Acceptability, Tolerability, and Estimates of Putative Treatment Effects of Probiotics as Adjunctive Treatment in Patients With Depression: A Randomized Clinical Trial

Key Points

Question  Are probiotics an acceptable, tolerable, and potentially efficacious adjunctive treatment for depression?

Findings  In this pilot randomized clinical trial including 49 participants, daily probiotic intake for 8 weeks resulted in greater improvements in depressive and anxiety symptoms compared with placebo.

Meaning  The acceptability, tolerability, and estimated effect sizes on key clinical outcomes are promising and encourage further investigation of probiotics as adjunctive treatment for people with major depressive disorder in a definitive efficacy trial.

Importance  The microbiota-gut-brain axis is a promising target for novel treatments for mood disorders, such as probiotics. However, few clinical trials have been conducted, and further safety and efficacy data are needed to support this treatment approach.

Objective  To provide acceptability and tolerability data and estimates of intervention effect size for probiotics as adjunctive treatment for patients with major depressive disorder (MDD).

Design, Setting, and Participants  In this single-center, double-blind, placebo-controlled pilot randomized clinical trial, adults aged 18 to 55 years with MDD taking antidepressant medication but having an incomplete response were studied. A random sample was recruited from primary and secondary care services and general advertising in London, United Kingdom. Data were collected between September 2019 and May 2022 and analyzed between July and September 2022.

Intervention  Multistrain probiotic (8 billion colony-forming units per day) or placebo daily for 8 weeks added to ongoing antidepressant medication.

Main Outcomes and Measures  The pilot outcomes of the trial were retention, acceptability, tolerability, and estimates of putative treatment effect on clinical symptoms (depression: Hamilton Depression Rating Scale [HAMD-17] and Inventory of Depressive Symptomatology [IDS] scores; anxiety: Hamilton Anxiety Rating Scale [HAMA] and General Anxiety Disorder [GAD-7] scores) to be used as indicators for a definitive trial.

Results  Of 50 included participants, 49 received the intervention and were included in intent-to-treat analyses; of these, 39 (80%) were female, and the mean (SD) age was 31.7 (9.8) years. A total of 24 were randomized to probiotic and 25 to placebo. Attrition was 8% (1 in the probiotic group and 3 in the placebo group), adherence was 97.2%, and there were no serious adverse reactions. For the probiotic group, mean (SD) HAMD-17 scores at weeks 4 and 8 were 11.00 (5.13) and 8.83 (4.28), respectively; IDS, 30.17 (11.98) and 25.04 (11.68); HAMA, 11.71 (5.86) and 8.17 (4.68); and GAD-7, 7.78 (4.12) and 7.63 (4.77). For the placebo group, mean (SD) HAMD-17 scores at weeks 4 and 8 were 14.04 (3.70) and 11.09 (3.22), respectively; IDS, 33.82 (9.26) and 29.64 (9.31); HAMA, 14.70 (5.47) and 10.95 (4.48); and GAD-7, 10.91 (5.32) and 9.48 (5.18). Standardized effect sizes (SES) from linear mixed models demonstrated that the probiotic group attained greater improvements in depressive symptoms according to HAMD-17 scores (week 4: SES, 0.70; 95% CI, 0.01-0.98) and IDS Self Report scores (week 8: SES, 0.64; 95% CI, 0.03-0.87) as well as greater improvements in anxiety symptoms according to HAMA scores (week 4: SES, 0.67; 95% CI, 0-0.95; week 8: SES, 0.79; 95% CI, 0.06-1.05), but not GAD-7 scores (week 4: SES, 0.57; 95% CI, −0.01 to 0.82; week 8: SES, 0.32; 95% CI, −0.19 to 0.65), compared with the placebo group.

Conclusions and Relevance  The acceptability, tolerability, and estimated effect sizes on key clinical outcomes are promising and encourage further investigation of probiotics as add-on treatment for people with MDD in a definitive efficacy trial.

Trial Registration  ClinicalTrials.gov Identifier: NCT03893162

Approximately 60% of people with major depressive disorder (MDD) experience some degree of nonresponse to first-line treatments, and approximately one-third continue to experience symptoms despite further treatment.¹ Increasing understanding of the involvement of the microbiota-gut-brain axis in the pathophysiology of MDD has made it a promising target for novel treatments, such as probiotics.

In a 2021 meta-analysis of 7 randomized clinical trials (RCTs) including 404 patients,² we found that probiotics appear to be effective in reducing depressive symptoms when administered adjunctively to antidepressants. However, several RCTs to date have not provided sufficient tolerability data, while others report poor adherence or retention.^2,3 Therefore, further safety and efficacy data are needed for probiotics to be considered a viable treatment option in clinical practice.

This study was part of a larger 8-week randomized double-blind placebo-controlled RCT. The aims of the main trial were mechanistic; however, the evaluation of the feasibility/pilot outcomes presented here was part of the protocol approved by the London-Surrey Research Ethics Committee. Written informed consent was obtained from all participants. The trial protocol can be found in Supplement 1. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.

A total of 50 outpatients with a primary diagnosis of MDD and with a Hamilton Depression Rating Scale (HAMD-17)⁴ score greater than 13 were recruited. All participants were taking an approved antidepressant at a stable dose for 6 or more weeks and were required to not make changes throughout the study. Exclusion criteria were bipolar disorder, psychosis, eating disorders, personality disorders, substance dependence, or suicidal ideation; serious medical illness, gastrointestinal disease or surgery; use of antibiotics or probiotics in the past 12 weeks; current or regular gastrointestinal medication use; smoking; pregnancy or breastfeeding; and a vegan diet. Race and ethnicity data were collected through self-report by choosing one of the following categories: Asian (non-Chinese; British or any other), Black (British or any other), Chinese, multiracial (any), White (British or any other), or other race (any).

Participants were randomized 1:1 to 4 capsules daily of probiotic (2 × 10⁹ colony-forming units per capsule) or matching placebo. The probiotic contained 14 strains of Bacillus subtilis, Bifidobacterium bifidum, Bifidobacterium breve, Bifidobacterium infantis, Bifidobacterium longum, Lactobacillus acidophilus, Lactobacillus delbrueckii subsp bulgaricus, Lactobacillus casei, Lactobacillus plantarum, Lactobacillus rhamnosus, Lactobacillus helveticus, Lactobacillus salivarius, Lactococcus lactis, and Streptococcus thermophilus (Bio-Kult Advanced; ADM Protexin) and was selected due to earlier evidence of antidepressant effects of these species.^2,5 Success of masking was evaluated by asking participants to guess their allocation at study end. For detailed intervention, randomization, and blinding procedures, see eMethods in Supplement 2.

Participants attended 3 visits (baseline, week 4, and week 8). The primary outcome for a future efficacy RCT was change in depressive scores at week 8 (HAMD-17 and Inventory of Depressive Symptomatology [IDS] Self Report).⁶ Other outcomes included changes in anxiety (Hamilton Anxiety Rating Scale [HAMA]⁷ and General Anxiety Disorder [GAD-7] scores)⁸ and clinical status (Clinical Global Impression [CGI])⁹ scores as well as adherence. Adverse events and gastrointestinal symptoms were also monitored (eMethods in Supplement 2).

We recruited 50 participants, consistent with recommendations for pilot studies aiming to perform power calculations.^10,11 Estimates of efficacy were calculated on the intent-to-treat (ITT) and per-protocol (PP) principles and measured the effect size of the between-group mean difference. Linear mixed models were performed with the outcomes as the dependent variables, treatment group, time, and time × group interaction as the fixed terms, and random intercept for participant. Standardized effect sizes (SES) with small sample size correction were calculated. For CGI Severity, as an ordinal outcome, generalized linear testing (ordinal logistic) was performed. CGI Improvement scores were analyzed with χ² likelihood ratio tests. Analyses were performed in SPSS version 28 (IBM), with 2-tailed significance level set at P < .05. The confounding effects of body mass index, age, weight, gastrointestinal complaints, alcohol intake, and dietary parameters were evaluated. Further details can be found in eMethods in Supplement 2.

Of 50 included participants, 49 received the intervention and were included in intent-to-treat analyses; of these, 39 (80%) were female, and the mean (SD) age was 31.7 (9.8) years. A total of 24 were randomized to probiotic and 25 to placebo. Three further participants in the placebo group dropped out, resulting in 46 completers and an attrition rate of 8%. Two participants commenced antibiotics and were included in the ITT analysis but not PP analysis (eFigure 1 in Supplement 2).

Participant characteristics are in Table 1. Baseline depression severity was moderate, and 45 of 49 participants (92%) were taking a selective serotonin-reuptake inhibitor (SSRI). Anxiety comorbidities were common, with 21 (43%) meeting criteria for generalized anxiety disorder. The only difference between groups was race, with all participants identifying as Asian (non-Chinese) allocated to probiotics (7 of 24 [29%] vs 0 of 25).

The masking was successful, with a nonsignificant correct guess rate between groups (8 [33%] in the probiotic group and 5 [23%] in the placebo group) and 22 participants (48%) selecting that they did not know (eTable 1 in Supplement 2). This rate is likely low due to the appropriate concealment and low adverse-effect profile of the intervention. Adherence was high, with 97.2% of doses taken as required (capsule count). The intervention was well tolerated, with no serious adverse reactions and no dropouts owing to adverse effects. A total of 16 participants reported adverse reactions (eTable 2 in Supplement 2). Of these, nausea and indigestion were experienced only in the probiotic group but were transient and did not require medication. Gastrointestinal symptom scores decreased in both groups over time and were not significant between groups (eFigure 2 in Supplement 2).

Depressive symptoms improved in both arms, with greater reductions in the probiotic group from week 4 (Figure). A strong association between treatment group and HAMD-17 scores was observed at week 4 (SES, 0.70; 95% CI, 0.01-0.98), IDS scores at week 8 (SES, 0.64; 95% CI, 0.03-0.87), and HAMA scores at both time points (week 4: SES, 0.67; 95% CI, 0-0.95; week 8: SES, 0.79; 95% CI, 0.06-1.05) but not GAD-7 scores (week 4: SES, 0.57; 95% CI, −0.01 to 0.82; week 8: SES, 0.32; 95% CI, −0.19 to 0.65) (Table 2). None of the covariates affected findings (data not shown), nor did sensitivity analyses evaluating the impact of non-SSRI medications or the clustering of Asian (non-Chinese) individuals in the probiotic arm (eTables 3 and 4 in Supplement 2). There were no notable differences between ITT and PP data sets (eTable 5 and eFigure 3 in Supplement 2). The probiotic group also showed stronger response on the CGI Improvement subscale (Figure) but not the CGI Severity subscale (week 4: odds ratio, 0.28; 95% CI, −0.85 to 1.42; week 8: odds ratio, 0.20; 95% CI, −0.99 to 1.39).

It has been suggested that probiotics may be beneficial as adjunctive treatment as they may help alleviate presentations that antidepressants are less effective against (eg, anxious, somatic).^2,5 To explore this, we analyzed the 9-item Anxiety/Arousal IDS subscale¹² using the PP data set at week 8 (eTable 6 in Supplement 2). We found significant effects, with a similar effect size to that of the main scale (SES, 0.75; 95% CI, 0.08-1.44), suggesting that the reduction in total scores may be driven by anxious and somatic symptoms.

To our knowledge, this is the first trial in a Western population to demonstrate the safety, acceptability, and therapeutic potential of a readily available and scalable probiotic intervention in patients with MDD. Compared with the placebo group, the probiotic group exhibited greater improvement in depressive symptoms with moderate effect sizes, which are comparable with those reported in earlier meta-analyses.^2,13 Participants in the probiotic arm experienced, on average, a reduction of 1 severity grade on both depression rating scales.

Anxiety symptoms have been little examined in probiotic trials in depression, despite their high prevalence in MDD (approximately 40% to 50%).¹⁴ In addition to the greater effects on clinician-rated anxiety, our exploratory analyses suggested that anxious-somatic symptoms may have been particularly improved by the probiotic. If confirmed in larger trials, these findings could provide an indication of which patients may benefit most from probiotic treatment.

The probiotic was well tolerated, with a low attrition rate, high adherence rate, and no serious adverse reactions. This safety and acceptability profile is better than those reported in earlier studies using different supplements.^2,3 Given the range of supplements available, clinical decisions should be guided not only by indicators of efficacy but also by safety and acceptability.

A limitation of this study is that we cannot ascertain whether the observed effects are specific to the interaction with SSRIs or generalizable to other treatments. Further, adherence was evaluated through capsule count, which, while the most commonly used method in clinical trials, can lead to overreporting. Nevertheless, this is the method used in most trials against which we compared adherence.

In summary, the preliminary findings from this pilot study suggest that 8-week adjunctive treatment with a multistrain probiotic is acceptable and tolerable for adults with MDD. The estimated effect sizes on key clinical outcomes are promising and encourage further investigation in a definitive efficacy trial.

Accepted for Publication: April 13, 2023.

Published Online: June 14, 2023. doi:10.1001/jamapsychiatry.2023.1817

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2023 Nikolova VL et al. JAMA Psychiatry.

Corresponding Author: Viktoriya L. Nikolova, PhD, Centre for Affective Disorders, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, De Crespigny Park, London SE5 8AF, United Kingdom (viktoriya.nikolova@kcl.ac.uk).

Author Contributions: Drs Nikolova and Stone had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: Nikolova, Cleare, Stone.

Drafting of the manuscript: Nikolova, Young.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Nikolova, Young.

Obtained funding: Nikolova, Cleare, Stone.

Administrative, technical, or material support: All authors.

Study supervision: Cleare, Young, Stone.

Conflict of Interest Disclosures: Dr Nikolova has received grants from the Medical Research Council and ADM Protexin during the conduct of the study as well as personal fees from Janssen outside the submitted work. Dr Cleare has received grants from the Medical Research Council and ADM Protexin during the conduct of the study as well as personal fees from Janssen, Compass Pathways, and Medscape outside the submitted work. Dr Young has received grants from the Medical Research Council and ADM Protexin during the conduct of the study; grants from Janssen, Lundbeck, and Compass Pathways; and personal fees from AstraZeneca, Boehringer Ingelheim, Eli Lilly, Sunovion, Allergan, Compass Pathways, Sage, Novartis, Neurocentrx, Sumitomo Dainippon Pharma, Bionomics, Livanova, Lundbeck, Servier, and Janssen outside the submitted work; and is Editor of the Journal of Psychopharmacology and Deputy Editor of BJPsych Open. Dr Stone has received grants from the Medical Research Council and ADM Protexin during the conduct of the study; research support from Takeda; charitable funding from Freya Green; and personal fees from Janssen outside the submitted work. No other disclosures were reported.

Funding/Support: This research was funded by a Medical Research Council Industrial CASE PhD Studentship with ADM Protexin as the industry partner and additional support from Freya Green.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Richard Day, MBBS, and Ashton Harper, MBBS (ADM Protexin), provided input into the design of the study pertaining to the intervention contents, dose, and contraindications. ADM Protexin provided the study product and placebo free of charge. The final manuscript was reviewed by ADM Protexin for confidentiality of company IP.

Disclaimer: This article represents independent research partly funded by the National Institute for Health and Care Research Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London.

Data Sharing Statement: See Supplement 3.

Additional Contributions: We kindly thank the study participants, ADM Protexin, the National Institute for Health and Care Research /Wellcome Lambeth, King’s College Hospital Clinical Research Facility, the FAST-R service by the National Institute for Health and Care Research Maudsley Biomedical Research Centre via King’s College London and South London and Maudsley NHS Foundation Trust, King’s College Clinical Trials Unit, the National Institute for Health and Care Research BioResource Consortium and South London and Maudsley NHS Foundation Trust Psychological Therapies Services Lewisham, Lambeth, and Croydon. We also thank all individuals who contributed to the conduct of the study at the Centre for Affective Disorders (Caroline Loveland, BA, Mutahira Qureshi, MBBS, Nefize Yalin, MD, PhD, Luke Jelen, MBBS, Tim Mantingh, MSc, Dimosthenis Tsapekos, PhD, Rebecca Strawbridge, PhD, Catherine Bird, MSc, and Zoe Kenealy). None of the contributors were compensated for their work.