Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis

European Neuropsychopharmacology (]]]]) ], ]]]–]]]

www.elsevier.com/locate/euroneuro

REVIEW

Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis Cora Scheffta, Laura L. Kilarskib, Tom Bschorc, Stephan Köhlera,n a

Charité Universitätsmedizin Berlin, Campus Mitte, Department for Psychiatry and Psychotherapy, Charitéplatz 1, 10117 Berlin, Germany b University Hospital of Cologne, Department of Psychosomatic Medicine and Psychotherapy, Cologne, Germany c Schlosspark-Klinik, Department of Psychiatry, Berlin, Germany and Department of Psychiatry and Psychotherapy, Technical University Dresden, Dresden, Germany Received 14 March 2017; received in revised form 16 June 2017; accepted 5 July 2017

KEYWORDS

Abstract

Unipolar depression; Augmentation; Nutritional supplements; Efficacy; Systematic review

In this article, we aimed to assess the efficacy of adjunctive administration of nutritional supplements to antidepressants by means of a systematic review and meta-analysis. The supplements included were inositol, vitamin D, folic acid, vitamin B12, S-adenosyl-Lmethionine (SAMe), omega-3 polyunsaturated fatty acids (n-3 PUFA) and zinc. A structured database search (MEDLINE, EBSCO, CENTRAL, Web of Science) was performed using terms for the respective substances in conjunction with terms for depression and the mode of treatment (“add-on” OR "adjunctive" OR "augmentation"). Meta-analyses, randomized controlled trials (RCTs) and non-randomized comparative studies that investigated the supplements as an add-on in the treatment of clinically diagnosed MDD were included. Agents had to be added to an existing antidepressant regime (augmentation) or started simultaneously with the antidepressant (acceleration). For n-3 PUFAs, folic acid and zinc, new meta-analyses were performed as part of this work. Our meta-analyses of 10 articles on n-3 PUFAs and four on zinc support their efficacy. For folic acid, our meta-analysis does not support efficacy. For n-3 PUFAs, sensitivity analysis showed no difference between acceleration and augmentation designs, but significant differences between individuals with or without comorbidities. For the remaining substances, only a few RCTs were available. The preliminary data on inositol was negative, while one RCT for vitamin D demonstrated positive results. For vitamin B12 one and for SAMe

n

Corresponding author. Fax: +49 30450517944. E-mail address: [email protected] (S. Köhler).

http://dx.doi.org/10.1016/j.euroneuro.2017.07.004 0924-977X/& 2017 Elsevier B.V. and ECNP. All rights reserved.

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

2

C. Schefft et al. two RCTs and a few open trials are available reporting positive and mixed results. To summarize, for most of the substances, the available data is not yet sufficient or inconclusive. & 2017 Elsevier B.V. and ECNP. All rights reserved.

1. 1.1.

Introduction Background

In about 50% of patients the initiation of antidepressant treatment does not lead to a satisfactory response, and even after several treatment approaches, rates of nonremitted patients are still around 30% (Rush et al., 2006; Trivedi et al., 2006; Walsh et al., 2002). In case of a nonresponse to an antidepressant treatment, adding a drug from a different pharmaceutical class to the antidepressant has been shown to have an augmentative effect (Bschor, 2010; Bschor et al., 2014; Bschor et al., 2003; Crossley and Bauer, 2007; Köhler et al., 2013; Nelson and Papakostas, 2009). Most evidence has been demonstrated for augmentation with lithium (Bschor, 2014) and atypical antipsychotics (Nelson and Papakostas, 2009). However, augmenting with one of these drugs might also increase undesired side effects and thus make discontinuation of treatment more likely (Shine et al., 2015). It is therefore desirable to find new agents that augment or accelerate response to antidepressant therapy yet do not inflict an additional systemic burden on the patient. Over the past two decades, the influence of nutritional factors on mental health conditions received growing attention through epidemiological and experimental findings, coining the name nutritional psychiatry for this nascent field of research (Sarris et al., 2015a). In this review, we analyzed the literature on nutritional supplements as add-on agents for the treatment of major depressive disorder (MDD) that include essential nutrients and nutritional supplements. Our aim was to systematically review the evidence from original studies and collate data into meta-analyses where appropriate. Herein, we describe details and limitations of the respec-

Augmentation; delayed initiation of second agent

tive studies and report standardized effect sizes for each data set. Previous reviews and meta-analyses have summarized data relevant to this review (Almeida et al., 2015; Mocking et al., 2016; Sarris et al., 2016). However, none of these were designed to include data exclusively from patients fulfilling criteria for unipolar depression based on DSM or ICD-10 criteria in adjunctive trials. We furthermore distinguished between acceleration and augmentation designs, since these can address two different issues of antidepressant treatment: 1) time lapse between treatment initiation and response and 2) non-response to treatment, respectively. Contrary to previous analyses, we assessed trials for potential risks of bias and corrected accordingly where possible.

2. 2.1.

Experimental procedures Study inclusion criteria

We conducted a systematic review of randomized controlled trials (RCTs) and open-label trials investigating the efficacy of adjunctive nutritional supplements in classic antidepressant treatments. Therein, we defined augmentation as the addition of a supplement to an ongoing antidepressant monotherapy. The simultaneous initiation of a common antidepressant and a supplement was termed acceleration (Fig. 1). In RCTs the comparator had to be the antidepressant only or a placebo added to the antidepressant. The selected substances were classified as non-prescription nutritional supplements, which have previously been studied in clinical trials as add-on substances to antidepressants. The authors agreed to focus on the following substances for which at least one RCT was available and which were most commonly reported as adjunctive agents in the therapy of unipolar depression (Freeman et al., 2010; Rechenberg, 2015):

Acceleration; both agents commenced simultaneously

insufficient response

Antidepressant

Antidepressant

Augmenting agent

Fig. 1

Accelerating agent

Augmentation and acceleration trials.

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Nutritional supplement augmentation in unipolar depression – – – – – – – –

Inositol Vitamin D Metabolites of the one-carbon-cycle Folate Vitamin B12 S-adenosyl-L-methionine (SAMe) Omega-3 polyunsaturated fatty acid (n-3 PUFA) Zinc

The level of evidence according to the NICE guidelines (NICE, 2012), (see below) was indicated for each substance.

2.2.

Inclusion and exclusion criteria

All studies written in English language that described an augmentation or acceleration design of any duration and adding one of the supplements to an antidepressant were included. Only original data from both RCTs as well as open-label trials were included. Suitable open-label trials were included if the number of available RCTs matching our criteria was less than three. Samples with neurological or medical comorbidities were not excluded. Only studies reporting primary diagnoses of major depressive disorder (MDD) or major depressive episode according to DSM-III, -IV or ICD-10 criteria and their mode of ascertainment were considered. Psychometric instruments to measure the primary outcome of the intervention had to be the Hamilton Depression Rating Scale (HDRS, (Hamilton, 1960)), Beck Depression Inventory (BDI, (Beck et al., 1961)) or the Montgomery-Åsberg Depression-Rating Scale (MADRS, (Montgomery and Asberg, 1979)).

2.3.

Literature search

Several databases (MEDLINE, PsycINFO (EBSCO), Web of Science, Cochrane Central Register of Controlled Trials (CENTRAL), clinicaltrials.gov) were searched by one reviewer (CS). The search was conducted combining relevant MeSH terms such as "Fatty Acids, Omega-3", "Eicosapentaenoic Acid", "Fish Oils", "Docosahexaenoic Acids", "Inositol", "Folic Acid", "Folinic Acid", "S-Adenosyl methionine", "Vitamin B12", "Vitamin D", "Zinc", "Depressive Disorder" and “add-on”, "adjunctive" "augmentation" with the respective Boolean operators. Backward and forward citation search was conducted for relevant studies within published research or review articles. Search results were integrated into one reference database following removal of duplicates. Studies were scanned for relevance by title and abstract and a selection was made by one author (CS) for full text review according to inclusion and exclusion criteria and was double-checked by a co-author (SK).

3 available, data on serum levels or other physiological measures and their association to clinical outcomes were extracted. For the trials included in meta-analyses, risk of bias was assessed applying the tool proposed by the Cochrane Collaboration (Higgins and Green, 2011). 2.4.3. Statistics Statistical analyses were performed in R using the package meta (Schwarzer et al., 2015). Standardized mean differences were calculated as Hedge's g based on endpoint score differences between treatment and placebo arms (Cuijpers et al., 2016; Higgins and Green, 2011). Intention to treat data was used, if available. Standard deviations were imputed based on other trials in the review, if not otherwise imputable (Furukawa et al., 2006; Higgins and Green, 2011). Meta-analysis was performed if three or more data-sets were available. Random effects models were fitted given expected heterogeneity of trials due to inclusion criteria. Studies were weighted by the inverse variance method, between study variance was assessed by the DerSimonian-Laird estimator. We assessed heterogeneity by calculating the I2 statistic and the Q statistic, which is reported with its respective degrees of freedom (Q(df)) and its p-value. Subgroup analysis and meta-regression were performed if ten or more data-sets were available. A priori, we determined factors for subgroup analyses to be “comorbidity” and the design related factor “acceleration or augmentation”. Smallstudy effects were assessed using a funnel plot and rank correlation method for assessing asymmetry (Begg and Mazumdar, 1994). We report standardized mean differences (SMDs) with their respective 95% confidence intervals [lower bound; upper bound], response and remission rates, where available. Means are reported with standard deviations (SD), medians with their respective range (maximum minimum).

2.5.

In order to uniformly label the quality of evidence, we classified according to the NICE guidelines (NICE, 2012).

Ia Evidence from a meta-analysis of at least three RCTs Ib Evidence from at least one RCT or a meta-analysis of less than three RCTs

IIa Evidence from at least one qualitatively well-controlled study without randomization

IIb Evidence from at least one quasi-experimental, descriptive study III Evidence from non-experimental observational studies IV Evidence based on expert opinions or clinical authorities

3. 2.4.

Quality of evidence

Results

Data analysis 3.1.

2.4.1. Definitions Response was defined as a score reduction of 50% in the HDRS, BDI and MADRS (Riedel et al., 2010). Remission according to the HDRS was defined by a cut-off value r 7, or as individually defined in the original studies. Remission cut-off values for BDI and MADRS scores were reported as individually defined, if applicable. 2.4.2. Data extraction We extracted data on sample sizes, sample characteristics (age, gender, comorbidities), inclusion and exclusion criteria, use of antidepressants and other drugs, dose and substance of supplementation, trial duration, design, primary and secondary outcomes and the statistical significance of the results according to the original publication. We extracted baseline and endpoint scores of HDRS, BDI and MADRS as continuous variables for the estimation of effect sizes or other data to derive effect sizes (i.e., t-, F-values). If

Search results

Our search strategy yielded a total of 2046 results. After filtering and removing duplicates 88 full text articles were selected based on titles and abstracts. Of these, 60 were excluded for not meeting selection criteria, the selection process being depicted in Fig. 2. The chosen articles included 25 RCTs and 5 open-label trials, in 29 publications. Additionally, we reviewed 13 meta-analyses collating the data, which are discussed alongside our results. An overview of the extracted data, SMDs and grading of evidence levels is summarized in Table 1 and Table 2. 3.1.1. Inositol Inositol is an isomer of glucose. It has been hypothesized to play a role in the treatment of depression since it is part of the phosphatidylinositol second messenger mechanism, which is a common intracellular pathway for neurotransmitter systems such

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

4

C. Schefft et al.

Search of databases (n= 2046) Use of respective filters, where applicable - English language - Human subjects - Articles, Clinical Trials, RCTs, Meta analyses

n=1178 Removal of duplicate articles n=305 n=873 original articles

Searching database results for eligible publications by title and abstract

n=88 full text articles

Addition of one publication identified via backward search (Mozzafari-Khosravi et al. (2013))

n=30 original datasets from 29 articles included

Exclusion of 60 due to - No baseline data n=2 - Unclear diagnoses n=2 - No adjunctive design n=21 - Review, protocol or another article format n=8 - Different outcome measures n= 9 - No original data (secondary analysis) n=5 -

Meta-Analyses n=13, included in discussion

Fig. 2 Flowchart of study selection process.

as norepinephrine or serotonin (Nemets et al., 2001). Attempts to use inositol as a mono-therapeutic agent in depression, however, did not consistently show beneficial effects (Levine, 1995, 1997; Levine et al., 1993; Mukai et al., 2014) 3.1.1.1. Inositol as an adjunct to antidepressants. Two randomized, placebo-controlled, parallel trials that compared the addition of 12 g of inositol with the addition of a placebo to an SSRI (selective serotonin reuptake inhibitor) could be identified (Levine et al., 1999; Nemets et al., 1999), each with a duration of 4 weeks. Levine et al. (1999) also included some patients with medical comorbidities. In the accelerative RCT (Levine et al., 1999), 36 patients received treatment with 12 g of inositol or a placebo simultaneously to initiating SSRI treatment. The augmentative trial from the same group (Nemets et al., 1999) included 42 patients who had not or only partially responded to at least 3 weeks of SSRI treatment. In each study, treatment and placebo groups decreased their HDRS scores to the same extent, no significant differences were observed. The standard mean differences of the accelerative (Levine et al., 1999) and the augmentative trial (Nemets et al.,

1999) were 0.15 [ 0.6; 0.9] and 0.19 [ 0.46; 0.85], respectively. Response or remission rates were not reported. 3.1.2. Vitamin D Vitamin D (in its biologically active form of 1,25-dihydroxycholecalciferol) is essential for sufficient calcium and phosphorus uptake in order to maintain respective serum levels. Epidemiological data on a possible link between vitamin D serum levels and depression were reviewed in a meta-analysis by Anglin et al. (Anglin et al., 2013), showing a non-significantly increased odds of depression in patients with low versus high level vitamin D in the cross-sectional studies and a significantly increased hazard ratio in the cohort studies (2.21 [1.40;3.40]). RCTs supplementing vitamin D in clinically depressed patients as a monotherapy without any standard antidepressant have previously been reviewed (Shaffer et al., 2014) and have shown results of both efficacy and non-efficacy (MozaffariKhosravi et al., 2013; Yalamanchili and Gallagher, 2012). 3.1.2.1. Vitamin D as an adjunct to antidepressants. Two trials, one RCT and one open-label trial matched the inclusion criteria (Khoraminya et al., 2013; Zanetidou et al., 2011). The RCT

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Studies on nutritional supplement addition in unipolar depression.

Substance Authors (Year)

N. age: mean Adjunct Antidepressant Duration, (SD).% female substance AC or AU and comparator

Inositol

Levine et al. (1999)

36, 27 com- inositol pleted. 12 g/d, placebo 47.8 (17). 70.3%

fluvoxamine, fluoxetine, paroxetine

4 weeks, AC

Nemets et al. (1999)

42, 36 completed. 50.5 (11). 61%

fluvoxamine, fluoxetine, paroxetine

4 weeks, AU

inositol 12 g/d, placebo

Vitamin D Khorami42, 40 com- vitamin D fluoxetine nya et al. pleted. 1500 IU/d, (2013) 38.9 (9.6). placebo 85%

Zanetidou 39 et al. 74 (6.2). (2011) 68%

Folate derivatives

8 weeks, AC

Inclusion criteria

Outcome measures, (primary outcome bold faced)

DSM-IV diagnosis of HDRS, version not ns. MDD diagnosed in clin- specified SMD= 0.15 ical interview. 3-7 0.9] days drug free prior to study DSM-IV diagnosis of 24-item HDRS ns. MDD, previously at SMD= 0.19 least 3 weeks of [-0.46; 0.85] treatment, 24-item HDRS Z 18 DSM-IV diagnosis of MDD. 17-item HDRS score Z 15, no antidepressant treatment or nutrient supplementation for past 2 months prior to inclusion

24-item HDRS, BDI, 25-hydoxycholecalciferol serum levels

vitamin D any 4 weeks, DSM-IV diagnosis of HDRS 300,000 antidepressant AU MDD ascertained by IU, openclinical interview, 65 label. years and older Unclear treatment in comparator group

Coppen 127, 109 com- folic acid fluoxetine and Bailey pleted. 0.5 mg/d, (2000) 43.1 (13.4). placebo 63%

Results at endpoint, Type of trial SMD (95%CI)

10 weeks, DSM-III-R diagnosis of AC MDD, new episode in at least six months. ascertained by clinician. HDRS-17 Z 20, no medication for past 9 weeks at baseline

17-item HDRS, response and remission ( r9 points), plasma folate and homocysteine levels

Level of evidence

RCT, double-blind

Ib

RCT, double-blind

Ib

RCT, double-blind

Ib

[-0.6;

sign. SMD= -1.25 [-1.93; -0.58]. Depression severity and 25(OH)-D level: negative correlation at baseline r= -0.63 (po0.001) sign. SMD= -0.82 [-1.49; -0.15]

Nutritional supplement augmentation in unipolar depression

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Table 1

Non-randomized, IIa controlled, openlabel trial

ns. RCT, double-blind SMD= -0.36 [-0.74; 0.02] Score reduction and remission rates only significant for female folic acid treated

Ib

5

6

Substance Authors (Year)

Başoğlu et al. (2009)

Antidepressant Duration, N. age: mean Adjunct (SD).% female substance AC or AU and comparator

42, 35 completed. 30 (18-66). 54.3%

Papakostas 148, 119 et al. completed. (2012) (1) 47.9 (11.6). 69.5%

folic acid escitalopram 2.5 mg/d, escitalopram monotherapy as comparator

Inclusion criteria

6 weeks, DSM-IV MDD ascerAC tained by SCID. New episode in at least six months. no medication for past 9 weeks at baseline. No folate or vitamin B12 deficiency at baseline

l-methyl- SSRI 60 days folate. 1) (2x30 days placebo on fixed for 60 d, 2) dose), placebo 30 AU d, then 7.5 mg/d for 30 d, 3) 7.5 mg/ d for 30 d then 15 mg/d for 30 d Papakostas 75, 61 com- l-methyl- see Papakostas 60 days et al. pleted. folate. 1) et al. (2012 (1) (2x30 days (2012) (2) 48.4 (12.1). placebo on fixed 70.6% for 60 d, 2) dose), placebo 30 AU d, then 15 mg/d for 30 d, 3) 15 mg/d for 60 d

DSM-IV diagnosis of MDD, SSRI treatment for at least 8 weeks, for past 4 weeks on stable dose, unsatisfactory response at point of inclusion

Outcome measures, (primary outcome bold faced)

MADRS, CGI, folate, vitamin B12 and homocysteine serum levels

17-item HDRS, QIDS-SR, CGI severity and improvement scales

Results at endpoint, Type of trial SMD (95%CI)

Level of evidence

patients (64.7% vs. 48.9%) sign. Randomized, con- IIa favoring escitalopram trolled, open-label monotherapy over trial combination. Median scores : 12 (6-35) vs. 16 (1033), z = -2.48 (po0.05). Response 80%: escitalopram, 35% for escitalopram plus folic acid ns. SMD at 30 days = RCT, double-blind. Ib 0.32 [-0.06 to 0.7] Sequential parallel comparison. Multicenter

see Papakostas et al. see Papakostas sign. SMD at 30 days See Papakostas (2012) (1) et al. (2012) (1) = -0.54 [-1.07; et al. (2012) (1) -0.01]. Response rate higher in treatment group. ns: Remission rate

Ib

C. Schefft et al.

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Table 1 (continued )

27, 20 com- folic acid flluoxetine pleted. 10 mg/d, 20 mg/d 35 (13.7), 85% placebo

Alpert et al. (2002)

22, 16 completed. 45.2 (11). 59%

6 weeks, AC

folinic acid 15 mg/d for 2 weeks then 30 mg/d for 6 weeks, no placebo control Venkatasu- 42, 30 com- folic acid bramanian, pleted. 1.5 mg vs. Kumar, and 33.15 (6.4). 5 mg/day, Pandey 100% no placebo (2013) control

fluoxetine, ser- 8 weeks, traline, paroxe- AU tine, venlafaxine

Bedson et al. (2014)

various antide- 12 weeks pressants at AC and AU optimal dosing: 70% SSRI, 30% other

Vitamin Almeida B12, B6 et al. and (2014) folate

479, 440 com- folic acid pleted 5 mg, (included in placebo primary analysis). 45 (13). 64%

153, 128 completed. 62.6 (7.8). 56.2%

fluoxetine

vitamin citalopram B12 0.5 mg + vitamin B6 25 mg + folic acid 2 mg/ day in one capsule, placebo

6 weeks, AC

DSM-IV major depressive episode ascertained by SCID, free of antidepressant and vitamins 30 days prior DMS-IV diagnosis of MDD ascertained by SCID, non-response to SSRI or venlafaxine treatment after at least 4 weeks, 17item HDRS Z 12

17-item HDRS, sign. SMD= -0.72 plasma folate, (-1.5; 0.06) homocysteine

Female patients only, ICD-10 moderate to severe depressive episode, 17-item HDRS score Z 18, previous 8 weeks without treatment

17-item HDRS sign. SMD (5mg vs. (remission: r 8), 1.5mg) = -0.7 BDI (remission (-1.32; -0.07). r 9) Remission rate high vs. low dose: 36.8% vs. 8.7% (p =0.03) by HDRS criterion. 26.3% vs. 8.7% (p =0.13) by BDI criterion BDI-II, CGI, MADRS ns. and others. Serum SMD (BDI-II) at 12 folate, vitamin weeks = 0.04 [-0.15; B12 and 0.22] homocysteine

ICD-10 moderate to severe depression. BDI-II score Z 17 at baseline. No folate or B12 deficiency, 75% on continuous medication at sufficient dosage

52 weeks, DSM-IV-TR major AC depressive episode, age over 50 years, MADRS Z 20, no antidepressant at baseline

RCT, blinding unclear

17-item HDRS, CGI sign. 27% responded, non-controlled, 18% remitted after open-label trial folate addition. No correlation between rise of folate level in the treatment group and HDRS score

MINI to assess remission, MADRS scores, relapse rates after remission by 12 weeks, at 26 and 52 weeks of follow-up, homocysteine plasma concentration

Ib

III

Randomized, dou- IIa ble-blind, dosecomparison without trial placebo control

Nutritional supplement augmentation in unipolar depression

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Resler et al. (2008)

RCT, double-blind, Ib multicenter

ns. RCT, double-blind SMD at 12 weeks = 0.15 [-0.17; 0.48] (n at 12 weeks =146). sign. lower odds of relapse at follow-up for treatment group patients that were remitted after 12 weeks

Ib

7

8

Substance Authors (Year)

Antidepressant Duration, N. age: mean Adjunct (SD).% female substance AC or AU and comparator

Inclusion criteria

S-adeno- Berlanga sylet al. methio- (1992) nine (SAMe)

63, 40 after 1-week placebo run-in and at completion. 40.9 (11.4). 80%

DSM-III-R diagnosis of 17-item a major depressive response episode, 17-item remission HDRS score of Z 18

S-adenoAlpert sylet al. methio- (2004) nine (SAMe)

30, 23 completed. 48.4 (13). 73.3%

200 mg imipramine SAMe per daily intramuscular injection or placebo injection

SAMe 400 mg for 2 weeks, then 800 mg for 4 weeks, no placebo control Papakostas 73, 55 com- SAMe et al. pleted. 800 mg/d, (2010) na. escalation 60% to 1600 mg/d after 2 weeks in all but 2 patients, placebo

De Berardis 25, 24 comet al. pleted. (2013) 32 (5.1). 56.8%

SAMe 800 mg/d, no placebo control

2 weeks add-on treatment, 6 weeks follow-up, AC

fluoxetine, par- 6 weeks, DMS-IV diagnosis of oxetine, citalo- AU MDD, 17-item HDRS pram, escitaloscore Z 14. Stable pram, sertradose of antidepresline or sant for 4 weeks at venlafaxine baseline

Outcome measures, (primary outcome bold faced)

Results at endpoint, Type of trial SMD (95%CI)

HDRS, ns. RCT, double-blind and SMD= -0.73 [-1.37; -0.09]. Significant acceleration of treatment onset in SAMe treated group days 4-12. Loss of benefit thereafter 17-item HDRS, sign. mean 17-item Non-controlled, MADRS, BDI, CGI, HDRS score reduction open-label trial SQ, from 17.7 (4.2) to 10 physical measures: (6.6); Folate, homocys- 50% response and teine, vitamin B12 43.3% remission rate levels

Level of evidence

Ib

III

fluoxetine, par- 6 weeks, oxetine, citalo- AU pram, escitalopram, sertraline, duloxetine or venlafaxine

DMS-IV diagnosis of 17-item HDRS, CGI MDD ascertained by SCID, 17-item HDRS Z 16, SSRI nonresponders, stable adequate dose of antidepressant for 4 weeks at baseline

Ib

venlafaxine, escitalopram, sertraline, bupropion, duloxetine, agomelatine, mirtazapine

DMS-IV diagnosis of MDD, HDRS Z 16, previous non-response after 8 weeks of treatment. Stable dose for at least 6 weeks

III

8 weeks, AU

ns. RCT, double-blind SMD= -0.77 [-1.24; -0.29]. Significantly higher response and remission rates in the SAMe- treated group. SAMe: 46% response, 36% remission. Placebo: 17% response, 12% remission 21-item HDRS, sign. 61% HDRS score Non-controlled, CGI-I, SHAPS, SDS reduction over base- open-label trial line after 8 weeks. 60% response and 36% remission by HDRS criterion

C. Schefft et al.

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Table 1 (continued )

Ranjbar et al. (2013) Nowak et al. (2003)

Siwek et al. (2009)

44, 38 completed. 37.3 (8.5). 89.5% 20, 14 completed. 42.8 (11.4). 57%

zinc 25 mg/d, placebo

citalopram, fluoxetine

12 weeks, DMS-IV-TR diagnosis of BDI AU MDD, medicated at baseline.

zinc 25 mg/d, placebo

clomipramine, amitriptyline, citalopram, fluoxetine

6 weeks, AC

60, 54 com- zinc pleted. 25 mg/d, 46 (6). placebo 67%

imipramine

DMS-IV diagnosis of MDD. At least one week washout period from previous medication 12 weeks, DMS-IV diagnosis of AC MDD, moderate or severe episode. One week washout period prior to beginning of trial

sign. RCT, double-blind SMD= -0.71 [-1.38; -0.05]

Ib

17-item HDRS, BDI sign. RCT, double-blind SMD= -0.96 [-2.10; 0.1]. No significant effect on BDI 17-item HDRS, sign. RCT, double-blind BDI, CGI, MADRS. SMD for treatment Custom composite resistant patients = scores for response -1.12 [-2.1; -0.17]. and remission cri- ns. teria, combining SMD non-treatment HDRS/CGI, MADRS/ resistant = CGI, BDI/CGI, -0.23 [-0.93; 0.48] assessment of previous treatment resistance as mediating variable

Ib

Ib

AC, Acceleration design; AU, Augmentation design; BDI, Beck Depression Inventory; CGI, Clinical Global Impression Scale; CI, Confidence interval; d, day; DSM, Diagnostic and Statistical Manual of Mental Disorders; HDRS, Hamilton Depression Rating Scale; ICD-10, International Classification of Diseases, tenth revision; MADRS, Montgomery-Åsberg Depression Rating Scale; MDD, Major Depressive Disorder; MINI, Mini International Neuropsychiatric Rating Scale; N, Sample size; na, Not available; ns, not significant; QIDS-SR, Quick Inventory of Depressive Symptomatology - Self Rated; RCT, Randomized Controlled Trial; SAMe, S-Adenosyl-L-Methionine; SCID, Structured Clinical Interview for DSM Disorders; SDS, Sheehan Disability Scale; SHAPS, Snaith-Hamilton Pleasure Scale; sign., significant; SMD, Standardized Mean Difference; SQ, Systemizing Quotient; SSRI, Selective Serotonin Reuptake Inhibitor.

Nutritional supplement augmentation in unipolar depression

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Zinc

9

10

C. Schefft et al.

Fig. 3 Random effects meta-analysis of three RCTs using adjunctive folic acid treatment. From the report by Bedson et al. (2014) MADRS data was entered. Means representing mean endpoint scores of the respective rating scales. Estimates of between-study variance given as τ2.

employed an accelerative, the open-label trial an augmentative design. Vitamin D administration differed between a daily oral dose of 1500 IU for 8 weeks and a single oral dose of 300 000 IU at study entry and 4 weeks of follow-up (Zanetidou et al., 2011). Khoraminya et al., 2013 compared an adjunctive 1500 IU/d of vitamin D to adjunctive placebo during fluoxetine treatment (Khoraminya et al., 2013). The vitamin D supplemented group scoring significantly lower on HDRS and BDI compared to the fluoxetine and placebo arm at week 8 (SMD= 1.25 [ 1.93; 0.58]). At baseline, 95% of all patients presented with vitamin D deficiency and baseline serum levels of vitamin D were inversely correlated to BDI and HDRS scores (r= 0.63, po0.01). No data was reported on a differential benefit from the intervention based on baseline vitamin D levels or their subsequent increase. In the open-label augmentation trial (Zanetidou et al., 2011), patients received a single oral dose of 300 000 IU of cholecalciferol in addition to their antidepressant. Treatment of the control group is unclear from the report. The vitamin D group showed lower HDRS scores compared to controls after four weeks (SMD = 0.82 [ 1.49; 0,15]). Neither the vitamin D plus antidepressant (t (df=46)=0.72, p =0.47) nor the control group (t(df=28)=0.35, p=0.73, two-sided, within-group t-tests, α=0.05) changed their scores significantly over time. No significant differences were found in response and remission rates.

3.1.3. Molecules of the one-carbon-cycle 3.1.3.1. Folate. Folate, vitamin B12 and S-adenosyl-L-methionine are all involved in a biochemical methylation process that is referred to as the one-carbon-cycle (Bottiglieri, 1996; Coppen and Bolander-Gouaille, 2005). A meta-analysis of eleven cross-sectional epidemiological and case-control studies involving more than 15000 patients by Gilbody et al., (2007) yielded an odds ratio of 1.42 [1.10; 1.83] for patients with low serum folate status to be diagnosed with depression. In predicting response to antidepressant treatment, folate levels were largely inversely correlated to outcome (Alpert et al., 2003; Fava et al., 1997). For the oral supplementation of folate, not only folic acid but also Lmethylfolate and folinic acid are available.

3.1.3.1.1. Folic acid, methylfolate or folinic acid as an adjunct to antidepressants. Eight trials in seven publications were retrieved (Alpert et al., 2002; Başoğlu et al., 2009; Bedson et al., 2014; Coppen and Bailey, 2000; Papakostas et al., 2012; Resler et al., 2008; Venkatasubramanian et al., 2013). Five of them were RCTs, one compared against antidepressant monotherapy (Başoğlu et al., 2009), one was a randomized, dose-ranging study (Venkatasubramanian et al., 2013) and one an open-label trial (Alpert et al., 2002). Folic acid was administered in five trials, two trials used L-methylfolate and the openlabel trial used folinic acid. All samples used mainly SSRIs as the background antidepressant medication.

3.1.3.1.2. Trials using adjunctive folic acid or folinic acid. Of the five trials that investigated folic acid (Başoğlu et al., 2009; Bedson et al., 2014; Coppen and Bailey, 2000; Resler et al., 2008; Venkatasubramanian et al., 2013), three (Bedson et al., 2014; Coppen and Bailey, 2000; Resler et al., 2008) were included in a random effects meta-analysis (see below). Two studies on folic acid were excluded from the meta-analysis since one reported outcome data that did not fit a normal distribution (Başoğlu et al., 2009) and the other compared different dosage levels but did not include a placebo control (Venkatasubramanian et al., 2013). Başoğlu et al. (Başoğlu et al., 2009) reported lower endpoint MADRS scores in the escitalopram monotherapy group compared to the folic acid supplemented group after 6 weeks. Venkatasubramanian (Venkatasubramanian et al., 2013) reported the higher dose (5 mg/d folic acid) to yield larger score decreases on both outcome measures HDRS and BDI (SMD (HDRS)= 0.7, [ 1.33; 0.07]) and higher remission rates than the low dose (1.5 mg/d folic acid). This acceleration trial had a duration of six weeks, using fluoxetine as the background antidepressant. Folinic acid was tested in a prospective open-label augmentation trial in 22 patients non-responsive to treatment (minimum 4 weeks of SSRI or venlafaxine) (Alpert et al., 2002). Mean reduction of HDRS scores was 6.3 (5.6), which was significant in this sample with 27% of patients responding and 18% remitting. 3.1.3.1.3. Risk of bias. Risk of selection and reporting bias was low for all reports. Blinding of outcome assessment was not detailed in two trials (Coppen and Bailey, 2000; Resler et al., 2008), while Bedson et al. (2014) primarily used a self-reporting instrument and was therefore at low risk of detection bias. Risk of attrition bias was estimated to be high for Resler et al. (2008) due to high drop-out rates and low for the other two studies. 3.1.3.1.4. Meta-analysis. The pooled effect estimate of a random-effects meta-analysis for three published trials did not significantly differ from zero (SMD= 0.26 [ 0.67; 0.15], z = 1.24, p=0.21, Fig. 3). Effect sizes were heterogenous (Q(df=2) =6.16, p =0.05, I2 =68%). All trials followed an acceleration design except for one (Bedson et al., 2014), which included both patients starting treatment and patients in active treatment. Dosing levels were 0.5 mg (Coppen and Bailey, 2000), 5 mg (Bedson et al., 2014) and 10 mg/d (Resler et al., 2008). Given the small number of trials, no subgroup analyses were performed.

3.1.3.1.5. Trials using adjunctive L-methylfolate (LMTHF). In two sequential double-blind placebo-controlled trials in SSRI treatment-resistant patients (augmentation design, (Papakostas et al., 2012)) escalating or stable doses of L-MTHF or a placebo were added to the SSRI in three parallel conditions. The first trial showed no efficacy of 7.5 mg/d of L-MTHF over placebo but a trend towards higher response rates for a dose of 15 mg/d (SMD at 30 days = 0.32 [ 0.06; 0.7]). The second trial's conditions tested the trend of efficacy for 15 mg/d and yielded a response rate of 32.3% in the L-MTHF group

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Nutritional supplement augmentation in unipolar depression compared to 14.6% in the placebo condition as well as higher score reductions (SMD at 30 days= 0.54 [ 1.07; 0.01]). 3.1.3.2. Vitamin B12. As mentioned above, the enzymatic cofactor vitamin B12 also plays a role in the one-carbon-cycle. As far as psychiatric and especially depressive symptoms are concerned, low vitamin B12 levels have been linked to depression in some epidemiological trials (Penninx et al., 2000; Tiemeier et al., 2002). One RCT was found to have investigated its augmentative potential (Almeida et al., 2014).

3.1.3.2.1. Vitamin B12 as an adjunct to antidepressants. We found one trial matching our criteria. However, it compared a combination compound of vitamins B12, B6 and folic acid to placebo adjunctive to citalopram over a one year period (n=153) (Almeida et al., 2014). Here, the adjunctive treatment with vitamins did not have a significant influence on the absolute MADRS scores over the course of treatment and at follow-up (SMD at 12 weeks= 0.15 [ 0.17; 0.48], means and SD were derived from a meta-analysis by the same authors (Almeida et al., 2015)). However, patients that were being administered the supplement and had remitted by the third month of the trial were less likely to relapse during follow-up than those receiving placebo. The odds of remission at 52 weeks in the vitamin-treated group were significantly higher for patients whose plasma homocysteine levels at baseline were above a cut-off of 10.4 mmol/l. This mediating effect is in line with previous research linking high homocysteine levels – indicative of a low folate and vitamin B12 status – to depression (Almeida et al., 2008; Bottiglieri et al., 2000; Folstein et al., 2007; Tiemeier et al., 2002). However, the authors also report a switch of antidepressants in the vitamin-treated group in 10.3% of patients at 26 weeks and in 16.1% at 52 weeks. No explicit remark is made by the authors on whether this differentially affected the patients above or below the homocysteine cut-off. Thus, it cannot be ruled out that the treatment benefit at follow-up in the high homocysteine subgroup is attributable to a change in antidepressant during that period. 3.1.3.3. S-adenosyl-L-methionine (SAMe). S-adenosyl-L-methionine (SAMe) is also involved in the one-carbon cycle. SAMe has predominantly been studied as a mono-therapeutic agent in the treatment of depression. A meta-analysis of 27 RCTs available in an evidence report by the Agency for Healthcare Research and Quality published in 2003 found that SAMe treatment significantly reduced HDRS scores compared to placebo by around 5.6 points, but did not outperform conventional antidepressants in studies using those as comparisons (Hardy et al., 2002).

3.1.3.3.1. SAMe as an adjunctive to antidepressants. Here we report the results of two RCTs (Berlanga et al., 1992; Papakostas et al., 2010) and two open-label trials (Alpert et al., 2004; De Berardis et al., 2013). Doses were administered orally or intramuscularly (i.m.) (Berlanga et al., 1992) and ranged from 200 mg to 1600 mg. The older RCT is the only accelerative trial meeting inclusion criteria for this review (Berlanga et al., 1992). Over 30% of the patients of the initial sample were excluded after remitting during a one-week placebo run-in phase. The remaining 40 patients received daily i.m. injections of 200 mg of SAMe or a placebo injection for 14 days. Simultaneously, imipramine treatment was initiated. After 14 days of adjunctive injections, the difference in HDRS scores between SAMe and placebo treatment reflected a trend towards SAMe superiority in this sample (SMD= 0.73 [ 1.37; 0.09]). Response rates were 65% in the SAMe and 45% in the placebo arm. No differences in response rates or HDRS scores were observed between study arms during follow-up (8 weeks). Papakostas et al., (2010) published results from an RCT including 73 patients who had been on a stable dosage of an SSRI or SSNRI (selective serotonin-norepinephrine reuptake inhibitor) for at least four weeks at the time of inclusion. Doses of 800 mg to 1600 mg/d of adjunctive SAMe were compared to placebo in continued antidepressant treatment. In both groups, mean HDRS scores, measured every week, declined over time and showed a trend towards lower scores in the SAMe treated

11 group (SMD= 0.77 [ 1.24; 0.29]). Significantly higher response (46% vs. 17.6%) and remission rates (36% vs 11.7%) in the SAMe treated group vs. placebo were observed. 3.1.3.3.2. Results from open-label trials. Alpert et al. (2004) and De Berardis et al. (2013) both report high response and remission rates for treatment-resistant patients in open-label trials (for details see Table 1). 3.1.4. Omega-3 fatty acids Omega-3 polyunsaturated fatty acids (n-3 PUFA) are essential fatty acids for humans as they cannot be produced by the body itself in sufficient quantities but need to be obtained from the diet. There are essentially two forms: eicosapentaenoic acid (EPA), synthesized from alpha-linolenic acid, and docosahexaenoic acid (DHA), which is in turn derived from EPA. N-3 PUFAs are constituents of the eukaryotic cell membrane and participate in G-Protein associated cell signaling. A variety of mechanisms have been proposed to mediate a potential antidepressant effect, such as anti-inflammatory and anti-oxidative, neurogenic and neurotrophic properties as well as an increase in serotonin and dopamine receptor expression (Su et al., 2013). A large body of research has been published on n-3 PUFAs in the treatment of depressive symptoms and is subject to previous reviews and meta-analyses (Appleton et al., 2006; Appleton et al., 2010; Appleton et al., 2015; Bloch and Hannestad, 2012; Grosso et al., 2014; Lin et al., 2012; Lin and Su, 2007; Martins, 2009; Mocking et al., 2016; Sarris et al., 2016; Sublette et al., 2011). Effects were largely favoring n-3 PUFA supplementation over placebo except for one meta-analysis that did not support efficacy (Bloch and Hannestad, 2012). A subsequent methodological critique and reanalysis of this article by two different groups revised and amended the original outcome, supporting n-3 PUFA efficacy. One of the points criticized in Bloch and Hannestad's work was the inclusion of results from subjects with depressive symptoms without a formal diagnosis of MDD (Lin et al., 2012; Martins et al., 2012). It has repeatedly been shown that the overall benefit of n-3 PUFA supplementation applies for the treatment of clinically diagnosed depression but not subclinical depressed mood (Appleton et al., 2010; Grosso et al., 2014; Lin et al., 2012; Lin and Su, 2007). Another unanimous finding was that a higher content of EPA seems to be associated with a larger antidepressant effect. However, more recent meta-analyses (Grosso et al., 2014; Mocking et al., 2016) imply that the predominant driver of the overall effect is the absolute dose of EPA, and not, as conjectured in earlier studies (Martins 2009, Sublette, Ellis et al. 2011), a particular ratio between EPA and DHA. All meta-analyses show significant heterogeneity across studies. We aimed to replicate and update the previous meta-analyses by including only individuals with MDD according to DSM or ICD criteria, adding a new trial (Shinto et al., 2016) and performing relevant subgroup analysis to account for heterogeneity. Further details such as risk of bias analysis, and sensitivity analysis are displayed in the supplement. 3.1.4.1. n-3 PUFA as an adjunctive to antidepressants: Metaanalysis. Our search yielded 10 data sets from RCTs published between the years 2002 and 2016 (summary Table 2). Fig. 4 displays the results of a random effects meta-analysis including all ten. A total of 402 patients were included and a significant effect in favor of n-3 PUFA supplementation in antidepressant treatment of MDD was shown (SMD= 0.48 [ 0.84; 0.11], z = 2.57, p=0.01). Effect sizes were heterogeneous (Q(df= 9)=24.69, p=0.003, I2 =64%). Sensitivity analysis in supplementary data. 3.1.4.1.1. Sub-group analysis. In order to account for the significant heterogeneity between studies (Q(df=9)=24.69, p=0.003, I2 =64%), comorbidities and supplementation design were taken into account. The subgroups sampling from populations with and without comorbidities differed significantly in a random effects model (χ2(df=1) =5, p=0.04). The random effects estimate and its p-values improved in the group of non-comorbid patients (SMD = 0.7 [ 0.81; 0.09], z=

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

12

Studies of n-3 PUFA supplementation in the treatment of unipolar depression. Authors (Year)

N. age: mean (SD), % female

Adjunct Antidepressant substance and comparator

Duration, AC or AU

Inclusion criteria

MozaffariKhosravi et al. (2013)

81, 62 completed. 35.1 (9.45). 60.8%

EPA 1g/d TCA, buproprion, MAOIs, vs. DHA SSRIs. Combinations 1g/d, thereof placebo: coconut oil

12 weeks, DSM-IV mild to moderate AU depressive episode ascertained by SCID, BDI 10-28, 17-item HDRS 8-18, no change in AD regimen 4 weeks prior

Outcome Risk of bias measures, (primary bold faced) 17-item HDRS at 6 weeks and at 12 weeks. Response and remission. 21-item HDRS, RBC n-3 PUFA content

Su et al. 28, 22 (2003) completed. 38.4 (11.16). 82%

EPA 4.40 g/d + DHA 2.20 g/d, placebo: olive oil

fluvoxamine, trazodone, moclobemide, fluoxetine. One patient in each group without AD treatment

8 weeks, one-week placebo run-in phase, AU

Shinto et al. (2016) Comorbid MS by McDonald criteria

39, 31 completed. 51.3 (10.8). 92.45%

EPA 1.95 g/d + DHA 1.35 g/d, placebo: soybean oil

various AD (bupropion, duloxetine, SSRIs, venlafaxine, trazodone)

Gertsik et al. (2012)

45, 42 after placeborun-in, 40.5 (10.2). Gender ratio na

EPA citalopram 1.8 g/d + DHA 0.4 g/d, placebo: olive oil

3 months, DSM-IV diagnosis of MDD AU ascertained by SCID, BDI score 10-30, stable dose of AD 3 months prior to enrollment. MS medication stable for 6 months prior to enrollment. Exclusion: BDI or MADRS Z30 8 weeks, DSM-IV diagnosis of MDD 21-item one-week ascertained by SCID, 21- HDRS, BDI, placebo item HDRS score Z17, par- MADRS run-in ticipants with previous phase, AC non-response to citalopram excluded, no fluoxetine or MAOI during previous 2 months

DSM-IV diagnosis of MDD, 21-item HDRS Z18, stable medication or psychotherpy within 4 weeks before enrollment

Type of trial

baseline: level of mean evidence (SD)

Low risk of bias in RCT, dou- 15.7 Ib all domains ble-blind (2.16) (HDRS17)

Unclear risk of selection, performance and detection bias.High risk of attrition bias. Low risk of reporting bias Low risk of bias in MADRS, BDI,RBC all domains. n-3 PUFA Reporting accordcontent ing to CONSORT criteria. Availability of patient level data

RCT, dou- 22.3 Ib ble-blind (3.9) (HDRS21)

RCT, dou- 18.5 ble-blind (4.75) (MADRS)

Unclear risk of RCT, selection bias, blinding unclear risk of unclear performance bias. High risk of detection, attrition and reporting bias

Ib

25.3 Ib (4.4) (HDRS21)

C. Schefft et al.

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Table 2

122, 117 completed. 58.4 (9.0). 33.6%

EPA 930 sertralin mg/d + DHA 750 mg/d, placebo: corn oil

Park et al. (2015)

35, 24 completed. 41.5 (3.7). 77%

EPA 1.14 various AD, 60% SSRI g/d + DHA 0.6 g/d, placebo: safflower oil with oleic acid

Jazayeri et al. (2008)

60, 48 completed. 34.8 (7.3). 66%

fluoxetine E-EPA 1g/d vs. fluoxetine 20 mg/d vs. fluoxetine 20 mg/d + E-EPA 1 g/d, doubledummy placebo

Nemets et al. (2002)

20, 19 E-EPA SSRIs, mirtazapine, MAOI, comple- 2 g/d, one patient without ted. 53.2 placebo: treatment

10 weeks, DSM-IV diagnosis of MDD two weeks ascertained by SCID or BDI placebo + II Z16 included sertraline run-in phase, AC

BDI-II, 17item HDRS, omega 3 RBC levels, 24 h ECG, remission r 8 on BDI-II

Low risk of bias in sequence generation, unclear allocation concealment.Low risks of detection, attrition and reporting bias.High risk of un-blinded patients (performance bias) 12 weeks, screening by CES-D-K, 17-item Low risk of selecAU score Z25, DSM-IV diag- HDRS, CGI- tion, performance nosis confirmed by a I, CGI-S, and detection clinician, chronic or CES-D-K bias. High risk of treatment-resistant attrition bias due depression excluded, to incomplete Z 65y excluded, no outcome data and physical or medical high drop-out comorbidities rate. 8 weeks, DSM-IV diagnosis of MDD 17-item Unclear risk of AC ascertained by SCID, 17- HDRS selection bias (no item HDRS score Z15, no details on randomedication for previous mization given). 6 weeks High risk of performance bias: no concealment of fishy taste. Low risk of detection bias. High risk of attrition bias. Selective reporting: no precision of outcome reported 4 weeks, DSM-IV diagnosis of MDD, 24-item Unclear risk of AU 24-item HDRS Z 18, at HDRS selection bias. least 3 weeks at current Low risk of perfor therapeutic dose mance, detection,

RCT, doubleblind

20.2 Ib (5.43) (HDRS17)

RCT, dou- 20.9 Ib ble-blind (6.05) (HDRS17)

Nutritional supplement augmentation in unipolar depression

RCT, dou- 30.47 Ib ble (6.2) dummy, (HDRS17) doubleblind

RCT, dou- 23.15 Ib ble-blind (2.9) (HDRS24)

13

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Carney et al. (2009). Comorbid coronary heart disease

14

Authors (Year)

N. age: mean (SD), % female

(8.3). 85% Bot et al. 25, 23 (2011). compleComorted. bid dia- 54.1 betes (11.5). mellitus 52% type I and II Da Silva 31, 29 et al. comple(2008) ted. 64.4 (range=4978). 58%

Adjunct Antidepressant substance and comparator not specified EPA 1 g/ d, placebo: rapeseed oil + medium chain triglycerides EPA 720 mg/d + DHA 480 mg/d, placebo: mineral oil

Duration, AC or AU

Inclusion criteria

Outcome Risk of bias measures, (primary bold faced)

Type of trial

baseline: level of mean evidence (SD)

reporting and attrition bias Unclear risk of RCT, dou- 26.35 SSRI (majority), TCA, SNRI, 12 weeks, diabetes I or II, on AD for MADRS, selection bias. ble-blind (8.5) adequacy of dose not AU at least two moths, DSM-IV baseline (MADRS) mentioned diagnosis MDD ascertained EPA levels Low risk of perforby CID Interview in RBC, mance, detection and reporting HbA1c bias. High risk of attrition bias SSRIs, 12 weeks, Parkinson's disease (PD) MADRS, (mainly sertraline), TCA, AU patients, DSM-IV MDD ascer- CGI. BDI trazodone. PD tatined by SCID, no signs of medication: levo dementia, only patients dopa, prami entered who had been takpexol, amantadine, COMTing AD for at least a year OR inhibitors had been refusing, sxcluded if PD symptoms worsened over time

Low risk of selec- RCT, dou- 23.5 tion and reporting ble-blind (4.62) bias. High risk of (MADRS) performance and attrition bias. Unclear risk of detection bias

Ib

Ib

AC, Acceleration design; AD, Antidepressant; AU, Augmentation design; BDI, Beck Depression Inventory; CES-D-K, Center for Epidemiological Studies Depression Scale - Korean Version; CGI-I/-S, Clinical Global Impression Scale - Global Improvement/ -Severity Scale; CHD, Coronary heart disease; CI, Confidence interval; CIDI, Composite International Diagnostic Interview; COMT, Catechol-O-methyl transferase; d, day; DHA, Docosahexaenoic acid; DSM, Diagnostic and Statistical Manual of Mental Disorders; E-EPA/EPA, Ethyl-eicosapentaenoic acid/ eicosapentaenoic acid; ECG, Electrocardiography; HDRS, Hamilton Depression Rating Scale; ICD-10, International Classification of Diseases, tenth revision; MADRS, Montgomery-Åsberg Depression Rating Scale; MAOI, Monoamine oxidase inhibitor; MDD, Major Depressive Disorder; MINI, Mini International Neuropsychiatric Rating Scale; MS, Multiple sclerosis; N, Sample size; n-3 PUFA, omega-3 polyunsaturated fatty acids; na, Not available; ns, not significant; PD, Parkinson's Disease; RBC, Red blood cells; RCT, Randomized Controlled Trial; SCID, Structured Clinical Interview for DSM Disorders; sign, significant; SMD, Standardized mean difference; SNRI, Serotonin and norepinephrine reuptake inhibitor; SSRI, Selective Serotonin Reuptake Inhibitor; TCA, Tricyclic antidepressant.

C. Schefft et al.

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Table 2 (continued )

Nutritional supplement augmentation in unipolar depression 3.45, po0.001), while the subgroup of studies sampling from comorbid populations showed no overall treatment effect (see Supplementary data). Heterogeneity was reduced by the subgroup division and was no longer significant in either of the subgroups. Distinguishing between acceleration and augmentation designs did not show significant differences between groups (Fig. S3). 3.1.5. Zinc Zinc is part of many metalloenzymes regulating processes such as protein synthesis, cell division, alcohol degradation or T-cell function (Trumbo et al., 2001). In an activity-dependent manner, zinc is released into the synaptic cleft where it can bind to specific sites of postsynaptic ionotropic NMDA-receptors, but also AMPA- and GABA receptors (Frederickson and Bush, 2001). For further reviews on the role of zinc as a neuromodulating agent, we refer to (Frederickson and Bush, 2001; Harrison and Gibbons, 1994). In a meta-analysis conducted by Swardfager et al. (Swardfager et al., 2013), the authors found zinc levels to be lowered in depressed patients with a linear association between plasma concentration and depression severity. 3.1.5.1. Zinc as an adjunctive to antidepressants. Three RCTs on zinc as an adjunctive to antidepressants were retrieved (Nowak et al., 2003; Ranjbar et al., 2013; Siwek et al., 2009). Each used a daily oral dose of 25 mg of zinc and two trials adopted an accelerative design (Nowak et al., 2003; Siwek et al., 2009), while one was augmentative. Trial duration was 6–12 weeks. In one trial (Siwek et al., 2009), study arms were stratified in treatment resistant and non-resistant patients by Thase and Rush criteria (Thase and Rush, 1997). We therefore included all three trials (one trial with two subgroups) in our meta-analysis. Given the small number of trials, no subgroup analyses were performed. 3.1.5.1.1. Risk of bias. All trials were at high risk of attrition bias since drop-out rates were high and no intention to treat analyses were performed. Risk of reporting bias was high for two trials since precision measures were either difficult to infer graphically (Siwek et al., 2009) or missing completely (Ranjbar et al., 2013), whereas Nowak et al. (2003) provides raw data. Risk of bias related to randomization and blinding procedures is unclear for two trials (Nowak et al., 2003; Ranjbar et al., 2013) since details are not reported, but is considered low for Siwek et al. (2009). 3.1.5.1.2. Meta-analysis. Fig. 5 shows the results of a random effects meta-analysis. The pooled effect estimate of adjunctive supplementation of 25 mg/d of zinc in antidepressant treatment is statistically significant (SMD= 0.66 [ 1.06; 0.26],z= 3.22, po0.01). No heterogeneity was detected between studies (Q(df=3)=2.64, p=0.45, I2 =0%). Repeating the analysis with BDI data did not substantially alter the point estimate (data not shown).

4. 4.1.

Discussion Main results and outlook

This article tried to provide a systematic review of the evidence on adjunctive nutritional supplementation in antidepressant therapy as a means of augmentation or acceleration. We reviewed 30 articles. We found that for folic acid, efficacy is not supported at a Ia level of evidence. For zinc and n-3 PUFAs there is level Ia evidence of efficacy in patients without comorbidities. For the remaining substances, efficacy is still underexplored. Evidence from two RCTs with small sample sizes does not currently support a beneficial role of inositol as an adjunctive agent in antidepressant treatment despite biological plausibility. For vitamin D, one RCT and one open-label trials have been published so far that both show a benefit of vitamin D as an adjunctive agent. A correlation of baseline vitamin D serum levels and depression severity could also be shown (Khoraminya et al., 2013). The trials reviewed differ in the dose of vitamin D, frequency

15 of administration, mean age of participants, depression severity and trial duration. Zanetidou et al. (2011) did not include a placebo comparison and there may be a risk of selective reporting bias. Future trials need to further assess this association and probe the robustness of the beneficial effect. As vitamin D deficiency appears to be more prevalent in the psychiatric population (Anglin et al., 2013), a control of vitamin D status should be considered when a patient who is otherwise not being followed by a physician presents with depressive symptoms. The same holds for vitamin B12 and folate, where deficiencies have been shown to be associated with symptoms of depression in cross-sectional studies (Gilbody et al., 2007; Penninx et al., 2000; Tiemeier et al., 2002). To avoid harmful neurologic and hematologic sequelae, supplementation by any means is only indicated in case of a deficiency. However, findings on vitamin B12's efficacy in treating depression are still too few, thereby prohibiting empirically founded statements on its potential in augmenting standard antidepressant therapy. The trial reviewed here showed no effect on depression severity (Almeida et al., 2014). Contrary to vitamin B12, a large body of data is available from studies assessing the effect of folic acids (Taylor et al., 2004). Here, collation of data across three RCTs does not currently support augmentative efficacy with Ia levels of evidence. Data on Lmethylfolate and folinic acid are preliminary and will require replication in larger samples. For SAMe, the RCTs described differ in their route of administration (oral vs intramuscular), dosage of SAMe, background antidepressant treatment, duration of supplementation, baseline depression severity and sample size. However, the effect sizes from both RCTs consistently favor SAMe supplementation and the trials were sufficiently powered to detect the observed effect size of -0.7. Augmentative efficacy was also observed in the open-label trials (Alpert et al., 2004; De Berardis et al., 2013). A large, high-quality RCT is needed before SAMe can be established as a standard augmentation agent. Since, as mentioned above, folate is a precursor in SAMe synthesis and already a well-established supplement with various applications, it would be interesting to see if SAMe, which is the downstream metabolite active in the brain is superior to folic acid in the augmentation of antidepressants. N-3 PUFAs have been discussed as promising antidepressant supplements, both as single and add-on agents, in several reviews. The most recent ones conclude that addition to antidepressants and the total dose of EPA are both associated with better outcome (Mocking et al., 2016; Sarris et al., 2016). Other analyses fall short of statistical significance (Appleton et al., 2015). We sought to replicate these meta-analyses by narrowing the inclusion criteria to rigorous diagnoses of MDD in adjunctive treatment data sets only. In subgroup analyses we accounted for heterogeneity and our results support the efficacy of n-3 PUFA supplementation in antidepressant treatment of otherwise healthy MDD patients. The significance of the effect size was robust to several adjustments (trim-and-fill, adjusting for detection bias). However, patients who suffer from medical or neurological comorbidities do not seem to benefit from supplementation, which can be attributed to a number of factors. The evidence supporting n-3 PUFA addition is currently based on small and mostly unregistered trials. A large clinical trial is needed to assess its potential, before n-3 PUFA supplementation can be recommended for clinical practice. Given that zinc's neuromodulating action in NMDA-receptor mediated transmission is known, this is yet another supplement that should be put under higher scrutiny within a sufficiently powered double-blind RCT. Our meta-analysis of four data sets shows a significant treatment effect. One RCT showed benefits in treatment-resistant patients. Due to the small number of trials in this meta-analysis and homogeneity of effect sizes, no further subgroup analysis was performed. However, the result should be interpreted with caution because of low power of the analysis.

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

16

C. Schefft et al.

Fig. 4 Forest plot of a meta-analysis of RCTs on adjunct omega-3 polyunsaturated fatty acid supplementation in antidepressant treatment. The two independent treatment groups of Mozaffari-Khosravi et al. (2013) were combined in order to include the common placebo group only once. For Jazayeri (2008) the conditions EPA +fluoxetine and fluoxetine +placebo were included. From DaSilva et al. (2008) only the data from patients on antidepressant medication was included. Means representing mean endpoint scores of the respective rating scales. Estimates of between-study variance given as τ2.

Fig. 5 Forest plot of the meta-analysis on adjunctive zinc in antidepressant treatment. All estimates are based on HDRS outcome data, except for Ranjbar et al. (2014), which reports only BDI data. Results from Siwek et al. (2009) were entered separately for treatment-resistant (TRD) patients and non-treatment resistant subgroups. Means representing mean endpoint scores of the respective rating scales. Estimates of between-study variance given as τ2.

Additionally, there is high risk of attrition bias in all studies included and most standard deviations are based on imputations. It remains to be clarified, how oral zinc supplementation affects cerebral zinc homeostasis, since serum and brain zinc concentrations seem to be independent and cognitive effects of zinc deficiency might be mediated by other mechanisms (Takeda, 2000, 2011). In this review, we maintained a distinction between acceleration and augmentation trials. However, not all original reports explicitly focused on accelerated treatment responses (earlier than 4 weeks) or augmentation in non-responding patients. The latter was explicitly addressed in 6 trials where L-methylfolate (Papakostas et al., 2012), folinic acid (Alpert et al., 2002) and SAMe (Alpert et al., 2004; De Berardis et al., 2013; Papakostas, 2009) showed augmentation potential in the scenario of a previous non-response according to defined criteria. Shinto et al. (2016), however, found no augmentation effect of n-3 PUFAs in antidepressant therapy in a sample of patients with comorbid MS. Acceleration of treatment response was reported for SAMe (Berlanga et al., 1992). In this study however, it is unclear if scoring patients every other day during the first two weeks of treatment reveals clinically significant findings. Apart from this, acceleration was observed for n-3 PUFAs in a trial by Nemets et al. (2002), yet not by Gertsik et al. (2012) nor Carney et al. (2009).

4.2.

Limitations

Several limitations of this review should be taken into consideration: Our search method was limited to the abovementioned databases and the English language. The substances we focused on might not reflect all supplements studied in this context (e.g. magnesium). Countries in which the trials were conducted (USA, Israel, Iran, Poland, New Zealand, etc.) might differ in baseline supplies of the respective nutrients as a function of culinary traditions, distance to the sea, scarcity or abundance of certain supplies or socioeconomic factors. Length of study and dosage levels were not factored into our metaanalyses since the number of studies included was too low to perform numerous meta-regressions and subgroup analyses. A positive correlation of dosage level has previously been reported in several other metaanalyses. Tolerability issues were beyond the focus of this article. Limited conclusions can be drawn from the open-label trials we included. The potential contribution of placebo effect and bias due to un-blinded outcome assessment have to be considered when interpreting these results. However, we decided to include them in order to adequately reflect the current state of research on the respective substances where few RCTs were available. We did not strictly adhere to PRISMA to guide our process of review, however, in the conduction

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Nutritional supplement augmentation in unipolar depression and reporting of this review we comply with most of the criteria in the PRISMA checklist (Moher et al., 2009). We state the deviation from it in our abstract, introduction and did not publish a review protocol.

4.3.

Conclusion

Highest levels of evidence are available for folic acid, n-3 PUFAs and zinc. In the case of folic acid the evidence points towards no efficacy, whereas for n-3 PUFAs several meta-analyses have shown efficacy and therefore should be assessed in a large clinical trial. The latter also holds true for zinc. Inositol has been shown to be of no efficacy in two studies. For L-methylfolate, SAMe, vitamin B12 and vitamin D there is too little data to draw firm conclusions. In future RCTs on nutritive agents, it would be desirable to adjust sample sizes accordingly and to operationalize both treatmentresistance and states of nutrient deficiency more precisely. To guide future research in the field, the International Society for Nutritional Psychiatry Research published a consensus statement emphasizing the importance of methodological rigor in RCTs and elucidation of the biological pathways involved. “Clinical trials of nutraceuticals should include assessment of biomarkers in tandem with clinical outcomes” (Sarris et al., 2015b).

Role of funding source There was no financial support and no competing interest.

Conflicts of interest none

Acknowledgements PD Dr. Stephan Köhler is a participant in the Charité Clinical Scientist Program funded by the Charité-Universitätsmedizin Berlin and the Berlin Institute of Health.

Appendix A.

Supporting information

Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/ j.euroneuro.2017.07.004.

References Almeida, O.P., Ford, A.H., Flicker, L., 2015. Systematic review and meta-analysis of randomized placebo-controlled trials of folate and vitamin B12 for depression. Int. Psychogeriatrics 27, 727–737. Almeida, O.P., Ford, A.H., Hirani, V., Singh, V., vanBockxmeer, F.M., McCaul, K., Flicker, L., 2014. B vitamins to enhance treatment response to antidepressants in middle-aged and older adults: results from the B-VITAGE randomised, double-blind, placebocontrolled trial. Br. J. Psychiatry 205, 450–457. Almeida, O.P., McCaul, K., Hankey, G.J., Norman, P., Jamrozik, K., Flicker, L., 2008. Homocysteine and depression in later life. Arch. Gen. Psychiatry 65, 1286–1294. Alpert, J.E., Mischoulon, D., Rubenstein, G.E., Bottonari, K., Nierenberg, A.A., Fava, M., 2002. Folinic acid (Leucovorin) as an adjunctive treatment for SSRI-refractory depression. Ann. Clin. Psychiatry: Off. J. Am. Acad. Clin. Psychiatr. 14, 33–38.

17 Alpert, J.E., Papakostas, G., Mischoulon, D., Worthington 3rd, J.J., Petersen, T., Mahal, Y., Burns, A., Bottiglieri, T., Nierenberg, A. A., Fava, M., 2004. S-adenosyl-L-methionine (SAMe) as an adjunct for resistant major depressive disorder: an open trial following partial or nonresponse to selective serotonin reuptake inhibitors or venlafaxine. J. Clin. Psychopharmacol. 24, 661–664. Alpert, M., Silva, R.R., Pouget, E.R., 2003. Prediction of treatment response in geriatric depression from baseline folate level: interaction with an SSRI or a tricyclic antidepressant. J. Clin. Psychopharmacol. 23, 309–313. Anglin, R.E.S., Samaan, Z., Walter, S.D., McDonald, S.D., 2013. Vitamin D deficiency and depression in adults: systematic review and meta-analysis. Br. J. Psychiatry: J. Ment. Sci. 202, 100–107. Appleton, K.M., Hayward, R.C., Gunnell, D., Peters, T.J., Rogers, P. J., Kessler, D., Ness, A.R., 2006. Effects of n-3 long-chain polyunsaturated fatty acids on depressed mood: systematic review of published trials. Am. J. Clin. Nutr. 84, 1308–1316. Appleton, K.M., Rogers, P.J., Ness, A.R., 2010. Updated systematic review and meta-analysis of the effects of n-3 long-chain polyunsaturated fatty acids on depressed mood. Am. J. Clin. Nutr. 91, 757–770. Appleton, K.M., Sallis, H.M., Perry, R., Ness, A.R., Churchill, R., 2015. Omega-3 fatty acids for depression in adults. Cochrane Database Syst. Rev. (Cd004692). Başoğlu, C., Ateş, M.A., Algül, A., Ipçioğlu, O.M., Geçici, Ö., Yilmaz, O., Semiz, Ü.B., Ebrinç, S., Gülsün, M., Özcan, Ö., Kiliç, S., Çetin, M., 2009. Adjuvant folate with escitalopram treatment and homocystein, folate, vitamin B-12 levels in patients with major depressive disorder. Klin. Psikofarmakol. Bul./Bull. Clin. Psychopharmacol. 19, 135–142. Beck, A.T., Ward, C.H., Mendelson, M., Mock, J., Erbaugh, J., 1961. An inventory for measuring depression. Arch. Gen. Psychiatry 4, 561–571. Bedson, E., Bell, D., Carr, D., Carter, B., Hughes, D., Jorgensen, A., Lewis, H., Lloyd, K., McCaddon, A., Moat, S., Pink, J., Pirmohamed, M., Roberts, S., Russell, I., Sylvestre, Y., Tranter, R., Whitaker, R., Wilkinson, C., Williams, N., 2014. Folate Augmentation of Treatment–Evaluation for Depression (FolATED): randomised trial and economic evaluation. Health Technol. Assess. 18 (vii-viii), 1–159. Begg, C.B., Mazumdar, M., 1994. Operating characteristics of a rank correlation test for publication bias. Biometrics 50, 1088–1101. Berlanga, C., Ortega-Soto, H., Ontiveros, M., Senties, H., 1992. Efficacy of S-adenosyl-L-methionine in speeding the onset of action of imipramine. Psychiatry Res. 44, 257–262. Bloch, M.H., Hannestad, J., 2012. Omega-3 fatty acids for the treatment of depression: systematic review and meta-analysis. Mol. Psychiatry 17, 1272–1282. Bot, M., Pouwer, F., Assies, J., Jansen, E.H., Beekman, A.T., de Jonge, P., 2011. Supplementation with eicosapentaenoic omega3 fatty acid does not influence serum brain-derived neurotrophic factor in diabetes mellitus patients with major depression: a randomized controlled pilot study. Neuropsychobiology 63, 219–223. Bottiglieri, T., 1996. Folate, vitamin B12, and neuropsychiatric disorders. Nutr. Rev. 54, 382–390. Bottiglieri, T., Laundy, M., Crellin, R., Toone, B.K., Carney, M.W.P., Reynolds, E.H., 2000. Homocysteine, folate, methylation, and monoamine metabolism in depression. J. Neurol., Neurosurg. Psychiatry 69, 228–232. Bschor, T., 2010. Therapy-resistant depression. Exp. Rev. Neurother. 10, 77–86. Bschor, T., 2014. Lithium in the treatment of major depressive disorder. Drugs 74, 855–862. Bschor, T., Bauer, M., Adli, M., 2014. Chronic and treatment resistant depression: diagnosis and stepwise therapy. Dtsch. Arztebl. Int. 111, 766–775. Bschor, T., Lewitzka, U., Sasse, J., Adli, M., Köberle, U., Bauer, M., 2003. Lithium augmentation in treatment-resistant depression:

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

18 clinical evidence, serotonergic and endocrine mechanisms. Pharmacopsychiatry 36 (Suppl 3), S230–S234. Carney, R., Freedland, K., Rubin, E., Rich, M., Steinmeyer, B., Harris, W., 2009. Omega-3 augmentation of sertraline in treatment of depression in patients with coronary heart disease: a randomized controlled trial. Jama, 1651–1657 pp. Coppen, A., Bailey, J., 2000. Enhancement of the antidepressant action of fluoxetine by folic acid: a randomised, placebo controlled trial. J. Affect. Disord 60, 121–130. Coppen, A., Bolander-Gouaille, C., 2005. Treatment of depression: time to consider folic acid and vitamin B12. J. Psychopharmacol. 19, 59–65. Crossley, N.A., Bauer, M., 2007. Acceleration and augmentation of antidepressants with lithium for depressive disorders: two metaanalyses of randomized, placebo-controlled trials. J. Clin. Psychiatry 68, 935–940. Cuijpers, P., Weitz, E., Cristea, I.A., Twisk, J., 2016. Pre-post effect sizes should be avoided in meta-analyses. Epidemiol. Psychiatr. Sci., 1–5. da Silva, T.M., Munhoz, R.P., Alvarez, C., Naliwaiko, K., Kiss, A., Andreatini, R., Ferraz, A.C., 2008. Depression in Parkinson's disease: A double-blind, randomized, placebo-controlled pilot study of omega-3 fatty-acid supplementation. Journal of Affective Disorders 111, 351–359. De Berardis, D., Marini, S., Serroni, N., Rapini, G., Iasevoli, F., Valchera, A., Signorelli, M., Aguglia, E., Perna, G., Salone, A., DiIorio, G., Martinotti, G., Di Giannantonio, A., 2013. S-adenosyl-Lmethionine augmentation in patients with Stage II TreatmentResistant Major Depressive Disorder: an open label, fixed dose, single-blind study. Sci. World J.. Fava, M., Borus, J.S., Alpert, J.E., Nierenberg, A.A., Rosenbaum, J. F., Bottiglieri, T., 1997. Folate, vitamin B12, and homocysteine in major depressive disorder. Am. J. Psychiatry 154, 426–428. Folstein, M., Liu, T., Peter, I., Buel, J., Arsenault, L., Scott, T., Qiu, W.W., 2007. The homocysteine hypothesis of depression. Am. J. Psychiatry 164, 861–867. Frederickson, C.J., Bush, A.I., 2001. Synaptically released zinc: physiological functions and pathological effects. Biometals 14, 353–366. Freeman, M.P., Mischoulon, D., Tedeschini, E., Goodness, T., Cohen, L.S., Fava, M., Papakostas, G.I., 2010. Complementary and alternative medicine for Major Depressive Disorder: a metaanalysis of patient characteristics, placebo-response rates, and treatment outcomes relative to standard antidepressants. J. Clin. Psychiatry 71, 682–688. Furukawa, T.A., Barbui, C., Cipriani, A., Brambilla, P., Watanabe, N., 2006. Imputing missing standard deviations in meta-analyses can provide accurate results. J Clin. Epidemiol. 59, 7–10. Gertsik, L., Poland, R., Bresee, C., Rapaport, M., 2012. Omega-3 fatty acid augmentation of citalopram treatment for patients with major depressive disorder. Journal of clinical psychopharmacology, 61–64. Gilbody, S., Lightfoot, T., Sheldon, T., 2007. Is low folate a risk factor for depression? Ameta‐analysis and exploration of heterogeneity. J. Epidemiol. Community Health 61, 631–637. Grosso, G., Pajak, A., Marventano, S., Castellano, S., Galvano, F., Bucolo, C., Drago, F., Caraci, F., 2014. Role of omega-3 fatty acids in the treatment of depressive disorders: a comprehensive metaanalysis of randomized clinical trials. Plos One, 9. Hamilton, M., 1960. A rating scale for depression. J. Neurol., Neurosurg., Psychiatry 23, 56–62. Hardy, M., Coulter, I., Morton, S.C., Favreau, J., Venuturupalli, S., Chiappelli, F., Rossi, F., Orshansky, G., Jungvig, L.K., Roth, E.A., Suttorp, M.J., Shekelle, P., 2002. S-Adenosyl-L-Methionine for Treatment of Depression, Osteoarthritis, and Liver Disease. Agency Healthc. Res. Qual. (US).

C. Schefft et al. Harrison, N.L., Gibbons, S.J., 1994. Zn2+: an endogenous modulator of ligand- and voltage-gated ion channels. Neuropharmacology 33, 935–952. Higgins, J., Green, Se, 2011. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (updated March 2011). Cochrane Collab. Jazayeri, S., Tehrani-Doost, M., Keshavarz, S., Hosseini, M., Djazayery, A., Amini, H., Jalali, M., Peet, M., 2008. Comparison of therapeutic effects of omega-3 fatty acid eicosapentaenoic acid and fluoxetine, separately and in combination, in major depressive disorder. The Australian and New Zealand journal of psychiatry, 192–198. Khoraminya, N., Tehrani-Doost, M., Jazayeri, S., Hosseini, A., Djazayery, A., 2013. Therapeutic effects of vitamin D as adjunctive therapy to fluoxetine in patients with major depressive disorder. Aust. New Zealand J. Psychiatry 47, 271–275. Köhler, S., Unger, T., Hoffmann, S., Steinacher, B., Fydrich, T., Bschor, T., 2013. Comparing augmentation with nonantidepressants over sticking to antidepressants after treatment failure in depression: a naturalistic study. Pharmacopsychiatry 46, 69–76. Levine, j, 1995. Double-blind, controlled trial of inositol treatment of depression. Am. J. Psychiatry 152, 792–794. Levine, J., 1997. Controlled trials of inositol in psychiatry. Eur. Neuropsychopharmacol. 7, 147–155. Levine, J., Gonsalves, M., Babur, I., Stier, S., Elizur, A., Kofman, O., Belmaker, R.H., 1993. Inositol 6 g daily may be effective in depression but not in schizophrenia. Hum. Psychopharmacol. Clin. Exp. 8, 49–53. Levine, J., Mishori, A., Susnosky, M., Martin, M., Belmaker, R.H., 1999. Combination of inositol and serotonin reuptake inhibitors in the treatment of depression. Biol. Psychiatry 45, 270–273. Lin, P.Y., Mischoulon, D., Freeman, M.P., Matsuoka, Y., Hibbeln, J., Belmaker, R.H., Su, K.P., 2012. Are omega-3 fatty acids antidepressants or just mood-improving agents? The effect depends upon diagnosis, supplement preparation, and severity of depression. Mol. Psychiatry 17, 1161–1163 (author reply 1163-1167). Lin, P.Y., Su, K.P., 2007. A meta-analytic review of double-blind, placebo-controlled trials of antidepressant efficacy of omega-3 fatty acids. J. Clin. Psychiatry 68, 1056–1061. Martins, J.G., 2009. EPA but not DHA appears to be responsible for the efficacy of omega-3 long chain polyunsaturated fatty acid supplementation in depression: evidence from a meta-analysis of randomized controlled trials. J Am. Coll. Nutr. 28, 525–542. Martins, J.G., Bentsen, H., Puri, B.K., 2012. Eicosapentaenoic acid appears to be the key omega-3 fatty acid component associated with efficacy in major depressive disorder: a critique of Bloch and Hannestad and updated meta-analysis. Mol. Psychiatry 17, 1144–1149 (discussion 1163-1147). Mocking, R., Harmsen, I., Assies, J., Ruhe, H., Koeter, M., Schene, A., 2016. Meta-analysis and meta-regression of Omega-3 polyunsaturated fatty acid supplementation for major depressive disorder. Bipolar Disord. 18 (177-177). Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., Group, P., 2009. Preferred reporting items for systematic reviews and metaanalyses: the PRISMA statement. PLoS Med. 6, e1000097. Montgomery, S.A., Asberg, M., 1979. A new depression scale designed to be sensitive to change. Br. J. Psychiatry 134, 382–389. Mozaffari-Khosravi, H., Nabizade, L., Yassini-Ardakani, S.M., Hadinedoushan, H., Barzegar, K.M., 2013. The effect of 2 different single injections of high dose of vitamin D on improving the depression in depressed patients with vitamin D deficiency: a randomized clinical trial. J. Clin. Psychopharmacol. 33, 378–385. Mukai, T., Kishi, T., Matsuda, Y., Iwata, N., 2014. A meta-analysis of inositol for depression and anxiety disorders. Human. Psychopharmacol. 29, 55–63.

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

Nutritional supplement augmentation in unipolar depression Nelson, J.C., Papakostas, G.I., 2009. Atypical antipsychotic augmentation in major depressive disorder: a meta-analysis of placebocontrolled randomized trials. Am. J. Psychiatry 166, 980–991. Nemets, B., Fux, M., Levine, J., Belmaker, R.H., 2001. Combination of antidepressant drugs: the case of inositol. Hum. Psychopharmacol. Clin. Exp. 16, 37–43. Nemets, B., Mishory, A., Levine, J., Belmaker, R.H., 1999. Inositol addition does not improve depression in SSRI treatment failures. J. Neural Transmission 106, 795–798 (Vienna, Austria: 1996). Nemets, B., Stahl, Z., Belmaker, R., 2002. Addition of omega-3 fatty acid to maintenance medication treatment for recurrent unipolar depressive disorder. The American journal of psychiatry, 477–479. NICE, 2012. National Institute for Health and Clinical Excellence (November 2012) The Guidelines Manual. National Institutef or Health and Clinical Excellence, London (Available from) 〈www. nice.org.uk〉. Nowak, G., Siwek, M., Dudek, D., Zieba, A., Pilc, A., 2003. Effect of zinc supplementation on antidepressant therapy in unipolar depression: a preliminary placebo-controlled study. Pol. J. Pharmacol. 55, 1143–1147. Park, Y., Park, Y.S., Kim, S.H., Oh, D.H., Park, Y.C., 2015. Supplementation of n-3 Polyunsaturated Fatty Acids for Major Depressive Disorder: A Randomized, Double-Blind, 12-Week, Placebo-Controlled Trial in Korea. Annals of Nutrition and Metabolism 66, 141–148. Papakostas, G., Shelton, R., Zajecka, J., Etemad, B., Rickels, K., Clain, A., Baer, L., Dalton, E., Sacco, G., Schoenfeld, D., Pencina, M., Meisner, A., Bottiglieri, T., Nelson, E., Mischoulon, D., Alpert, J., Barbee, J., Zisook, S., Fava, M., 2012. Lmethylfolate as adjunctive therapy for SSRI-resistant major depression: results of two randomized, double-blind, parallelsequential trials. Am. J. Psychiatry, 1267–1274. Papakostas, G.I., 2009. Evidence for S-adenosyl-L-methionine (SAMe) for the treatment of major depressive disorder. J. Clin. Psychiatry 70, 18–22. Papakostas, G.I., Mischoulon, D., Shyu, I., Alpert, J.E., Fava, M., 2010. S-Adenosyl Methionine (SAMe) augmentation of serotonin reuptake inhibitors for antidepressant nonresponders with major depressive disorder: a double-blind, randomized clinical trial. Am. J. Psychiatry 167, 942–948. Penninx, B.W., Guralnik, J.M., Ferrucci, L., Fried, L.P., Allen, R.H., Stabler, S.P., 2000. Vitamin B(12) deficiency and depression in physically disabled older women: epidemiologic evidence from the Women's Health and Aging Study. Am. J. Psychiatry 157, 715–721. Ranjbar, E., Kasaei, M.S., Mohammad-Shirazi, M., Nasrollahzadeh, J., Rashidkhani, B., Shams, J., Mostafavi, S.-A., Mohammadi, M.R., 2013. Effects of zinc supplementation in patients with major depression: a randomized clinical trial. Iran. J. Psychiatry 8, 73–79. Ranjbar, E., Shams, J., Sabetkasaei, M., M-Shirazi, M., Rashidkhani, B., Mostafavi, A., Bornak, E., Nasrollahzadeh, J., 2014. Effects of zinc supplementation on efficacy of antidepressant therapy, inflammatory cytokines, and brain-derived neurotrophic factor in patients with major depression. Nutritional neuroscience, 65–71. Rechenberg, K., 2015. Nutritional interventions in clinical depression. Clin. Psychol. Sci (2167702614566815). Resler, G., Lavie, R., Campos, J., Mata, S., Urbina, M., García, A., Apitz, R., Lima, L., 2008. Effect of folic acid combined with fluoxetine in patients with major depression on plasma homocysteine and vitamin B12, and serotonin levels in lymphocytes. Neuroimmunomodulation, 145–152. Riedel, M., Moller, H.J., Obermeier, M., Schennach-Wolff, R., Bauer, M., Adli, M., Kronmuller, K., Nickel, T., Brieger, P., Laux, G., Bender, W., Heuser, I., Zeiler, J., Gaebel, W., Seemuller, F.,

19 2010. Response and remission criteria in major depression–a validation of current practice. J. Psychiatr. Res. 44, 1063–1068. Rush, A.J., Trivedi, M.H., Wisniewski, S.R., Nierenberg, A.A., Stewart, J.W., Warden, D., Niederehe, G., Thase, M.E., Lavori, P.W., Lebowitz, B.D., McGrath, P.J., Rosenbaum, J.F., Sackeim, H.A., Kupfer, D.J., Luther, J., Fava, M., 2006. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am. J. Psychiatry 163, 1905–1917. Sarris, J., Logan, A.C., Akbaraly, T.N., Amminger, G.P., BalanzaMartinez, V., Freeman, M.P., Hibbeln, J., Matsuoka, Y., Mischoulon, D., Mizoue, T., Nanri, A., Nishi, D., Ramsey, D., Rucklidge, J.J., Sanchez-Villegas, A., Scholey, A., Su, K.P., Jacka, F.N., International Society for Nutritional Psychiatry, R., 2015a. Nutritional medicine as mainstream in psychiatry. Lancet Psychiatry 2, 271–274. Sarris, J., Logan, A.C., Akbaraly, T.N., Paul Amminger, G., BalanzaMartinez, V., Freeman, M.P., Hibbeln, J., Matsuoka, Y., Mischoulon, D., Mizoue, T., Nanri, A., Nishi, D., Parletta, N., Ramsey, D., Rucklidge, J.J., Sanchez-Villegas, A., Scholey, A., Su, K.P., Jacka, F.N., 2015b. International Society for Nutritional Psychiatry Research consensus position statement: nutritional medicine in modern psychiatry. World Psychiatry 14, 370–371. Sarris, J., Murphy, J., Mischoulon, D., Papakostas, G.I., Fava, M., Berk, M., Ng, C.H., 2016. Adjunctive nutraceuticals for depression: a systematic review and meta-analyses. Am. J. Psychiatry 173, 575–587. Schwarzer, G., Carpenter, J.R., Rücker, G., 2015. Meta-Analysis with R. Springer International Publishing. Shaffer, J.A., Edmondson, D., Wasson, L.T., Falzon, L., Homma, K., Ezeokoli, N., Li, P., Davidson, K.W., 2014. Vitamin D supplementation for depressive symptoms: a systematic review and meta-analysis of randomized controlled trials. Psychosom. Med. 76, 190–196. Shine, B., McKnight, R.F., Leaver, L., Geddes, J.R., 2015. Long-term effects of lithium on renal, thyroid, and parathyroid function: a retrospective analysis of laboratory data. Lancet 386, 461–468. Shinto, L., Marracci, G., Mohr, D.C., Bumgarner, L., Murchison, C., Senders, A., Bourdette, D., 2016. Omega-3 fatty acids for depression in multiple sclerosis: a randomized pilot study. Plos One, 11. Siwek, M., Dudek, D., Paul, I.A., Sowa-Kućma, M., Zięba, A., Popik, P., Pilc, A., Nowak, G., 2009. Zinc supplementation augments efficacy of imipramine in treatment resistant patients: a double blind, placebo-controlled study. J. Affect. Disord. 118, 187–195. Su, K.-P., Wang, S.-M., Pae, C.-U., 2013. Omega-3 polyunsaturated fatty acids for major depressive disorder. Exp. Opin. Investig. Drugs 22, 1519–1534. Sublette, M.E., Ellis, S.P., Geant, A.L., Mann, J.J., 2011. Metaanalysis of the effects of Eicosapentaenoic Acid (EPA) in clinical trials in depression. J. Clin. Psychiatry 72, 1577–1584. Swardfager, W., Herrmann, N., Mazereeuw, G., Goldberger, K., Harimoto, T., Lanctôt, K.L., 2013. Zinc in depression: a metaanalysis. Biol. Psychiatry 74, 872–878. Takeda, A., 2000. Movement of zinc and its functional significance in the brain. Brain Res. Rev. 34, 137–148. Takeda, A., 2011. Zinc signaling in the hippocampus and its relation to pathogenesis of depression. Mol. Neurobiol. 44, 166–174. Taylor, M., Carney, S., Goodwin, G., Geddes, J., 2004. Folate for depressive disorders: systematic review and meta-analysis of randomized controlled trials. J. Psychopharmacol. (Oxf., Engl.), 251–256. Thase, M.E., Rush, A.J., 1997. When at first you don't succeed: sequential strategies for antidepressant nonresponders. J. Clin. Psychiatry 58 (Suppl 13), 23–29. Tiemeier, H., van Tuijl, H.R., Hofman, A., Meijer, J., Kiliaan, A.J., Breteler, M.M., 2002w. Vitamin B12, folate, and homocysteine in

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004

20 depression: the Rotterdam study. Am. J. Psychiatry 159, 2099–2101. Trivedi, M.H., Rush, A.J., Wisniewski, S.R., Nierenberg, A.A., Warden, D., Ritz, L., Norquist, G., Howland, R.H., Lebowitz, B., McGrath, P. J., Shores-Wilson, K., Biggs, M.M., Balasubramani, G.K., Fava, M., Team, S.D.S., 2006. Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: implications for clinical practice. Am. J. Psychiatry 163, 28–40. Trumbo, P., Yates, A.A., Schlicker, S., Poos, M., 2001. Dietary reference intakes: vitamin A, vitamin K, arsenic, boron, chromium, copper, iodine, iron, manganese, molybdenum, nickel, silicon, vanadium, and zinc. J. Am. Diet. Assoc. 101, 294–301. Venkatasubramanian, R., Kumar, C.N., Pandey, R.S., 2013. A randomized double-blind comparison of fluoxetine

C. Schefft et al. augmentation by high and low dosage folic acid in patients with depressive episodes. J. Affect. Disord. 150, 644–648. Walsh, B., Seidman, S.N., Sysko, R., Gould, M., 2002. Placebo response in studies of major depression: variable, substantial, and growing. JAMA 287, 1840–1847. Yalamanchili, V., Gallagher, J.C., 2012. Treatment with hormone therapy and calcitriol did not affect depression in older postmenopausal women: no interaction with estrogen and vitamin D receptor genotype polymorphisms. Menopause 19, 697–703 (New York, N. Y. ). Zanetidou, S., Belvederi Murri, M., Buffa, A., Malavolta, N., Anzivino, F., Bertakis, K., 2011. Vitamin D supplements in geriatric major depression. Int. J. Geriatr. Psychiatry 26, 1209–1210.

Please cite this article as: Schefft, C., et al., Efficacy of adding nutritional supplements in unipolar depression: A systematic review and meta-analysis. European Neuropsychopharmacology (2017), http://dx.doi.org/10.1016/j.euroneuro.2017.07.004