately has ii. Do patients with ID and SID have different patterns of treat-
been limited. ment response?
iii. Does antidepressant efcacy differ between patients with ID
1.1. Subtyping depression in patients with alcohol use disorders and SID?

An early approach to guide treatment and predict depression

outcome in patients with an alcohol disorder was to classify de-
pression as primary or secondary according to whether it devel- 2. Methods
oped before or after the onset of heavy drinking (Schuckit, 1985).
In the 1990's, the typology evolved to also incorporate information 2.1. Selection criteria
about a past history of depression during abstinence (Schuckit
et al., 1997). The term independent was used for depression that Studies were chosen according to the following criteria:
began before the onset of alcohol dependence or during sustained
abstinence while depressive syndromes occurring only during a 1. Studies reported longitudinal data on alcohol use and depres-
period of active alcohol dependence were labelled substance-in- sion in adults over at least 8 weeks.
duced (Schuckit et al., 1997). Structured clinical assessment tools 2. Change in mean score on a validated depression scale was
such as the Structured Clinical Interview for DSM (SCID) (Spitzer reported.
et al., 1992) and the Psychiatric Research Interview for Substance 3. Subjects had an active alcohol use disorder (alcohol depen-
and Mental Disorders (PRISM) (Hasin et al., 1996) have helped to dence or alcohol abuse) diagnosed according to DSM or ICD
operationalise these denitions and introduce greater diagnostic criteria.
reliability. 4. Mean baseline score Z10 on the 17-item Hamilton Depression
Rating Scale (HDRS) (Hamilton, 1960), or equivalent severity on
1.2. Implications of subtypes another depression rating scale. Given the uncertain nosology
of depression in alcohol use disorders, subjects were not re-
It has been argued that the distinction between independent quired to have a diagnosis of a depressive disorder. The low
depression (ID) and substance-induced depression (SID) has im- severity threshold was chosen to avoid excluding studies in
portant implications for treatment and prognosis: SID is con- which baseline measures were taken after a short period of
sidered a self-limiting condition that remits with abstinence while abstinence. A published score comparison table available at
 J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759 49

http://www.ids-qids.org/index2.html (University of Pittsburgh, during the selection process.

2014) was used to calibrate severity across scales.
5. Treatment for depressive symptoms was provided. A liberal
denition of such treatment was applied, and it could include
any form of medication or psychosocial treatment with plau-
sible efcacy in treating depressive symptoms.
Unpublished and incomplete studies were not assessed for in-
clusion. Studies were excluded if they were conducted in special
sub-populations (for example, prisoners) or those with another
psychiatric or substance use disorder (for example, borderline
personality disorder or nicotine dependence). Studies exclusively
comprised of subjects with dysthymic disorder were also excluded
as this condition has a different natural history compared to other
depressive syndromes (Judd et al., 2002).
2.2. Information sources
Information sources were MEDLINE, Embase and Cochrane
databases, searched on 9 November 2014. The reference lists from
review articles (Baker, 2012; Nunes and Levin, 2004; Ostacher,
2007; Pettinati, 2004; Salloum and Jones, 2008; Sullivan et al.,
2005) were used to hand-search for additional articles. Search
items were alcohol drinking; alcohol-induced disorder; alcohol-
related disorder; alcoholics; alcoholism and depression; anti-
depressive agents. English-language publications on clinical trials
from 1980 onwards were considered.
2.3. Study selection
The PRISMA diagram in Fig. 1 illustrates the ow of studies
Authors 1 (JF) and 2 (SA) independently screened titles and
abstracts to exclude obviously irrelevant articles. Duplicate cita-
tions were deleted. The full text of articles in the screened sample
was extracted and reviewed by authors 1 and 2. Studies meeting
inclusion criteria were then compiled into the nal sample by
consensus.
2.4. Data items and summary measures
For analyses on the change in depression during treatment,
results for each treatment arm of studies were aggregated. The
mean depression score at baseline (taken as the earliest reported
timepoint) and study completion, and their respective standard
deviations were recorded. Study completion was chosen as the
endpoint rather than a xed time interval because studies did not
report on outcomes at the same time points, and most change was
expected to occur early in treatment (Baker et al., 2014). The main
summary measure was the standardised mean change in depres-
sion score from baseline to study completion, using the pooled
standard deviation of baseline and follow up scores.
In two studies (Hernandez-Avila et al., 2004; Moak et al., 2003)
the baseline standard deviation was very low, probably due to
scores clustering just above the inclusion threshold. To avoid in-
troducing bias, the baseline standard deviations were not used for
these two studies and the follow-up standard deviations alone
were used instead.
Depression type was dened according to whether studies were
in subjects with ID, SID or a mixture of both types. Information on
demographics, duration and adjunctive treatment was also ex-
tracted as shown in Table 1.

Records identified through

searching Embase (n=376), Additional records identified
identification

MEDLINE (n=499) and Cochrane through other sources

(n=387) databases
(n=6)
(n=1260)

Unique studies identified after duplicate

citations removed (n=799)
screening

Records excluded as
Records screened obviously irrelevant
(n=799) (n=661)

Full-text articles excluded

eligibility

Full- text articles assessed (n=116)

for eligibility
(n=138) Did not meet depression criteria or
did not adequately report depression
outcomes (n=54)

Outcomes not reported over at least

Excluded from 8 weeks (n=7)
Studies included in
meta-analysis due
qualitative synthesis Mixed substance use disorder
to lack of data on sample (n=7)
(n=22)
meta- analysis

intent-to-treat
sample (n=6); Low quality (n=10)
outcome measure Not meeting alcohol criteria (n=6)
not Hamilton
Depression Rating Studies included Foreign language studies with no
translation available (n=6)
Scale(n=5) in meta- analysis (n=11)
Other reasons (n= 26)

Fig. 1. Selection of systematic review sample: PRISMA ow diagram.

 50
Table 1
Characteristics of studies included in review.

Study Setting Sample com- Inclusion criteria Depression mea- Depression type Treatment conditions Adjunctive Main outcomes Quality and bias
position sures and base- treatment
line score (sd)

Adamson Outpatient ad- n 138, 46% DSM-IV Alcohol depen- MADRS at baseline Independent and sub- Citalopram vs placebo Manualised clin- No difference between ci- Jadad score 5
et al. diction clinics. male dence and MDD and 24 weeks stance-induced depression. ical case talopram and placebo.
(2015) No abstinence Baseline Independent started be- management Substance-induced de-
requirement MADRS 31 (5.8) fore heavy drinking or pression had better de-
present during 4 4 weeks pression outcomes
abstinence
Agyapong Specialist in- n 54, 46% DSM-IV Alcohol depen- BDI at baseline and Not assessed Text message support vs Routine outpatient No difference between text Jadad score 3
et al. patient dual di- male dence or abuse and 6 months treatment as usual follow up message support and
(2013) agnosis service. MDD Baseline treatment as usual
Baseline taken BDI 31.8 (8.6)
at week 3 of
inpatient stay

J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759

Altintoprak Specialised in- n 44, 61% DSM-IV Alcohol depen- HDRS-17 after 10 Not assessed but only pa- Amitriptyline vs No difference between Jadad score 4.
et al. patient addic- male dence and MDD 14 days detoxica- tients with HDRS Z14 after mirtazapine amitriptyline and Analyses were
(2008) tion unit. Base- tion and at detoxication were mirtazapine based on study
line taken after 8 weeks randomised completers only.
detoxication Baseline
HDRS 23.9
(4.6)
Brown et al. Intensive out- n 166, 67% DSM-IV Alcohol BDI at baseline and Independent and sub- Cognitive behavioural Abstinence-or- No difference between Jadad score 2.
(2011) patient. Mini- male dependenceBDI Z 15 52 weeks stance-induced depression. treatment for depression iented group cognitive behavioural Description of ran-
mum 2 days Independent started42 vs relaxation training treatment treatment and relaxation domisation proce-
abstinence Baseline BDI 25.3 weeks before onset of hea- training dure unclear.
(8) vy substance use or con-
tinued during 4 weeks
abstinence
Cornelius Psychiatric in- n 51, 51% DSM-IIIR alcohol de- HDRS-24 at base- Primary depression, de- Fluoxetine vs placebo Supportive Fluoxetine more effective Jadad score 3
et al. patients. 9 days male pendence and MDD line (pre-washout) ned as the condition [] psychotherapy than placebo for
(1997) abstinence persisting after comple- and 12 weeks responsible for occasioning depression
tion of detoxication Baseline the evaluation
HDRS 33.1 pre-
detoxication
and 18.6 (8.2)
after
detoxication
Davis et al. Subgroup ana- n 536, 36% PDSQ alcohol disorder QIDS at baseline Not assessed Open-label citalopram Subjects with alcohol use High quality study.
(2010) lysis from male and 14 weeks disorder had similar de- Medication was
STAR*D. Psy- MDD pression outcomes to those supplied at no cost
chiatric out- HDRS17 Z 14 Baseline without by drug
patients. No ab- QIDS 15.9 (4) manufacturer.
stinence
requirement
Dorus et al. Inpatient veter- n 457, 100% DSM-III alcohol BDI after 3 weeks Not assessed Lithium vs placebo Individual and No difference between li- Jadad score 3 but
(1989) ans. 3 weeks male dependence abstinence and 52 group therapy plus thium and placebo not clear if results
abstinence weeks alcoholics presented are from
Baseline anonymous intent to treat
BDI 14.0 sample.
Gual et al. Outpatients. n 83, 53% DSM-IV/ICD-10 alcohol HDRS-17 after 2 or Secondary depression Sertraline vs placebo No difference between Jadad score 4
(2003) Abstinence male dependence and MDD/ more weeks ab- sertraline and placebo
2 weeks DD. Patients with a pri- stinence and at 24
mary psychiatric dis- weeks
order were excluded Baseline
HDRS 13.4
 (4.9)
Hernandez- Outpatients free n 41, 49% DSM-IV alcohol de- HDRS-17 at base- Depression diagnosed 1 1 week placebo lead in Manual-guided No difference between ne- Jadad score 4.
Avila et al. of heavy drink- male pendence major line (after cessa- weeks after heavy drinking then nefazodone vs supportive therapy fazodone and placebo Very low baseline
(2004) ing in past week depressionHDRS Z 17 tion of heavy ceased placebo depression stan-
drinking for dard deviation.
1 week) and 10 Partially funded by
weeks drug manufacturer.
Baseline
HDRS 19.6 (1.9)
at intake and
16.3 (2.3) at
randomisation
Janiri et al. Mixed inpatient n 50, 80% DSM-III-R alcohol HDRS-17 after de- Not assessed Fluoxetine vs placebo Psychological in- No difference between Jadad score 4 but
(1996) and outpatient. male dependence toxication and terviews and alco- uoxetine and placebo not clear if results
Abstinence 8 weeks holics anonymous presented are for
7 days Baseline intent-to-treat
HDRS 16.7 sample
(11.8)
Kranzler Outpatients. No n 328, 64% DSM-IV alcohol depen- HDRS-17 at day Patients stratied by de- Sertraline vs placebo Manualised sup- No difference between Jadad score 4.

J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759

et al. abstinence male dence and MDD, 7 to 14 during pression response during portive therapy sertraline and placebo for Funded by drug
(2006) requirement HDRSZ 17 at screening placebo lead-in placebo lead-in either group manufacturer.
period, and at 10
weeks
Baseline
HDRS 20.6
(5.1) for group A
and 12.6 (5.1) for
group B
Mason et al. Outpatients. n 28, 86% DSM-III-R alcohol HDRS-24 after 13 Secondary depression, de- Desipramine vs placebo 4-weekly clinic Desipramine superior to Jadad score 4.
(1996) Abstinence male dependence weeks abstinence ned as onset after onset of monitoring placebo for depression Completion rate
7 days and at week 26 alcoholism only 23%
Baseline
HDRS 19.8
(10.4)
McGrath Psychiatric out- n 69, 49% DSM-IIIR alcohol de- HDRS-21 at base- Primary depression, de- 1-week placebo lead-in. Non-manualised Imipramine superior to Jadad score 4
et al. patients. No ab- male pendence or abuse and line and 12 weeks ned as onset prior to al- Placebo non-responders weekly individual placebo for depression
(1996) stinence MDD, DD or DDNOS HDRS 14.9 cohol abuse or during based on Clinical Global counselling and
requirement (5.2) 6 months abstinence Impression randomised to alcoholics
imipramine vs placebo anonymous
Moak et al. Outpatients. No n 82, 61% DSM-III-R mild to mod- HDRS-21 at base- Independent depression, 1-week placebo lead-in. Manualised CBT Sertraline was more effec- Jadad score 5.
(2003) abstinence male erate alcohol depen- line and 12 weeks dened as depression pre- Placebo non-responders tive than placebo for fe- Very low baseline
requirement dence and MDD/DD, dating alcoholism or per- (HDRS 17) after lead-in male but not male patients standard deviation.
HDRS21 Z 17 sisting during 4 weeks randomly assigned to ser- Drug manufacturer
abstinence, or independent traline vs placebo supplied drug for
Baseline affective disorder in rst study.
HDRS 19.1 (2.5) degree relative
Muhonen Outpatients. No n 80, 55% DSM-IV alcohol depen- MADRS at baseline Not assessed Memantine vs Medical No difference between Jadad score 5
et al. abstinence male dence and MDD; and 26 weeks escitalopram management memantine and
(2008) requirement BDI Z17 Baseline escitalopram
MADRS 26.3
(4.3)
Oslin (2005) Outpatients age n 74, 80% DSM-IV alcohol depen- HDRS-17 after at Substance-induced or pri- Naltrexone vs placebo Sertraline plus No difference between Jadad score 1.
55 Ab- male dence and MDD least 3 days ab- mary. Primary depression manualised com- naltrexone and placebo. Drug manufacturer
stinence 3 days stinence and at 12 before onset of alcohol de- pliance enhance- supplied drug for
weeks pendence or during ment therapy study.
Baseline 3 months abstinence
HDRS 21.8 (5.4)
Paparrigo- Inpatients. No n 120, 74% DSM-IV alcohol HDRS at admission Secondary depression. Open label tiagabine vs Cognitive-beha- Tiagabine was safe and Jadad score 1. Not
poulos abstinence male dependence for detoxication Subjects with depression treatment as usual viour oriented well tolerated clear if results pre-
et al. requirement and at 52 weeks. prior to onset of alcoholism therapy sented are for

51
 52
Table 1 (continued )

Study Setting Sample com- Inclusion criteria Depression mea- Depression type Treatment conditions Adjunctive Main outcomes Quality and bias
position sures and base- treatment
line score (sd)

(2010) HDRS version not or during abstinence were intent-to-treat

stated excluded sample. Multiple
Baseline other quality
HDRS 39.8 (7.8) issues.
pre-
detoxication
Petrakis Outpatient ve- n 254, 97% DSM-IV alcohol HDRS-17 after at Not assessed Placebo vs naltrexone vs Compliance en- All groups had similar im- Jadad score 3.
et al. terans. Ab- male dependence least 3 days ab- disulram vs hancement ther- provement in depression Medication effect
(2007) stinence 3 days stinence and at 12 naltrexone disulram apy; analyses were
weeks antidepressants based on intent-to-
Baseline treat but follow-up
HDRS 10.1 (5.7) HDRS scores were
only reported for

J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759

completers.
Pettinati Outpatients. n 170, 62% DSM-IV alcohol depen- HDRS-17 after at Independent depression, Placebo vs sertraline vs Manualised in- Patients receiving com- Jadad score 4.
et al. Abstinence male dence and MDD; least 3 days ab- dened using SCID-P cri- naltrexone vs dividual CBT bined naltrexone sertra- Drug manufacturer
(2010) 3 days HDRS17 4 10 stinence and at 14 teria, i.e. depression prior sertraline naltrexone line had better depression supplied drug for
weeks to onset of heavy drinking outcomes than those re- study.
Baseline and/or during 4 weeks ceiving placebo or either
HDRS 23.0 abstinence treatment alone
(6.4)
Pettinati Outpatients. n 53, 52% DSM-III-R alcohol de- HDRS-24 at ran- Not assessed Sertraline vs placebo Manualised coun- No difference between Jadad score 4
et al. Two-week lead- male pendence and lifetime domisation (after sellor-delivered sertraline and placebo
(2001)a in period to history of MDD (29/53 2-week lead-in) TSF
maximise had current MDD) and at 14 weeks
abstinence Baseline
HDRS 12.6 (8.1)
after
detoxication
Roy-Byrne Outpatients. No n 64, 45% DSM-III-R alcohol de- HDRS-17 at base- Independent depression, Nefazodone vs placebo Group cognitive- Nefazodone group had Jadad score 3.
et al. abstinence male pendence and MDD line and at 12 dened as a history of de- behavioural skills lower depression scores at Analyses of treat-
(2000) requirement weeks pression during 4 weeks training plus baseline and follow-up ment effect did not
Baseline abstinence psychoeducation adjust for baseline
HDRS 23.9 score differences.
(5.2)
Yoon et al. Mixed. No ab- n 184, 77% DSM-IV alcohol depen- HDRS-17 at base- Not assessed Open label mirtazapine Supportive Mirtazapine was well tol- Minor quality is-
(2006) stinence male dence and MDD; line and 8 weeks psychotherapy erated and associated with sues noted.
requirement HDRS17 Z 8 Baseline large reduction in depres-
HDRS 22.5 sion score
(8.4)

List of abbreviations: DSM American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders; BDI Beck Depression Inventory; HDRS Hamilton Depression Rating Scale; MADRS Montgomery Asberg
Depression Rating Scale; MDD major depressive disorder; DD dysthymic disorder; CBT cognitive behaviour therapy; TSF twelve-step facilitation.
a
Results for lifetime depression subgroup presented.
 J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759
2.5. Assessment of bias and quality
Individual study bias was addressed by recording follow-up
completion rates, and including in the meta-analysis only studies
where reported outcomes were clearly from the intention-to-treat
sample. External funding sources were noted. Study quality in
randomized trials was assessed according to the adequacy of se-
quence generation, allocation concealment, blinding and com-
pleteness of reporting (Higgins and Green, 2008; Jadad et al.,
1996). For observational studies, a quality checklist was used
(Tooth et al., 2005). A funnel plot screened for publication bias.
2.6. Synthesis of results
A pooled estimate of the standardised mean change in de-
pression score from baseline to follow-up was derived using a
random effects model tted with RevMan software (Cochrane
Collaboration, 2008). This approach uses the pooled standard de-
viation of baseline and follow-up scores as the denominator of the
standardized mean difference calculation, but ignores the corre-
lation between these scores. However it is likely baseline and
follow-up scores are in fact positively correlated, and ignoring this
correlation can bias the estimation of effect sizes (Dunlap et al.,
1996). Therefore a sensitivity analysis was performed to test the
effect of correlated scores on the pooled effect size estimate. A
plausible correlation gure of 0.4 was imputed, based on raw
clinical trial data from a study conducted in our research unit
(Adamson et al., 2015). The sensitivity analyses were performed
with SAS v9.2 PROC MIXED (SAS Institute, Cary NC, USA) em-
ploying a method for random effects meta-analysis outlined by
Sheu and Suzuki (2001). Variance estimates for each individual
study outcome were obtained by the following formula (Dunlap
et al., 1996):
Var (d) = [2 (1 r )/n] + [d2/(2n 2)] (1)
where d is the standardized mean change for each study, r is
the correlation between baseline and follow up scores within
subjects and n is the sample size.
Sensitivity analyses also investigated the inuence of individual
studies on the pooled effect size estimate.
Exploration of moderators of change in depression during
treatment was then performed as a meta-regression analysis using
PROC MIXED. Due to the small number of studies, moderator
analyses were limited to baseline HDRS score, depression type,
gender composition of the sample and manualisation of adjunctive
psychological treatment.
2.7. Antidepressant efcacy
For the studies that compared an antidepressant to placebo, a
pooled estimate of the treatment effect was derived using a ran-
dom effects model in RevMan. In this analysis, the mean difference
in change scores between antidepressant and placebo groups was
calculated. This was then standardised using the pooled standard
deviation of follow-up scores. If follow-up standard deviations
were not quoted, the standardised mean difference was derived
from the F statistic from ANCOVA models. This approach to ana-
lysing change scores followed guidelines outlined in the Cochrane
Handbook for Systematic Reviews of Interventions (Higgins and
Green, 2008, pages 269270) and it maintained consistency with a
previous meta-analysis in this population (Nunes and Levin, 2004).
A subgroup analysis then compared the pooled effect size between
ID and SID/mixed studies.
Sensitivity analyses investigated the inuence of individual
studies on the pooled effect size estimate.
53
2.8. Psychological treatments
A brief narrative review of psychological treatment studies was
performed. Due to the small number of randomised trials on
psychological treatments that met inclusion criteria, no attempt
was made to summarise results in a meta-analysis.
2.9. Predictors of treatment outcome
In addition to the meta-regression analyses described above, a
narrative review of predictors of treatment outcome at a within-
study level was performed. This section focused on gender and
alcohol use, as these predictors were most commonly studied and
reported.
3. Results
3.1. Search results and quality of studies
The search yielded 22 individual studies included in the sys-
tematic review. Table 1 presents a summary of these studies.
Twenty studies were randomized trials, of which 18 were
pharmacotherapy studies and two investigated non-pharmacolo-
gical treatments. The remaining two studies were open label
pharmacotherapy trials.
Of the 22 studies, 16 studies clearly presented data at baseline
and follow-up for the intention-to-treat sample. Follow-up com-
pletion rates in these studies ranged from 48% to 93% (median
70%). The remaining 6 studies (Altintoprak et al., 2008; Dorus
et al., 1989; Janiri et al., 1996; Mason et al., 1996; Paparrigopoulos
et al., 2010; Petrakis et al., 2007) were excluded from the meta-
analysis because their reported outcomes were not clearly iden-
tied as relating to the intention-to-treat sample.
The quality of randomized trials selected for the meta-analysis
sample was good as judged by the Jadad quality score (range 35,
median 4). For the two open-label studies, quality was acceptable
to good.
3.2. Dening and measuring depression in subjects with an alcohol
use disorder
Depression and alcohol use disorders were diagnosed with the
Structured Clinical Interview for DSM (SCID) (Spitzer, et al., 1992)
in 14 studies; the Psychiatric Research Interview for Substance and
Mental Disorders (PRISM) (Hasin et al., 1996) in two studies; the
Schedule for Clinical Assessment in Neuropsychiatry (SCAN) (Wing
et al., 1990) in one study; and the Diagnostic Interview Schedule
for DSM (Robins et al., 1981) in one study. The remaining studies
used unstructured or semi-structured interviews.
Sixteen studies, all pharmacotherapy trials, used versions of the
Hamilton Depression Rating Scale (HDRS) (Hamilton, 1960) as a
primary outcome. The standard 17-item version was used in 10/16
studies, the 24-item version in 3/16, the 21-item version in 2/16
and the remaining study did not state the version used. Three
studies used the Beck Depression Inventory (BDI), (Beck et al.,
1961). Two studies used the Montgomery-Asberg Depression
Rating Scale (MADRS), (Montgomery and Asberg, 1979) and the
remaining study used the QIDS-SR (Rush et al., 2003).
As shown in Table 1, there was variation between studies in the
timing of baseline depression measures and the inclusion or ex-
clusion of patients who responded during the initial assessment
phase of the study. Most studies had a mean baseline depression
score in the mild to moderate depression range (Zimmerman et al.,
2013); for example, among the studies using the 17-item HDRS,
the score at baseline ranged from 12.6 to 23.9 (median 19.6).
54 J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759

Operational denitions of depression types varied between

studies, and these are summarised in Table 1. For simplicity, the
terms independent and substance-induced are used in the present
review while studies on primary and secondary depression are
grouped with those on ID and SID respectively.
Eleven studies determined depression type at baseline. Of
these, ve studies included only subjects with ID (Cornelius et al.,
1997; McGrath et al., 1996; Moak, et al., 2003; Pettinati et al., 2010;
Roy-Byrne et al., 2000). Three studies (Adamson, et al., 2015;
Brown et al., 2011; Oslin, 2005) explicitly identied whether
subjects had ID or SID, and included both types. The remaining
three studies were on subjects with SID (Gual et al., 2003; Mason
et al., 1996; Paparrigopoulos et al., 2010).

3.3. Change in depression during treatment
A meta-analysis was initially planned on the 16 studies that
reported on change in depression from baseline to study com-
pletion for an intention-to-treat sample. However there was very
high heterogeneity in this sample (Q152.21, df 15, p o0.001;
I2 90%). To address the heterogeneity, two strategies were
adopted:
1. The meta-analysis was limited to the 11/16 studies that used
the HDRS; studies using other scales were excluded. The 11
studies were all pharmacotherapy trials: nine were randomised
placebo-controlled trials and the remaining two were open
label trials of a single medication.
2. The 11 studies were divided into those with a high or low
mean depression score at baseline according to whether the
baseline HDRS score was 17 or more. For studies using the 21 or
24-item HDRS, the threshold score was 18 or 22 respectively.
Note that the high baseline group corresponds to moderate to
severe depression and the low baseline group corresponds to
mild depression based on a standard severity classication
(Zimmerman, et al., 2013).
Fig. 3. Funnel plot for change in depression during treatment. Standardised mean
change in depression score from baseline to follow up is shown on the x-axis and
its standard error on the y-axis.
In the studies with high baseline depression severity (n 8),
there was little heterogeneity in change in depression during
treatment (Q7.84, df 7, p .28; I2 11%). High heterogeneity
(Q14.05, df 2, p o.001, I2 86%) persisted in the low baseline
depression severity studies (n 3).
Fig. 2 shows a forest plot of the change in depression from
baseline to study completion for each study. As shown in Fig. 2,
there was a consistently large reduction in depression during
treatment in the high baseline depression studies (standardised
mean change from baseline to study completion 1.92) and a
smaller and more variable reduction (standardised mean change
0.91) in the low baseline depression studies. These quoted effect
sizes may be underestimates, because some studies did not per-
form baseline measures until sometime after subjects stopped
drinking, and depressive symptoms typically improve rapidly early
in treatment (Baker et al., 2013).
As shown in Fig. 3, the funnel plot for high baseline depression
studies showed no strong evidence of asymmetry, providing no
indication of publication bias. There were only three studies with

Fig. 2. Forest plot of standardised mean change in depression score from baseline to study completion, for studies with high baseline depression (Hamilton Depression
Rating Scale scoreZ 17) and low baseline depression (Hamilton Depression Rating Scale score o 17). For Kranzler et al. (2006), results for the two subgroups (early re-
sponders and non-responders) were aggregated.
 J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759 55

Fig. 4. Forest plot of antidepressant efcacy for studies on independent depression in comparison to studies on substance-induced or undifferentiated depression. The plot
shows the standardised mean difference of change in depression for antidepressant therapy compared to placebo. Results for group A and group B in Kranzler et al. (2006)
are reported separately, and the four groups in Pettinati et al. (2010) are aggregrated into two, according to whether subjects received sertraline or not.

low baseline depression therefore no funnel plot for this subgroup
is shown.
Sensitivity analyses were performed examining the inuence of
within-subject correlation over repeated measures, and the in-
uence of individual studies on the pooled effect size estimate.
Changing the estimate of within-subject correlation between
baseline and follow up scores to 0.4 produced a change in the
pooled effect size estimate of less than 0.01, and had little inu-
ence on its standard error. This suggests the estimate of the pooled
effect is likely to be reliable even if the variances of individual
study effects are mis-specied. Sequentially deleting each of the
high-baseline studies had minimal effect on the pooled effect size
estimate for this subgroup. However for the low baseline studies,
deletion of any of the studies had a large effect on the pooled ef-
fect size estimate, suggesting this gure has low reliability.
Nine studies reported depression scores at multiple follow-up
timepoints (Altintoprak et al., 2008; Brown et al., 2011; Cornelius
et al., 1997; Kranzler et al., 2006; Moak et al., 2003; Paparrigo-
poulos et al., 2010; Pettinati et al., 2010; Roy-Byrne et al., 2000;
Yoon et al., 2006). In all of these studies, most of the change in
depression occurred within the rst 36 weeks. Two studies
showed a small further improvement between weeks 6 and 12
(Pettinati et al., 2010; Yoon et al., 2006) while the remaining stu-
dies showed no further improvement. No studies showed sig-
nicant improvement in the mean level of depressive symptoms
beyond 12 weeks.
3.4. Moderators of depression outcome
Moderators of change in depression score from baseline to
follow up were investigated using a linear mixed model approach
for meta-regression as outlined in Section 2.6. Given the low
number of studies these analyses were under-powered and should
be regarded as exploratory.
As expected on the basis of the subgroup analysis conducted in
RevMan, higher raw baseline HDRS scores were signicantly as-
sociated with greater reduction in depression score during treat-
ment (b 0.010, SE 0.029, p 0.007).
Depression type was not a signicant predictor of change in
depression during treatment (b 0.04, SE 0.37, and p 0.91).
Among the studies with high baseline depression severity, the
standardized mean change for the four ID studies ranged from 1.63
to 2.24 and for the remaining four studies it ranged from 1.70 to
2.22. However none of this second group of studies was explicitly
in SID patients. In the one study on patients with SID included in
the meta-analysis (Gual et al., 2003), subjects had mild depression
at baseline (mean HDRS 13.4) and there was a mean reduction in
HDRS score of about 7 points during follow up, a standardized
mean change of 1.4.
For the remaining moderator variables there was no evidence
of an effect for percentage of male subjects (b0.012, SE 0.016,
p 0.47), abstinence at baseline (b0.38, SE 0.37, p 0.33) or
whether adjunctive psychological treatment was manualised
(b 0.18, SE0.38, p .64).
3.5. Antidepressant efcacy
A meta-analysis of antidepressant efcacy was conducted on
the nine placebo-controlled studies included in the meta-analysis
reported in Section 3.3, and one additional study (Mason et al.,
1996). The study by Mason et al. had been excluded from the
earlier analyses because it did not report intention-to-treat data
for change in depression, but it did report these data for anti-
depressant efcacy. Therefore in total there were ten placebo-
controlled antidepressant studies, investigating sertraline (ve
studies), nefazodone (two studies), imipramine (one study), desi-
pramine (one study) and uoxetine (one study). Five of these
studies were in patients with ID and the remaining ve studies
56 J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759
Fig. 5. Funnel plot for antidepressant studies, grouped according to depression
type. Standardised mean difference is shown on the x-axis and its standard error on
the y-axis.
were in subjects with SID (Gual et al., 2003; Mason et al., 1996) or
mixed/undifferentiated samples (the other three studies). Mean
baseline HDRS score severity (Zimmerman et al., 2013) was in the
mild range in three studies, moderate in ve studies and severe in
the remaining two studies.
As shown in Fig. 4, signicant heterogeneity in antidepressant
effect was present in the SID/undifferentiated depression group of
studies but not for the ID studies. The pooled effect size for anti-
depressant medication compared to placebo across all studies was
0.17 (
0.05, 0.39). When this was broken down into subgroups,
the pooled effect size was 0.25 (0.06, 0.44) in the ID studies and
0.08 (
0.31, 0.47) in the remaining studies. A funnel plot (Fig. 5)
did not present evidence of publication bias. In sensitivity analyses
exploring the effect of deleting individual studies, the revised ef-
fect size estimates ranged from 0.21 to 0.35 for ID studies and from

0.06 to 0.21 for the SID/undifferentiated studies, suggesting the
overall effect size estimates for each subgroup were relatively
robust.
3.6. Efcacy of psychological treatments
A recent systematic review of psychological treatments for al-
cohol misuse among people with depression or anxiety disorders
found some support for motivational interviewing and cognitive-
behavioural interventions as treatments for both substance use
and comorbid psychiatric symptoms (Baker et al., 2012). However
the authors of this review noted there was limited published re-
search available in this area.
The present review identied only two psychotherapy studies
that met inclusion criteria although the majority (15/20) of the
remaining studies provided psychosocial treatment as part of the
routine intervention for all patients.
A randomised trial compared cognitive behavioural treatment
vs a relaxation therapy control treatment for patients with alcohol
dependence and elevated depressive symptoms (Brown et al.,
2011). In this trial, depression was categorised as independent or
substance-induced at baseline. There was no overall difference in
depression outcome between the two treatment conditions during
12 months of follow up, with both treatment groups having a
substantial reduction in symptoms. Patients with current or past
ID had higher depression scores during follow up than patients
with SID. SID was noted to be an unstable diagnosis, with 24% of
SID subjects exhibiting major depression associated with at least
4-weeks abstinence during follow up, thereby meeting criteria for
ID.
Agyapong et al. (2013) studied the effect of supportive text
messages as part of the follow up for patients with depression and
a comorbid alcohol use disorder. In this randomised trial the text
message intervention appeared to have an effect on depression at
3 months but the 6-month follow up suggested there was no en-
during benet from the intervention after it nished.
3.7. Predictors of depression outcome: narrative review
Two studies (Adamson et al., 2015; Brown et al., 2011) directly
compared depression outcomes between ID and SID subjects.
Brown et al. (2011) reported ID was associated with signicantly
poorer depression outcomes than SID for both treatment arms.
However the study also demonstrated that about one in four
subjects initially diagnosed with SID had ongoing major depres-
sion during a 4-week abstinence during follow-up, thereby
meeting criteria for ID. Adamson et al. (2015) reported on a pla-
cebo-controlled study of citalopram in patients receiving open-
label naltrexone. In this study, subjects with SID had more im-
provement in both drinking and depression during follow up, with
an effect size of about 0.7 for the difference in depression out-
comes between patients with ID and those with SID at 24 weeks
(Foulds et al., 2015). This study also reported that personality traits
including self-directedness and harm avoidance predicted de-
pression outcome, while specic depression characteristics such as
melancholic features and number of past episodes did not.
In relation to gender, one 12-week study of sertraline compared
to placebo reported a signicant overall treatment effect for fe-
male but not male subjects (Moak et al., 2003). Six other studies
(Adamson et al., 2015; Brown et al., 2011; Conner and Srensen,
2005; Janiri et al., 1996; Kranzler et al., 2006; McGrath et al., 1996)
found no difference in depression outcome between men and
women.
Moak et al. (2003) used drinking in the previous week as a
time-dependant predictor of depression scores. In this study, less
drinking in the previous week was associated with lower de-
pression scores. In a small study of imipramine vs placebo,
McGrath et al. (1996) found depression response was associated
with better drinking outcomes, and among patients who improved
during treatment, 13/20 responders experienced improvement in
alcohol use prior depression, while depression response occurred
rst in 4/20. Improvement occurred simultaneously for the re-
maining subjects. In a study of uoxetine vs placebo in patients
referred from an inpatient psychiatric facility, change in drinking
was correlated with change in depression in patients receiving
placebo but not uoxetine (Cornelius et al., 1997). Oslin (2005)
reported that more frequent bouts of heavy drinking were asso-
ciated with less improvement in depression in a sample of older
adults. Conversely Roy-Byrne et al. (2000) reported that a reduc-
tion in drinking was not an independent predictor of depression
outcome.
In summary, data from this set of clinical trials do not provide
strong evidence on the causal relationship between alcohol use
and depression.
4. Discussion
4.1. Findings
In patients with an alcohol use disorder and depressive
symptoms, the mean level of depression improved considerably in
the rst 36 weeks of treatment, and it reached a plateau by about
3 months. The amount of improvement was consistent across
studies on subjects with high levels of depression at baseline, and
equated to a reduction in symptom score of about two standard
deviations from baseline to follow-up, or around 1015 points on
 J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759
the Hamilton Depression Rating Scale.
Studies on ID showed this condition improves rapidly with
follow up, even for patients assigned to a control group. There was
limited evidence on the clinical course of SID, possibly because
depressive symptoms in patients with SID are believed to be self-
limiting, therefore treatment strategies for this group have infre-
quently been studied.
Two individual studies were identied that directly compared
treatment outcomes between ID and SID groups (Adamson et al.,
2015; Brown et al., 2011). In both of these studies, patients with
SID had better depression outcomes, although in the study by
Adamson et al., this appeared to be related to SID patients redu-
cing their alcohol consumption more than ID patients. There was
also evidence that SID is not a stable diagnosis, with about one
quarter of patients initially labelled with SID meeting criteria for
ID within the next 12 months (Brown et al., 2011). No studies
demonstrated that patients with SID show more improvement in
depression than those with ID if they reduce drinking.
Across all antidepressant studies, the pooled effect size of an-
tidepressant therapy compared to placebo was estimated at less
than 0.2, with the condence interval for this estimate including
zero. Studies in subjects with ID had a pooled effect size of 0.25.
Among the remaining studies the effect size was close to zero, but
these studies were generally in subjects with lower depression
severity and they had more variable depression outcomes.
It is important to note that most of the studies included in the
meta-analysis sample were on subjects who were on average only
mildly to moderately severely depressed. However in patients
without an alcohol use disorder, the efcacy of antidepressants is
largely conned to those with severe to very severe depression
(Fournier et al., 2010). Therefore it is unsurprising that the effect of
antidepressant therapy was found to be at best modest in the
studies included in this review.
Other than baseline depression severity, change in drinking is
the most obvious candidate predictor of change in depression
among patients with an alcohol use disorder. Several studies re-
ported that greater reduction in drinking was associated with
better depression outcomes, but overall the evidence provided on
the causal relationship between alcohol use and depression from
this set of studies was not strong. The inability to draw strong
conclusions about the causal relationship between alcohol use and
mood symptoms from the clinical trial data in this review relates
to two problems with these data. Firstly, most of the change in
both drinking and mood occurred between baseline and the rst
recorded follow-up point, therefore the temporal ordering of
change in depression and drinking was not adequately captured in
most of these studies. Secondly, advanced statistical modelling
techniques to explore causal relationships in longitudinal data
generally require large samples (Tanaka, 1987), while only one of
the studies included in this review had more than 200 subjects. A
secondary analysis of data from Project MATCH, a large landmark
alcohol treatment study (Conner and Srensen, 2005), found evi-
dence for a bidirectional association in which depressive symp-
toms at baseline predicted early drinking intensity, and continued
drinking predicted subsequent depression. However the subjects
on Project MATCH had on average only mild depressive symptoms,
so it is not clear whether these results can be generalised to more
severely depressed samples.
4.2. Limitations
Firstly, as noted, there were relatively few studies on the
treatment of SID therefore it is difcult to make condent state-
ments about its natural history and treatment response from this
dataset. Secondly there was heterogeneity in the amount of
change in depression among studies with low levels of depression
57
at baseline, suggesting ndings were unreliable for this group.
Thirdly, although studies in this review provided some form of
treatment for depression, most of the studies were conducted in
addiction treatment settings rather than in psychiatric or primary
care clinics. As such the outcomes reported may not be gen-
eralisable to these other populations. However we note that the
ndings from the STAR*D study, a large open-label trial conducted
in psychiatric and medical settings found that patients with al-
cohol comorbidity had similar depression outcomes to those
without comorbidity (Davis et al., 2010).
4.3. Comparison with existing literature
It has been suggested that the early improvement in depressive
symptoms after entering treatment is mainly conned to patients
with SID (Schuckit et al., 1997), in contrast to our ndings sug-
gesting patients with ID also improve rapidly even if they are as-
signed to a control group or placebo. Our ndings are consistent
with a recent randomised trial in hazardous drinkers with de-
pressive symptoms (Baker et al., 2013). In this study, substantial
improvements in both drinking and depressive symptoms oc-
curred during the initial screening and assessment phase of the
study. Furthermore early improvement in depression occurred ir-
respective of whether depression or alcohol was believed to be the
primary problem.
The relatively good depression outcomes seen in patients who
were assigned to placebo or a control group suggest the standard
care provided to these groups may in itself have been an effective
treatment for depression. Therefore there may be value in future
studies elucidating which components of standard care are most
useful in relation to depression outcomes.
A previous meta-analysis of antidepressant therapy for de-
pression associated with substance use disorders, including alco-
hol, found a pooled effect size of 0.38; studies that diagnosed
depression after at least one week of abstinence and those with a
lower placebo response showed greater antidepressant effect
(Nunes and Levin, 2004). Our ndings for studies on ID are broadly
consistent with the ndings of Nunes and Levin (2004), albeit with
a slightly smaller estimated effect size. However our ndings in
groups of patients with undifferentiated depression or SID were
less favourable.
4.4. Typology of depression in alcohol use disorders: a false
dichotomy?
Given what is known about the causal links between alcohol
use disorders and depression (Kendler et al., 1993), a dichotomous
classication into ID vs SID is almost certainly an over-
simplication. The typology also relies on patients accurately re-
membering the timing of heavy drinking, abstinence and depres-
sion in the past. This information is unlikely to be reliable given
that past episodes of mental disorder are in general inaccurately
recalled (Moftt et al., 2010) and heavy drinking is likely to worsen
recall. Classifying a patient's depressive episode as independent or
substance-induced based on a whole-life timeline also relies on
the unlikely premise that the causal relationship between drinking
and depression is static over time for any individual. Although the
construct appears to have some value, it requires renement and a
broader search for predictors of depression outcome needs to be
undertaken. A better classication system would help to inform
prognosis and would help to allocate treatments such as anti-
depressants only to patients who are likely to benet.
4.5. Clinical implications and directions for future research
The challenge in treating patients with an alcohol use disorder
58 J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759
who are depressed is not primarily related to short-term man-
agement, since studies consistently report rapid improvement
largely irrespective of patient characteristics and the treatment
provided. Rather, the challenge is to understand why recovery
quickly reaches a plateau, and to develop strategies to pro-
spectively identify, and effectively treat, patients at risk of ongoing
depression after the initial phase of treatment. Closer examination
of the predictors of long term depression outcome may be in-
formative in this regard. For example there is scarce literature on
personality variables despite good rationale to believe these fac-
tors may be relevant to depression treatment outcomes (Mulder,
2002).
The role of antidepressant therapy in depressed patients with
an alcohol use disorder is still not well dened, and the ndings
presented in this review are based on fewer than one thousand
patients in total. Studies that directly compare the effect of anti-
depressant therapy between patients with ID and SID would be
informative. Furthermore, with an estimated effect size less than
0.3, almost all studies on this population to date were under-
powered to detect an effect of this size. Future studies on this
population are therefore likely to need larger samples. Alter-
natively, and perhaps preferably, studies should concentrate on
recruiting patients with severe depression which persists after
detoxication.
4.6. Conclusion
In patients with an alcohol use disorder, symptoms of depres-
sion show substantial early improvement with all forms of treat-
ment, even if depression is believed to be independent of the al-
cohol use disorder.
The evidence for antidepressant therapy in depressed patients
with an alcohol use disorder points towards a small effect in pa-
tients with independent depression, and little or no effect in other
populations.
Role of funding source
Not applicable.
Conict of interest
The authors declare no conict of interest.
Acknowledgements
The authors wish to thank Sean Zloch for assistance with gure editing.
References
Adamson, S.J., Sellman, J.D., Foulds, J.A., Frampton, C.M., Deering, D., Dunn, A., Berks,
J., Nixon, L., Cape, G., 2015. A randomized trial of combined citalopram and
naltrexone for nonabstinent outpatients with co-occurring alcohol dependence
and major depression. J. Clin. Psychopharmacol. 35 (2), 143149.
Agyapong, V.I., McLoughlin, D.M., Farren, C.K., 2013. Six-months outcomes of a
randomised trial of supportive text messaging for depression and comorbid
alcohol use disorder. J. Affect. Disord. 151 (1), 100104.
Altintoprak, A.E., Zorlu, N., Coskunol, H., Akdeniz, F., Kitapcioglu, G., 2008. Effec-
tiveness and tolerability of mirtazapine and amitriptyline in alcoholic patients
with co-morbid depressive disorder: a randomized, double-blind study. Hum.
Psychopharmacol.: Clin. Exp. 23 (4), 313319.
Baker, A.L., 2012. Psychological interventions for alcohol misuse among people with
co-occurring depression or anxiety disorders: a systematic review. J. Affect.
Disord. 139 (3), 217229. http://dx.doi.org/10.1016/j.jad.2011.08.004.
Baker, A.L., Kavanagh, D.J., Kay-Lambkin, F.J., Hunt, S.A., Lewin, T.J., Carr, V.J.,
McElduff, P., 2014. Randomized controlled trial of MICBT for co-existing alcohol
misuse and depression: outcomes to 36-months. J. Subst. Abuse Treat. 46 (3),
281290.
Baker, A.L., Kay-Lambkin, F.J., Gilligan, C., Kavanagh, D.J., Baker, F., Lewin, T.J., 2013.
When does change begin following screening and brief intervention among
depressed problem drinkers? J. Subst. Abuse Treat. 44 (3), 264270.
Baker, A.L., Thornton, L.K., Hiles, S., Hides, L., Lubman, D.I., 2012. Psychological in-
terventions for alcohol misuse among people with co-occurring depression or
anxiety disorders: a systematic review. J. Affect. Disord. 139 (3), 217229.
Beck, A.T., Ward, C., Mendelson, M., 1961. Beck depression inventory (BDI). Arch.
Gen. Psychiatry 4 (6), 561571.
Brown, R.A., Ramsey, S.E., Kahler, C.W., Palm, K.M., Monti, P.M., Abrams, D., Du-
breuil, M., Gordon, A., Miller, I.W., 2011. A randomized controlled trial of cog-
nitive-behavioral treatment for depression versus relaxation training for alco-
hol-dependent individuals with elevated depressive symptoms. J. Stud. Alcohol
Drugs 72 (2), 286.
Brown, S.A., 1988. Changes in depression among abstinent alcoholics. J. Stud. Al-
cohol Drugs Suppl., 49 (05), 412.
Cochrane Collaboration, 2008. Review Manager (RevMan). 5.0. Copenhagen. The
Nordic Cochrane Centre: The Cochrane Collaboration.
Conner, K.R., Srensen, S., 2005. Initial depression and subsequent drinking during
alcoholism treatment. J. Stud. Alcohol Drugs 66 (3), 401.
Cornelius, J.R., Salloum, I.M., Ehler, J.G., Jarrett, P.J., Cornelius, M.D., Perel, J.M.,
Thase, M.E., Black, A., 1997. Fluoxetine in depressed alcoholics: a double-blind,
placebo-controlled trial. Arch. Gen. Psychiatry 54 (8), 700.
Davidson, K.M., 1995. Diagnosis of depression in alcohol dependence: changes in
prevalence with drinking status. Br. J. Psychiatry 166 (2), 199204. http://dx.
doi.org/10.1192/bjp.166.2.199.
Davis, L.L., Wisniewski, S.R., Howland, R.H., Trivedi, M.H., Husain, M.M., Fava, M.,
McGrath, P.J., Balasubramani, G.K., Warden, D., Rush, A.J., 2010. Does comorbid
substance use disorder impair recovery from major depression with SSRI
treatment? An analysis of the STAR*D level one treatment outcomes. Drug
Alcohol Depend. 107 (23), 161170.
Dorus, W., Ostrow, D.G., Anton, R., Cushman, P., Collins, J.F., Schaefer, M., Charles, H.,
Desai, P., Hayashida, M., Malkerneker, U., 1989. Lithium treatment of depressed
and nondepressed alcoholics. JAMA: J. Am. Med. Assoc. 262 (12), 16461652.
Dunlap, W.P., Cortina, J.M., Vaslow, J.B., Burke, M.J., 1996. Meta-analysis of experi-
ments with matched groups or repeated measures designs. Psychol. Methods 1
(2), 170.
Foulds, J.A., Sellman, J.D., Adamson, S.J., Boden, J.M., Mulder, R.T., Joyce, P.R., 2015.
Depression outcome in alcohol dependent patients: an evaluation of the role of
independent and substance-induced depression and other predictors. J. Affect.
Disord. 174, 503510.
Fournier, J.C., DeRubeis, R.J., Hollon, S.D., et al., 2010. Antidepressant drug effects
and depression severity: a patient-level meta-analysis. JAMA 303 (1), 4753.
http://dx.doi.org/10.1001/jama.2009.1943.
Gual, A., Balcells, M., Torres, M., Madrigal, M., Diez, T., Serrano, L., 2003. Sertraline
for the prevention of relapse in detoxicated alcohol dependent patients with a
comorbid depressive disorder: a randomized controlled trial. Alcohol Alcohol.
38 (6), 619625.
Hamilton, M., 1960. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry
23 (1), 56.
Hasin, D.S., Grant, B.F., 2002. Major depression in 6050 former drinkers: association
with past alcohol dependence. Arch. Gen. Psychiatry 59 (9), 794800. http://dx.
doi.org/10.1001/archpsyc.59.9.794.
Hasin, D.S., Trautman, K.D., Miele, G.M., Samet, S., Smith, M., Endicott, J., 1996.
Psychiatric research interview for substance and mental dsorders (PRISM):
reliability for substance abusers. Am. J. Psychiatry 153 (9), 11951201.
Hernandez-Avila, C.A., Modesto-Lowe, V., Feinn, R., Kranzler, H.R., 2004. Nefazo-
done treatment of comorbid alcohol dependence and major depression. Alco-
hol.: Clin. Exp. Res. 28 (3), 433440.
Higgins, J.P., Green, S., 2008. Cochrane Handbook for Systematic Reviews of Inter-
ventions. 5. Wiley Online Library, Chichester, England, Hokoben, New Jersey.
Jadad, A.R., Moore, R.A., Carroll, D., Jenkinson, C., Reynolds, D.J.M., Gavaghan, D.J.,
McQuay, H.J., 1996. Assessing the quality of reports of randomized clinical
trials: is blinding necessary? Control. Clin. Trials 17 (1), 112.
Janiri, L., Gobbi, G., Mannelli, P., Pozzi, G., Serretti, A., Tempesta, E., 1996. Effects of
uoxetine at antidepressant doses on short-term outcome of detoxied alco-
holics. Int. Clin. Psychopharmacol. 11 (2), 109117.
Judd, L.L., Akiskal, H.S., Schettler, P.J., Endicott, J., Maser, J., Solomon, D.A., Leon, A.C.,
Rice, J.A., Keller, M.B., 2002. The long-term natural history of the weekly
symptomatic status of bipolar I disorder. Arch. Gen. Psychiatry 59 (6), 530537.
Kendler, K.S., Heath, A.C., Neale, M.C., Kessler, R.C., Eaves, L.J., 1993. Alcoholism and
major depression in women. A twin study of the causes of comorbidity. Arch.
Gen. Psychiatry 50 (9), 690698 (Research Support, U.S. Govt, P.H.S.).
Kiefer, F., Barocka, A., 1999. Secondary depression in weaned alcoholics: implica-
tions of Leschs typology of chronic alcoholism. Alcohol Alcohol. 34 (6),
916917. http://dx.doi.org/10.1093/alcalc/34.6.916.
Kranzler, H.R., Mueller, T., Cornelius, J., Pettinati, H.M., Moak, D., Martin, P.R., An-
thenelli, R., Brower, K.J., OMalley, S., Mason, B.J., Hasin, D., Keller, M., 2006.
Sertraline treatment of co-occurring alcohol dependence and major depression.
J. Clin. Psychopharmacol. 26 (1), 1320.
Lejoyeux, M., Lehert, P., 2011. Alcohol-use disorders and depression: results from
individual patient data meta-analysis of the acamprosate-controlled studies.
Alcohol Alcohol. 46 (1), 6167.
Mason, B.J., Kocsis, J.H., Ritvo, E.C., Cutler, R.B., 1996. A double-blind, placebo-
controlled trial of desipramine for primary alcohol dependence stratied on the
presence or absence of major depression. J. Am. Med. Assoc. 275, 761767.
McGrath, P.J., Nunes, E.V., Stewart, J.W., Goldman, D., Agosti, V., Ocepek-Welikson,
K., Quitkin, F.M., 1996. Imipramine treatment of alcoholics with primary de-
pression: a placebo-controlled clinical trial. Arch. Gen. Psychiatry 53 (3), 232.
Moak, D.H., Anton, R.F., Latham, P.K., Voronin, K.E., Waid, R.L., Durazo-Arvizu, R.,
 J.A. Foulds et al. / Journal of Affective Disorders 185 (2015) 4759
2003. Sertraline and cognitive behavioral therapy for depressed alcoholics:
results of a placebo-controlled trial. J. Clin. Psychopharmacol. 23 (6), 553562.
Moftt, T.E., Caspi, A., Taylor, A., Kokaua, J., Milne, B.J., Polanczyk, G., Poulton, R.,
2010. How common are common mental disorders? Evidence that lifetime
prevalence rates are doubled by prospective versus retrospective ascertain-
ment. Psychol. Med. 40 (06), 899909. http://dx.doi.org/10.1017/
S0033291709991036.
Montgomery, S.A., Asberg, M., 1979. A new depression scale designed to be sen-
sitive to change. Br. J. Psychiatry 134 (4), 382389.
Muhonen, L.H., Lonnqvist, J., Juva, K., Alho, H., 2008. Double-blind, randomised
comparison of memantine and escitalopram for the treatment of major de-
pressive disorder with comorbid alcohol dependence. J. Clin. Psychiatry 69 (3),
392399.
Mulder, R.T., 2002. Personality pathology and treatment outcome in major de-
pression: a review. Am. J. Psychiatry 159 (3), 359371. http://dx.doi.org/
10.1176/appi.ajp.159.3.359.
Nunes, E.V., Levin, F.R., 2004. Treatment of depression in patients with alcohol or
other drug dependence: a meta-analysis. J. Am. Med. Assoc. 291 (15),
18871896. http://dx.doi.org/10.1001/jama.291.15.1887.
Oslin, D.W., 2005. Treatment of late-life depression complicated by alcohol de-
pendence. Am. J. Geriatr. Psychiatry 13 (6), 491500.
Ostacher, M.J., 2007. Comorbid alcohol and substance abuse dependence in de-
pression: Impact on the outcome of antidepressant treatment. Psychiatr. Clin.
N. Am. 30 (1), 6976. http://dx.doi.org/10.1016/j.psc.2006.12.009.
Paparrigopoulos, T., Tzavellas, E., Karaiskos, D., Malitas, P., Liappas, I., 2010. An open
pilot study of tiagabine in alcohol dependence: tolerability and clinical effects. J.
Psychopharmacol. 24 (9), 13751380.
Petrakis, I., Ralevski, E., Nich, C., Levinson, C., Carroll, K., Poling, J., Rounsaville, B.,
2007. Naltrexone and disulram in patients with alcohol dependence and
current depression. J. Clin. Psychopharmacol. 27 (2), 160165.
Pettinati, H.M., 2004. Antidepressant treatment of co-occurring depression and
alcohol dependence. Biol. Psychiatry 56 (10), 785792.
Pettinati, H.M., 2013. Current status of co-occurring mood and substance use dis-
orders: a new therapeutic target. Am. J. Psychiatry 170 (1), 2330. http://dx.doi.
org/10.1176/appi.ajp.2012.12010112.
Pettinati, H.M., Oslin, D.W., Kampman, K.M., Dundon, W.D., Xie, H., Gallis, T.L.,
Dackis, C.A., OBrien, C.P., 2010. A double-blind, placebo-controlled trial com-
bining sertraline and naltrexone for treating co-occurring depression and al-
cohol dependence. Am. J. Psychiatry 167 (6), 668675. http://dx.doi.org/
10.1176/appi.ajp.2009.08060852.
Pettinati, H.M., Volpicelli, J.R., Luck, G., Kranzler, H.R., Rukstalis, M.R., Cnaan, A.,
2001. Double-Blind Clinical Trial of Sertraline Treatment for Alcohol Depen-
dence. J. Clin. Psychopharmacol. 21 (2), 143153.
Robins, L.N., Helzer, J.E., Croughan, J., Ratcliff, K.S., 1981. National Institute of Mental
Health diagnostic interview schedule: its history, characteristics, and validity.
Arch. Gen. Psychiatry 38 (4), 381.
Roy-Byrne, P.P., Pages, K.P., Russo, J.E., Jaffe, C., Blume, A.W., Kingsley, E., Cowley, D.
S., Ries, R.K., 2000. Nefazodone treatment of major depression in alcohol-de-
pendent patients: a double-blind, placebo-controlled trial. J. Clin.
59
Psychopharmacol. 20 (2), 129136.
Rush, A.J., Trivedi, M.H., Ibrahim, H.M., Carmody, T.J., Arnow, B., Klein, D.N., Mar-
kowitz, J.C., Ninan, P.T., Kornstein, S., Manber, R., 2003. The 16-Item Quick In-
ventory of Depressive Symptomatology (QIDS), clinician rating (QIDS-C), and
self-report (QIDS-SR): a psychometric evaluation in patients with chronic major
depression. Biol. Psychiatry 54 (5), 573583.
Salloum, I.M., Jones, Y.O., 2008. Efcacy of pharmacotherapy for comorbid major
depression and substance use disorders: A review. Curr. Psychiatry Rev. 4 (1),
1427.
Schuckit, M.A., 1985. The Clinical Implications of Primary Diagnostic Groups Among
Alcoholics. Arch. Gen. Psychiatry 42 (11), 1043. http://dx.doi.org/10.1001/
archpsyc.1985.01790340021003.
Schuckit, M.A., 1995. Alcoholism and affective disorder: clinical course of depres-
sive symptoms. Am. J. Psychiatry 152, 4552.
Schuckit, M.A., 2006. Comorbidity between substance use disorders and psychiatric
conditions. Addiction 101, 7688. http://dx.doi.org/10.1111/j.1360-
0443.2006.01592.x.
Schuckit, M.A., Tipp, J.E., Bergman, M., Reich, W., Hesselbrock, V.M., Smith, T.L.,
1997. Comparison of induced and independent major depressive disorders in
2,945 alcoholics. Am. J. Psychiatry 154 (7), 948957.
Sheu, C.-F., Suzuki, S., 2001. Meta-analysis using linear mixed models. Behav. Res.
Methods Instrum. Comput. 33 (2), 102107.
Spitzer, R.L., Williams, J.B.W., Gibbon, M., First, M.B., 1992. The Structured Clinical
Interview for DSM-III-R (SCID) I: history, rationale, and description. Arch. Gen.
Psychiatry 49 (8), 624629. http://dx.doi.org/10.1001/
archpsyc.1992.01820080032005.
Sullivan, L.E., Fiellin, D.A., OConnor, P.G., 2005. The prevalence and impact of al-
cohol problems in major depression: a systematic review. Am. J. Med. 118 (4),
330341.
Tanaka, J.S., 1987. How big is big enough? Sample size and goodness of t in
structural equation models with latent variables. Child Dev., 134146.
Tooth, L., Ware, R., Bain, C., Purdie, D.M., Dobson, A., 2005. Quality of reporting of
observational longitudinal research. Am. J. Epidemiol. 161 (3), 280288.
University of Pittsburgh, 2014. Inventory of Depressive Symptomatology (IDS) and
Quick Inventory of Depressive Symptomatology (QIDS). From http://www.ids-
qids.org/index2.html.
Wing, J.K., Babor, T., Brugha, T., Burke, J., Cooper, J.E., Giel, R., Jablenski, A., Regier, D.,
Sartorius, N., 1990. SCAN: Schedules for Clinical Assessment in Neu-
ropsychiatry. Arch. Gen. Psychiatry 47 (6), 589593.
Yoon, S.-J., Pae, C.-U., Kim, D.-J., Namkoong, K., Lee, E., Oh, D.-Y., Lee, Y.-S., Shin, D.-
H., Jeong, Y.-C., Kim, J.-H., 2006. Mirtazapine for patients with alcohol depen-
dence and comorbid depressive disorders: a multicentre, open label study.
Prog. Neuro-Psychopharmacol. Biol. Psychiatry 30 (7), 11961201.
Zimmerman, M., Martinez, J.H., Young, D., Chelminski, I., Dalrymple, K., 2013. Se-
verity classication on the Hamilton depression rating scale. J. Affect. Disord.
150 (2), 384388.