Skip to main content
NCBI home page
Communications Medicine logoLink to Communications Medicine
. 2026 Jul 16;6:399. doi: 10.1038/s43856-026-01782-5

The effect of SSRI/SNRI antidepressant treatment on the gut microbiota of patients with major depressive disorder

Eugenia Emile Natasha 1,17,#, Danique Mulder 1,#, Leon Fehse 2, Nils R Winter 3, Lukas Fisch 3, Marius Welzel 2, Corinna Bang 4, Susanne Meinert 3,5, Kira Flinkenflügel 3, Tiana Borgers 3, Janik Goltermann 3, Elisabeth J Leehr 3, Carsten Culmsee 6,7, Frederike Stein 8, Florian Thomas-Odenthal 8, Paula Usemann 8, Lea Teutenberg 8, Igor Nenadic 8, Benjamin Straube 8, Nina Alexander 8, Hamidreza Jamalabadi 8, Andreas Jansen 8, Robert Nitsch 5, Andreas Lügering 9, Andre Franke 4, Udo Dannlowski 3,10, Tilo Kircher 6,8, Dominik Heider 2, Tim Hahn 3, Janna N Vrijsen 1,11,12, Philip van Eijndhoven 1, Indira Tendolkar 1, Andreas Reif 13,14,15, Sharmili Edwin Thanarajah 13,14,15, Silke Matura 13, Alejandro Arias Vasquez 1, Mirjam Bloemendaal 13,16,✉
PMCID: PMC13376374  PMID: 42463908

Abstract

Background

The gut microbiome has been linked to major depressive disorder (MDD), yet it remains unclear whether antidepressant treatment influences these associations. This study aimed to clarify the role of serotonin reuptake inhibitors (SSRI/SNRI) in shaping gut microbiome changes observed in MDD.

Methods

We conducted cross-sectional analyses in two independent patient cohorts (total N = 1802) and a meta-analysis across both cohorts, comparing the gut microbiome of MDD patients with and without SSRI/SNRI treatment.

Results

Here we show that SSRI/SNRI treatment is consistently associated with reduced Clostridium sensu stricto 1 abundance. This effect is specific to SSRI/SNRI treatment and not observed with other psychotropic medications. Importantly, reductions in Clostridium sensu stricto 1 in MDD compared to unaffected controls are explained by SSRI/SNRI medication status.

Conclusions

Antidepressant treatment is an important factor shaping gut microbiome alterations linked to MDD, underscoring the need to account for medication effects and potentially informing future microbiome-based strategies to improve treatment response.

Subject terms: Depression, Psychiatric disorders

Plain language summary

This study examined whether commonly used antidepressants (called SSRIs and SNRIs) influence the gut microbiome in people with depression. The gut microbiome is a community of bacteria in the intestines that may affect mental health. We analyzed data from more than 1,800 people, including individuals with depression who were taking or not taking SSRIs and SNRIs. We found that people using these medications consistently had lower levels of gut bacterium called Clostridium sensu stricto 1. This effect was not seen with other psychiatric medications. This means we identified a bacterium that is influenced by antidepressant use. In the future, this knowledge may help test whether the gut microbiome can improve treatment outcomes and support more personalized care.

Natasha, Mulder et al. investigate gut microbiome in major depressive disorder to disentangle disorder-associated signals from the effects of antidepressants. SSRI and SNRI antidepressant use is associated with lower abundance of Clostridium sensu stricto 1, suggesting that part of the microbial signature of depression is driven by medication.

Introduction

Major depressive disorder (MDD) is a debilitating condition that significantly reduces quality of life1. Despite its complex neurobiology, recent research highlights the gut microbiota as a potential contributor to MDD pathophysiology2. Microbial alterations may inform microbiota-targeting interventions to alleviate symptoms or improve treatment outcomes. However, findings on gut microbial changes in MDD have been inconsistent3,4, limiting identification of targetable features. These variations seem to arise partly due to within and between cohort variability in clinical presentation, host genetics, and environmental factors5,6. Another key source of variability is the use of psychotropic medications. Though psychotropics such as antidepressants and antipsychotics have demonstrated effects on the gut microbiota in animal models and in vitro, their importance has been underappreciated in clinical research7–11.

Psychotropic medication use is often inadequately controlled for in human cohort studies, due to limitations in clinical phenotyping and sample size. For example, population studies have examined the effects of antidepressants on the gut microbiome without considering disease status12,13. Furthermore, clinical cross-sectional studies often include both medicated and unmedicated patients without stratifying the analyses by medication status10, while longitudinal observational studies typically compare medicated patients only to unaffected individuals14,15. Additionally, psychotropic medications are frequently analyzed as a single group16,17, neglecting pharmacokinetic differences between drug classes that may result in varying effects on the gut microbiota7,18,19. Together, these designs limit insights into medication-specific effects.

In this work, using large clinical cohorts, we identify gut microbial genera whose associations with MDD differ depending on antidepressant treatment status, focusing on the most commonly prescribed antidepressant classes for MDD: Selective Serotonin and Norepinephrine Reuptake Inhibitors (SSRIs/SNRIs). In the discovery cohort (MACS, n = 1568), several MDD-associated genera show associations that are modified by SSRI/SNRI treatment, findings that are subsequently replicated in an independent cohort (MIND-Set, n = 234) and confirmed through meta-analysis. These results highlight SSRI/SNRI treatment as a modifier of the association between the gut microbiome and MDD, laying the groundwork for targeted studies investigating how antidepressant-induced microbial changes relate to treatment response and potentially informing microbiome-based therapeutic strategies in psychiatry.

Methods

Study populations

Cohort description

Marburg-Münster Affective Cohort Study (MACS) cohort

The MACS cohort20 is an observational, prospective, case-control study, conducted at the Department of Psychiatry and Psychotherapy at the University of Münster and University of Marburg, Germany. The study, described previously20,21, aims to investigate underlying neurobiological and environmental risk factors for affective disorders. The cohort comprises patients with an acute or lifetime diagnosis of an affective disorder and a group of unaffected individuals who reported no history of any psychiatric disorders. Individuals aged 18–67 years, of Caucasian ancestry, receiving in- or outpatient treatment from participating local clinics in Marburg or Münster, or responding to newspaper advertisements, were recruited between 2014 and 2018. Study procedures were conducted at the Marburg University Hospital and Münster University Hospital. Psychiatric diagnoses, including an MDD diagnosis, were determined by trained clinicians using the Structural Clinical Interview for DSM-IV, axis 1 disorders (SCID-I). Individuals were included if they had an active or lifetime major depressive episode, with or without somatic or other psychiatric comorbidities. For this project, only unaffected controls and individuals with an acute or lifetime MDD diagnosis, with or without comorbid anxiety disorder, were included. For a complete description of sample selection, see Supplementary Fig. 1. Upon the completion of baseline participation, individuals were invited to participate in a non-mandatory follow-up examination after 2 years. The study received ethical approval from the Ethics Committees of the Medical Faculties, University of Marburg (07/14), and University of Münster (2014-422-b-S). All participants provided written informed consent allowing the use of their data for the research questions examined in this study, and no additional institutional review board approval was required.

Measuring Integrated Novel Dimensions in Neurodevelopmental and Stress-related Mental Disorders (MIND-Set) cohort

The Measuring Integrated Novel Dimensions in Neurodevelopmental and Stress-related Mental Disorders (MINDSet) cohort22 was used as a replication cohort. The MIND-Set cohort is an observational, cross-sectional study, which aims to elucidate shared and specific mechanisms of stress-related and neurodevelopmental psychiatric disorders. Individuals aged 18 years and older were recruited through the Department of Psychiatry of the Radboud University Medical Center, Nijmegen, the Netherlands between 2016 and 2021. At the time of recruitment, participants were outpatients at the Department of Psychiatry of the Radboud University Medical Center. The SCID-I for mood and anxiety disorders was conducted by trained clinicians to confirm the MDD diagnosis, alongside other standardized instruments to assess the presence of other psychiatric and neurodevelopmental disorders (see Van Eijndhoven et al.22 for a description of all measurement instruments). Individuals were included if they fulfilled the criteria for a stress-related disorder (mood, anxiety or substance use disorders) and/or neurodevelopmental disorder (attention deficit/hyperactivity disorder (ADHD) or autism spectrum disorder (ASD)). For this project, only unaffected controls and individuals with an acute or lifetime MDD diagnosis, with or without comorbid psychiatric disorder(s), were included. For a complete description of sample selection, see Supplementary Fig. 1. A group of individuals free from psychiatric disorders was recruited from the local general population, screened via telephone interviews using the same diagnostic tools. Exclusion criteria comprised active psychotic disorders, IQ estimation <70, sensorimotor disabilities, insufficient Dutch language comprehension, or taking antibiotics at the time of recruitment. The study received ethical approvals from the local medical ethics committee (NL55618.091.015). All participants provided written informed consent allowing the use of their data for the research questions examined in this study, and no additional institutional review board approval was required

From both cohorts, we included participants who provided a fecal sample and had available information regarding MDD diagnosis, medication use, and relevant confounders (section Covariates).

Medication use

In the MACS cohort, a trained clinician registered current psychotropic medication use (ATC codes N05* and N06*) through a structured interview. Other medication use was not logged in detail. In the MIND-Set cohort, a pharmacy-employee collected medication history and current use, which was verified by a clinician during the study intake. We grouped psychotropic medications into three categories: SSRIs (ATC codes N06AB*), SNRIs (venlafaxine-N06AX16, milnacipran-N06AX17, duloxetine-N06AX21, vortioxetine-N06AX26), and other psychotropic agents (other N05* and N06*). Supplementary Data 1 provides the full list of psychotropic medications and their corresponding ATC codes.

Diagnosis and medication group definitions

We grouped participants based on the MDD diagnosis: individuals with an MDD diagnosis (active or a lifetime episode) were assigned to the MDD group. Individuals without any psychiatric diagnoses were assigned to the unaffected control group. Within the patient group, we stratified patients into three groups based on their use of SSRIs/SNRIs and other psychotropic medication (Fig. 1A). Patients not treated with any psychotropic medication were assigned to the unmedicated group. Patients treated with at least one SSRI or SNRI, either with or without co-medication with other psychotropic medications, were assigned to the SSRI/SNRI± group. Patients treated with only other psychotropic medication (no SSRI/SNRI) were assigned to the other psychotropic-only group. Non-psychotropic medications were not considered for this grouping. We focused on SSRIs and SNRIs, the most commonly prescribed pharmacotherapies for MDD, which primarily act by targeting the serotonin transporter (SERT). This choice enhances mechanistic interpretability because SERT is highly expressed in the gastrointestinal tract, where it modulates gut motility23,24, and influences the microbial community25.

Fig. 1. Analysis flowchart.

Fig. 1

Open in a new tab

The flowchart depicts the analysis structure. A Grouping based on treatment with psychotropic medications into three main groups: unmedicated, SSRI/SNRI-treated and other psychotropics. B MDD-associated microbial features were identified in the MACS cohort. All models were adjusted for age, gender, BMI, sequencing depth, sequencing batch, and collection site. C All MDD-associated microbial features were tested for associations with SSRI/SNRI treatment in the MACS and MIND-Set cohorts. These models were also adjusted for psychiatric comorbidities and depression symptom severity. MDD Major Depressive Disorder, FDR False Discovery Rate, SSRI Selective Serotonin Reuptake Inhibitor, SNRI Serotonin-Norepinephrine Reuptake Inhibitor.

Microbiota data acquisition and preprocessing

The protocol for DNA extraction, sequencing and preprocessing has been described in detail elsewhere for MACS26 and MIND-Set27. In short, participants from the MACS cohort were asked to collect a fecal sample on the day of their baseline intake interview or within two weeks after. Baseline samples were used by default; if unavailable, the 2-year follow-up samples were used along with corresponding clinical and questionnaire data. The DNA was extracted from the fecal samples, and the V1-V2 regions of the 16 s rRNA gene were amplified using the 27 F − 338 R primer pair (F:AGAGTTTGATCMTGGCTCAG, R:GCTGCCTCCCGTAGGAGT). Sequencing was performed using the Illumina MiSeq v3 platform (2x300bp), carried out in two batches, with samples from different timepoints processed in separate batches. Sequencing yielded an average of 36,335 [4008–11,3711] paired-end reads per sample. Raw sequences were pre-processed using the Natrix pipeline28, which clustered the reads into OTUs before using the BLAST algorithm against the SILVA database for taxonomic assignment29. The resulting feature table contained a total of 48,495 OTUs, corresponding to 636 genera.

For the MIND-Set cohort, all participants collected a fecal sample at home. DNA was extracted from the fecal samples, and the V4 region of the 16s rRNA gene was amplified using the 515F-806R primer pair (F:GTGCCAGCMGCCGCGGTAA, R:GGACTACHVGGGTWTCTAAT). Sequencing was performed using the Illumina NovaSeq6000 platform (2x250bp) in one batch, yielding an average of 768,466 [226,130 –1,266,720] paired-end reads per sample. Raw sequences were pre-processed using the QIIME2-DADA2 pipeline30,31, which inferred ASVs and then assigned taxonomy using a Naïve Bayes classifier pre-trained on the SILVA v138 reference database. The resulting feature table contained a total of 21,078 ASVs, corresponding to 243 genera.

Microbial diversity & taxonomy

All analyses were performed in R (version 4.4.132), using the mia, microbiome, vegan, and phyloseq packages33–35. Alpha diversity, reflecting within-sample richness and evenness, was quantified using the Inverse Simpson and Pielou’s indices36. These indices were selected as they showed the lowest correlation with other indices, indicating complementary information. Beta diversity, representing global between-sample differences in microbial composition, was quantified using pairwise Aitchison distances derived from center log-ratio (CLR) transformed abundance data37.

For taxonomic analyses, the microbial feature table was aggregated to the genus level. Genera detected in less than 10% of all the samples were excluded, which has shown to improve reproducibility across studies38. The remaining genera were centered log-ratio (CLR)-transformed to account for zero inflation and compositionality in the data39.

Statistics and reproducibility

To investigate the extent to which microbial differences in MDD are modified by SSRI/SNRI treatment, we conducted a multi-step analysis. First, we identified microbial associations (alpha/beta diversity, taxonomy) with MDD diagnosis in the larger, well-powered MACS cohort (N = 1568; Fig. 1B). Next, both in the MACS and MIND-Set (N = 234) cohorts, we assessed whether MDD-associated genera were also associated with SSRI/SNRI± treatment (Fig. 1B). We then performed a meta-analysis of the association between the gut microbiome and SSRI/SNRI treatment across the two cohorts. Finally, we performed post-hoc analyses to examine whether these associations were specifically driven by SSRI/SNRI treatment.

Microbial diversity and taxonomy

Associations with MDD

Associations between the microbiome and MDD were tested in MACS. For alpha diversity, we applied logistic regression (MDD diagnosis as the outcome, alpha diversity as the predictor) and quantile regression (median alpha diversity as the outcome, MDD diagnosis as the predictor). Logistic regression is appropriate for binary outcomes and captures non-linear associations40. Quantile regression makes minimal distributional assumptions and is robust to outliers, making it well-suited for sparse microbiome data41. Beta diversity differences between MDD groups were assessed with PERMANOVA using the Aitchison distance matrix as the outcome and MDD diagnosis as the predictor. To identify MDD-associated genera, we applied the same logistic and quantile regression models as for alpha diversity, fitting one model per genus. All models were adjusted for covariates (section Covariates), and p-values were adjusted for false discovery rate (FDR) using the Benjamini-Hochberg procedure, with significance set at pFDR < 0.05.

Associations with SSRI/SNRI treatment

Associations between the microbiome and SSRI/SNRI treatment were tested in both MACS and MIND-Set. For all diversity indices and microbial genera significantly associated with MDD, we repeated the same analysis scheme within the (stratified) MDD group, contrasting the SSRI/SNRI± against the unmedicated group. Finally, to improve statistical power, we performed a random-effects meta-analysis across both cohorts, combining the effect sizes from logistic and quantile regression assessing the association between SSRI/SNRI± treatment and the MDD-associated genera. Effect-sizes were weighted by within-study variance and between-study heterogeneity42.

SSRI/SNRI specificity

Two post-hoc analyses were performed: first, to test whether associations were specific to SSRI/SNRI treatment, we compared SSRI/SNRI-only and other-psychotropic-only groups against unmedicated patients. Second, to determine whether case-control differences were influenced by medication, we compared the SSRI/SNRI± and unmedicated MDD groups with unaffected controls. Both quantile and logistic regression models were assessed (Fig. 1C).

Covariates

All statistical models were adjusted for age, sex, and BMI as covariates, given their known associations with gut microbiome composition and/or MDD12,43. Sequencing depth (the number of reads retained after pre-processing) was included in all models, as it was significantly associated with the number of detected microbial features in both the MACS (r = .23, p < .001) and MIND-Set (r = .09, p < .001) cohorts. In the MACS cohort, sequencing batch was associated with the number of detected OTUs (F = 36.92, p < .001; Supplementary Fig. 2) and was therefore included as an additional covariate in all models.

Models examining associations with SSRI/SNRI treatment were further adjusted for depression symptom severity, assessed with the BDI-II44 in MACS and the IDS-SR45 in MIND-Set, as severity differed significantly between medication groups in both cohorts (MACS: F = 61.32, p < .001; MIND-Set: F = 14.68, p < .001). These models were also adjusted for the number of psychiatric comorbidities. Supplementary Data 2 provides a graphic overview of covariate adjustment across models.

Sensitivity analyses

To assess the robustness of associations between SSRI/SNRI treatment and gut microbiota composition in MDD, sensitivity analyses were performed to control for remission status, appetite change, somatic comorbidity and three dietary factors (Supplementary Data 2). In both cohorts, we examined whether associations persisted after accounting for remission status, categorized as ‘active’, ‘partially remitted’, and ‘remitted’ in MACS, and as ‘active’ and ‘remitted’ in MIND-Set. In the MACS cohort, additional sensitivity analyses adjusted for appetite change, somatic comorbidity, and dietary factors. Appetite change was assessed using item A3 from the SIGH-HAMD46 (“Has your appetite been greater than when you feel well or okay?”). Somatic comorbidity was encoded as a binary variable reflecting the presence or absence of any somatic diagnosis. Dietary fiber intake (g), total calorie intake (kJ), and soft drink consumption¹⁴ were derived from a validated German semiquantitative food frequency questionnaire (FFQ2)47. Sensitivity analyses were applied to all microbial features significantly linked to SSRI/SNRI use, using logistic and quantile regression models, with the sensitivity variable included as an additional covariate.

Results

Cohort characteristics

Table 1 and Supplementary Data 3 provide a demographic, clinical and technical overview of the MACS and MIND-Set cohorts. The MACS cohort included 811 unaffected controls and 757 MDD patients. Among MDD patients, 352 were treated with SSRIs/SNRIs (n = 177 SSRI/SNRI-only; n = 175 SSRI/SNRI + other psychotropics), 101 used only other psychotropics, and 304 were unmedicated. Four unaffected controls using psychotropic medications were excluded. The MIND-Set cohort comprised 67 controls and 167 MDD patients. Among MDD patients, 42 were treated with SSRIs/SNRIs (n = 28 SSNRI/SNRI-only and n = 14  SSRI/SNRI + other psychotropics), 46 used only other psychotropics, and 79 were unmedicated. An overview of individual SSRI/SNRI types used in MACS and MIND-Set is presented in Fig. 2 and Supplementary Fig. 3.

Table 1.

Demographic characteristics of the MACS and MIND-set cohorts

N Age (mean (SD)) Sex (M/F) BMI (mean (SD)) Depression symptom severitya Remission status (active episode/partial remission/remission) # psychiatric comorbidities (1/2/3/4 + )
MACS Unaffected control 811 35.7 (13.3) 282/529 24.4 (4.66) 3.86 (4.2) - -
MDD All 757 37.1 (13.0) 257/500 26.2 (5.88) 16.9 (11.3) 294/187/275 585/172/-/-
Unmedicated 304 35.6 (12.6) 89/216 25.6 (6.01) 13.1 (9.67) 74/57/172 246/58/-/-
SSRI/SNRI ± 352 38.1 (13.0) 122/230 26.7 (5.95) 19.6 (11.4) 180/91/81 257/95/-/-
Other psychotropics only 101 38.6 (14.0) 46/55 26.4 (5.08) 19.4 (12.4) 40/39/22 82/19/-/-
MIND-Set Unaffected control 67 39.0 (16.6) 33/0/34 23.9 (4.4) 5.15 (4.5) - -
MDD All 167 42.0 (14.7) 87/0/80 26.6 (10.2) 34.2 (13.0) 92/0/75 30/51/34/52
Unmedicated 79 36.7 (13.5) 45/0/34 24.6 (4.7) 29.7 (11.9) 35/0/44 11/30/13/25
SSRI/SNRI± 42 49.9 (13.7) 19/0/23 28.2 (4.9) 38.9 (13.6) 29/0/13 11/13/8/10
Other psychotropics only 46 44.3 (14.1) 23/0/23 28.6 (17.6) 37.8 (11.7) 28/0/18 8/8/13/17
Open in a new tab

aDepression symptom severity is measured using BDI in MACS and using the IDS-SR in MIND-Set.

MDD Major Depressive Disorder, SSRI/SNRI ±  SSRI/SNRI +/- other psychotropic medication.

Fig. 2. SSRI and SNRI treatment.

Fig. 2

Open in a new tab

An overview of the distribution of medication groups in the MACS (A) and MIND-Set (C) cohorts, along with the individual SSRI and SNRI types within the MACS (B) and MIND-Set (D) cohorts.

In both cohorts, MDD patients were older and had a higher BMI than controls. The SSRI/SNRI group was also older, with a higher BMI and depression severity than the unmedicated group. In MACS, the SSRI/SNRI group had a higher proportion of males than the unmedicated group (35% vs. 29%) (Supplementary Data 4).

Microbial associations with MDD

We found no significant associations between MDD diagnosis and alpha diversity (Supplementary Data 5), but there was a significant association between Aitchison’s distance and MDD diagnosis, F = 1.660, R2 = 0.001, p = 0.011 (Supplementary Data 6). The full regression models, including covariate effects, are presented in Supplementary Data 16–31.

Nine out of 188 genera showed a statistically significant association with MDD diagnosis in both logistic and quantile regression models (Supplementary Fig. 4 and 5, Supplementary Data 7–9). Confirming previous findings in this cohort26, seven genera showed a higher abundance in individuals with MDD, including Hungatella and Eggerthella, while the remaining two, Clostridium sensu stricto 1 and Lachnospiraceae FCS020 group, showed a lower abundance. An additional eight genera were identified only through logistic regression, and five genera were only identified through quantile regression (Supplementary Data 8 and 9).

Microbial associations with SSRI/SNRI treatment

MACS

In the MACS cohort, there were no significant differences in Aitchison’s distance between the SSRI/SNRI± and unmedicated patient groups, F = 1.078, R2 = .002, p = .232 (Supplementary Data 6). Out of the nine MDD-associated genera, one was also associated with SSRI/SNRI± treatment: Clostridium sensu stricto 1 showed a significantly lower abundance in SSRI/SNRI ± -treated compared to unmedicated patients, both in the logistic and quantile regression (Fig. 3, Supplementary Data 10). None of the other covariates (age, sex, BMI, anxiety symptoms, depression symptoms, collection site and sequencing depth,) were significantly associated with Clostridium sensu stricto 1 abundance in the quantile regression model (Supplementary Data 18).

Fig. 3. Meta-analysis forest plot of the associations between genera and SSRI/SNRI treatment.

Fig. 3

Open in a new tab

Outcomes of the random-effects meta-analysis associations between SSRI/SNRI treatment and MDD-associated genera from logistic regression (left) and quantile regression (right). The percentages represent the study weight. The squares represent log odds ratios (logistic regression) and beta coefficients (quantile regression). The confidence bars represent the study-level 95% confidence interval. The diamond width represents the confidence interval (95% CI) of the (overall) summary estimate.

MIND-Set

In the MIND-Set cohort, there were no significant differences in Aitchison’s distance between the SSRI/SNRI± and unmedicated patient groups, F = 1.029, R2 = 0.012, p = 0.306 (Supplementary Data 6). Taxonomic analyses revealed a significantly lower abundance of Clostridium sensu stricto 1 in SSRI/SNRI-treated compared to unmedicated patients in the quantile regression analysis (Fig. 3). In logistic regression analysis the association was only nominally significant (Supplementary Data 10).

Meta-analysis

Random-effects meta-analysis showed a significant negative association between SSRI/SNRI± treatment and Clostridium sensu stricto 1, consistent across logistic and quantile regression analyses (Fig. 3). Anaerotruncus showed a nominal positive association with SSRI/SNRI± treatment in both quantile and logistic regression meta-analyses, although this did not survive correction for multiple testing and was not consistently observed in the individual cohorts. No associations with SSRI/SNRI± treatment were detected for the other seven MDD-associated genera (Supplementary Data 11).

SSRI/SNRI specificity

The identified associations with Clostridium sensu stricto 1 abundance appear to be specific to the use of SSRI/SNRIs. Across both cohorts, Clostridium sensu stricto 1 abundance was significantly lower in the SSRI/SNRI-only group but showed no differences in the other-psychotropic-only group, compared to unmedicated patients (Supplementary Data 12).

Moreover, the reduced Clostridium sensu stricto 1 abundance observed in MDD patients compared to controls also seemed to be driven by SSRI/SNRI treatment. When comparing the SSRI/SNRI± and the unmedicated MDD groups to unaffected controls, stronger associations were found in the SSRI/SNRI± group for most MDD-associated genera (Supplementary Fig. 5, Supplementary Data 13). Specifically, across both cohorts, the association with Clostridium sensu stricto 1 was significant and exhibited larger effect sizes in the SSRI/SNRI± group, whereas no significant association was observed, and effect sized were markedly smaller, in the unmedicated group (Supplementary Fig. 5, Supplementary Data 13).

Sensitivity analyses

The association between Clostridium sensu stricto 1 and SSRI/SNRI treatment remained consistent after adjustment for remission status, appetite changes, and somatic comorbidities. Dietary adjustment (fiber intake, caloric intake, and soft drink consumption) was possible in a subset of participants from the MACS cohort where these data were available (N = 222). Effect sizes remained comparable to the unadjusted models (Supplementary Data 14), suggesting that dietary differences between treatment groups do not explain the observed reduction in Clostridium sensu stricto 1. Associations were, however, no longer statistically significant, which is likely attributable to the reduced sample size rather than confounding by diet.

Discussion

In this study, we aimed to identify the extent to which SSRIs and SNRIs modify gut microbial differences associated with MDD, thereby distinguishing disorder-related signals from those more likely shaped by pharmacological treatment. We identified nine genera with differential abundance between MDD patients and controls, including elevated Eggerthella and Hungatella. This replicates findings in a largely overlapping sub-sample of the MACS cohort26,48 and from independent studies4,27,49, especially the independently replicated findings underscoring their consistency as markers of the disorder. Among these taxa, Clostridium sensu stricto 1 showed a distinct pattern in relation to medication: its reduced abundance in MDD patients appeared to be driven by SSRI/SNRI treatment, with lower levels observed in treated compared to unmedicated patients, and its association with MDD diagnosis only evident in the treated group. This identifies Clostridium sensu stricto 1 as a potentially medication-sensitive genus, while the other MDD associated genera, which were not affected by medication status, may reflect other disorder-related factors instead, such as host genetics of lifestyle. Anaerotruncus showed only nominal associations in the meta-analysis across cohorts, suggesting caution in interpretation. This association does align with a previously reported association between Anaerotruncus colihominis and psychotropic medications use16. Therefore, further research into Anaerotruncus, particularly regarding medication specificity and the functional implications of its increased abundance, is warranted.

Clinical studies examining the impact of SSRIs and SNRIs on the gut microbiome show mixed results: some cross-sectional studies show decreased Clostridium or Clostridiaceae abundance with SSRI/SNRI treatment10,13,50, whereas others find no such differences but report associations with other taxa16,17,51. These inconsistencies may result from small sample sizes, and incomplete adjustment for clinical and demographic factors like symptom severity, psychiatric comorbidities, socioeconomic status, and geographical location. Since medicated and unmedicated patients often differ in these factors, insufficient control can obscure antidepressants’ effects on gut microbiota. Our study overcomes this by stratifying patients by medication status and adjusting for demographic and clinical covariates, allowing more precise microbiome-antidepressant association estimates.

Mechanistically, SSRIs and SNRIs act by inhibiting the serotonin transporter (SERT) in the central nervous system, but they also exert unintentional effects in the gut that may directly or indirectly affect the microbiome. For example, SSRIs and SNRIs can influence gut motility via serotonin receptor modulation, potentially causing diarrhea or constipation and indirectly altering the microbiome52. In addition, these antidepressants have direct antimicrobial effects in vitro: they can inhibit bacterial growth by penetrating bacterial membranes, disrupting efflux pump function, and compromising membrane integrity53–55. This may particularly affect Gram-positive bacteria, including Clostridium sensu stricto 1, due to their membrane structure and limited efflux capacity in comparison to Gram-negative bacteria56. Supporting this, metagenomic analyses in clinical cohorts revealed that the SSRI escitalopram reduced the abundance of Clostridium species and induced upregulation of bacterial survival pathways, including efflux pumps, antibiotic resistance, and sporulation genes, paralleling experimental findings15. These antimicrobial actions may impact host-microbe interactions, including peripheral serotonin metabolism. That is, Gram positive spore-forming bacteria (such as Clostridium sensu stricto 1) regulate gut-derived serotonin by producing metabolites that stimulate serotonin synthesis by the enterochromaffin cells57,58. Reductions in the abundance of these bacteria during SSRI/SNRI treatment may therefore alter gut serotonin dynamics, as shown preclinically for the spore-forming Turicibacter sanguinis57. Consistently, clinical studies have shown that SSRI/SNRI treatment induced decreases in spore-forming bacteria alongside shifts in peripheral serotonin metabolism15,59. The findings from this study depend on 16S-derived taxonomic profiles, limiting the ability to draw conclusions about functional changes. Establishing the functional relevance of these findings will require complementary approaches, including experimental validation and/or metagenomic, metatranscriptomic, or proteomic analyses.

Emerging evidence furthermore links SSRI-induced shifts in microbial composition and metabolic activity to treatment outcomes, including lower abundance of sporulation genes and higher plasma tryptophan in SSRI responders14,15,59. Notably, an open-label trial in treatment-resistant depression found that adjunctive administration of Clostridium butyricum (a member of the Clostridium sensu stricto 1 genus) improved treatment response60, suggesting that abundance of this species, whether pre-existing or medication-induced, may modulate treatment efficacy. Prospective studies incorporating repeated multi-omics sampling and detailed clinical phenotyping are needed to establish whether these associations are causal. This will guide the development and implementation of adjunctive interventions, such as targeted probiotics (e.g., C. butyricum) or dietary strategies, to improve therapeutic outcomes.

Our study benefits from a large, well-characterized patient cohort and independent replication, strengthening confidence in and generalizability of the findings. Nonetheless, several limitations remain. First, limited data on medication dose, treatment duration, and follow-up constrained our ability to distinguish acute from chronic medication effects and to link microbial changes to symptom trajectories. Our findings likely reflect long-term microbial adaptations rather than short-term effects, which may involve different mechanisms or temporal dynamics. Second, limited information on side effects, medication dosage, comedication and treatment duration, restricted our ability to assess the effect of medication burden. However, given that observed associations were not replicated in patients treated with non-SSRI/SNRI psychotropics, it seems unlikely that the signal is explained by general medication burden alone. Third, while evidence shows that microbial susceptibility varies not only by antidepressant class, but also by the specific compound7,9,11, subgroups of individual antidepressants were generally small and highly imbalanced, precluding assessment of compound-level effects on the gut microbiome. Still, we could distinguish that the observed associations are specific to SSRIs/SNRIs and do not generalize to other classes of psychotropic medication. Finally, reliance on fecal samples limits insight into the small intestine (and its microbiome), where SSRIs and SNRIs are thought to exert their primary effects due to higher expression of the serotonin transporter (SERT)61. As a result, drug-microbiome interactions occurring in upper intestinal sites may not be captured.

Conclusion

This study shows that SSRI/SNRI treatment shapes gut microbial alterations in MDD, with reductions in Clostridium sensu stricto 1 driven by antidepressant use, suggesting a role for gut serotonin metabolism. These findings clarify the drivers of microbial alterations in depression and lay the groundwork for microbiome-targeted interventions to improve treatment outcomes in MDD.

Supplementary information

Transparent Peer Review file (477.8KB, pdf)
Supplemental Information (1.4MB, pdf)
43856_2026_1782_MOESM3_ESM.pdf (110.5KB, pdf)

Description of Additional Supplementary Files

Supplementary Data 1-31 (721.6KB, xlsx)

Acknowledgments

Biosamples and corresponding data were sampled, processed and stored in the Marburg Biobank CBBMR. This work used the Dutch national e-infrastructure with the support of the SURF Cooperative using grant no. EINF-10820.

Author contributions

MB, AAV, EEN, and DM contributed to the conceptualization and methodology of the study. Formal analysis was performed by EEN and DM, with validation by MB and AAV. The original draft was written by DM and EEN. Visualization was carried out by DM and EEN. Supervision was provided by MB and AAV. Funding acquisition was secured by MB, AAV, TK, and UD. The manuscripts was reviewed and edited by MB, AAV, Leon Fehse, NRW, Lukas Fisch, MW, CB, SM, KF, TB, JG, EL, CC, FS, FTO, PU, LT, IN, BS NA, HJ, AJ, RN AL, AF, UD TK, DH, TH, JV, PvE, IT, AR, SET and SM.

Peer review

Peer review information

. Communications Medicine thanks Bangmin Yin and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work received funding from grants to Mirjam Bloemendaal from European Union’s Horizon Europe research and innovation programme under the Marie Curie Postdoctoral Fellowship Grant Agreement (#82601081) and the Brain & Behavior Research Foundation Young Investigator Grant (#32385). Alejandro Arias Vasquez received funding from the NWO-ENW Open Competition Domain Science Round 2025-2026 (OCENW.XL.25.054). Mirjam Bloemendaal and Alejandro Arias Vasquez were funded through the NWO-ENW Open Competition Domain Science Package 22-3 (OCENW.M.22.177). In addition, this work received funding from consortia grants to Tilo Kircher from the German Research Foundation (DFG) FOR 2107, SFB/TRR 393 (“Trajectories of Affective Disorders”, project grant no 521379614), and the Germany’s Excellence Strategy (EXC 3066/1 “The Adaptive Mind”, Project No. 533717223), as well as the DYNAMIC center, funded by the LOEWE program of the Hessian Ministry of Science and Arts (grant number: LOEWE1/16/519/03/09.001(0009)/98). Udo Dannlowski was funded by the German Research Foundation (DFG, grant FOR2107 DA1151/5-1, DA1151/5-2, DA1151/9-1, DA1151/10-1, DA1151/11-1 to UD; SFB/TRR 393, project grant no 521379614) and the Interdisciplinary Center for Clinical Research (IZKF) of the medical faculty of Münster (grant Dan3/022/22 to UD). Open Access funding enabled and organized by Projekt DEAL.

Code availability

The code supporting the findings of this study, along with software versions, is available at 10.5281/zenodo.20394496.

Data availability

The raw clinical metadata, and microbial community datasets generated and analyzed during this study are available from Tim Hahn (hahnt@uni-muenster.de) upon reasonable request. Anonymized data can be accessed directly, provided that the intended use of the data is clearly specified in writing at the time of request. Requests will be reviewed within a timeframe of up to four weeks. All summary statistics produced in this study are included with the manuscript as Supplementary Data 16–31. The source data for Fig. 2 are provided in Supplementary Data 14, and the source data for Fig. 3 are provided in Supplementary Data 11.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Eugenia Emile Natasha, Danique Mulder.

These authors jointly supervised this work: Alejandro Arias Vasquez, Mirjam Bloemendaal.

Supplementary information

The online version contains supplementary material available at 10.1038/s43856-026-01782-5.

References

  • 1.Marx, W. et al. Major depressive disorder. Nat. Rev. Dis. Prim.9, 44 (2023). [DOI] [PubMed] [Google Scholar]
  • 2.Valles-Colomer, M. et al. The neuroactive potential of the human gut microbiota in quality of life and depression. Nat. Microbiol.4, 623–632 (2019). [DOI] [PubMed] [Google Scholar]
  • 3.Gao, M. et al. Gut microbiota composition in depressive disorder: a systematic review, meta-analysis, and meta-regression. Transl. Psychiatry13, 379 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Nikolova, V. L. et al. Perturbations in gut microbiota composition in psychiatric disorders: a review and meta-analysis. JAMA Psychiatry78, 1343–1354 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Goldberg, D. The heterogeneity of “major depression. World Psychiatry10, 226–228 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Kurilshikov, A. et al. Large-scale association analyses identify host factors influencing human gut microbiome composition. Nat. Genet.53, 156–165 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Ait Chait, Y., Mottawea, W., Tompkins, T. A. & Hammami, R. Unravelling the antimicrobial action of antidepressants on gut commensal microbes. Sci. Rep.10, 17878 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Borgiani, G. et al. The bidirectional interaction between antidepressants and the gut microbiota: are there implications for treatment response? Int. Clin. Psychopharmacol.40, 3–26 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Cussotto, S., Clarke, G., Dinan, T. G. & Cryan, J. F. Psychotropics and the microbiome: a chamber of secrets…. Psychopharmacol. (Berl.)236, 1411–1432 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Gamboa, J. et al. Impact of antidepressants on the composition of the gut microbiome: A systematic review and meta-analysis of in vivo studies. J. Affect. Disord.369, 819–833 (2025). [DOI] [PubMed] [Google Scholar]
  • 11.Maier, L. et al. Extensive impact of non-antibiotic drugs on human gut bacteria. Nature555, 623–628 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Falony, G. et al. Population-level analysis of gut microbiome variation. Science352, 560–564 (2016). [DOI] [PubMed] [Google Scholar]
  • 13.Jackson, M. A. et al. Gut microbiota associations with common diseases and prescription medications in a population-based cohort. Nat. Commun.9, 2655 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Jiang, Y. et al. The role of gut microbiota and metabolomic pathways in modulating the efficacy of SSRIs for major depressive disorder. Transl. Psychiatry14, 1–11 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Wang, Y. et al. Multi-omics reveal microbial determinants impacting the treatment outcome of antidepressants in major depressive disorder. Microbiome11, 195 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Brushett, S. et al. Gut feelings: the relations between depression, anxiety, psychotropic drugs and the gut microbiome. Gut Microbes15, 2281360 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Dilmore, A. H. et al. Medication use is associated with distinct microbial features in anxiety and depression. Mol. Psychiatry30, 2545–2557 (2025). [DOI] [PMC free article] [PubMed]
  • 18.Vich Vila, A. et al. Impact of commonly used drugs on the composition and metabolic function of the gut microbiota. Nat. Commun.11, 362 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Lukić, I. et al. Antidepressants affect gut microbiota and Ruminococcus flavefaciens is able to abolish their effects on depressive-like behavior. Transl. Psychiatry9, 133 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Iovoli, F. et al. Exploring the complex interrelation between depressive symptoms, risk, and protective factors: A comprehensive network approach. J. Affect. Disord.355, 12–21 (2024). [DOI] [PubMed] [Google Scholar]
  • 21.Vogelbacher, C. et al. The Marburg-Münster Affective Disorders Cohort Study (MACS): A quality assurance protocol for MR neuroimaging data. NeuroImage172, 450–460 (2018). [DOI] [PubMed] [Google Scholar]
  • 22.van Eijndhoven, P. et al. Measuring integrated novel dimensions in neurodevelopmental and stress-related mental disorders (MIND-SET): Protocol for a cross-sectional comorbidity study from a research domain criteria perspective. JMIRx Med.3, e31269 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kułak-Bejda, A., Bejda, G. & Waszkiewicz, N. Antidepressants for irritable bowel syndrome—A systematic review. Pharmacol. Rep.69, 1366–1379 (2017). [DOI] [PubMed] [Google Scholar]
  • 24.Shine, J. M. et al. Understanding the effects of serotonin in the brain through its role in the gastrointestinal tract. Brain145, 2967–2981 (2022). [DOI] [PubMed] [Google Scholar]
  • 25.Singhal, M. et al. Serotonin transporter deficiency is associated with dysbiosis and changes in metabolic function of the mouse intestinal microbiome. Sci. Rep.9, 2138 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Fehse, L. et al. From Gut to Brain: Evidence for a Causal Contribution of Gut-Microbiota to Major Depressive Disorder in Humans. Preprint at 10.1101/2024.12.05.24318549 (2024).
  • 27.Mulder, D. et al. Gut microbiota composition links to variation in functional domains across psychiatric disorders. Brain. Behav. Immun.120, 275–287 (2024). [DOI] [PubMed] [Google Scholar]
  • 28.Welzel, M. et al. Natrix: a Snakemake-based workflow for processing, clustering, and taxonomically assigning amplicon sequencing reads. BMC Bioinforma.21, 526 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Quast, C. et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res.41, D590–D596 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Bolyen, E. et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol.37, 852–857 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods13, 581–583 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing.
  • 33.Borman, T., Ernst, F., Shetty, S. & Lahti, L. mia: Microbiome analysis. 10.18129/B9.bioc.mia, R package version 1.18.0, https://bioconductor.org/packages/mia. (2024).
  • 34.McMurdie, P. J. & Holmes, S. phyloseq: An R package for reproducible interactive analysis and graphics of microbiome census data. PLoS ONE8, e61217 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Oksanen, J. et al. vegan: Community Ecology Package. 210.32614/CRAN.package.vegan (2001).
  • 36.Cassol, I., Ibañez, M. & Bustamante, J. P. Key features and guidelines for the application of microbial alpha diversity metrics. Sci. Rep.15, 622 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Aitchison, J. The statistical analysis of compositional data. J. R. Stat. Soc. Ser. B Stat. Methodol.44, 139–160 (1982). [Google Scholar]
  • 38.Nearing, J. T. et al. Microbiome differential abundance methods produce different results across 38 datasets. Nat. Commun.13, 342 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Gloor, G. B., Macklaim, J. M., Pawlowsky-Glahn, V. & Egozcue, J. J. Microbiome Datasets Are Compositional: And This Is Not Optional. Front. Microbiol. 8, (2017). [DOI] [PMC free article] [PubMed]
  • 40.Stoltzfus, J. C. Logistic regression: a brief primer. Acad. Emerg. Med.18, 1099–1104 (2011). [DOI] [PubMed] [Google Scholar]
  • 41.Koenker, R. & Hallock, K. F. Quantile Regression. J. Econ. Perspect.15, 143–156 (2001). [Google Scholar]
  • 42.Dettori, J. R., Norvell, D. C. & Chapman, J. R. Fixed-Effect vs Random-Effects Models for Meta-Analysis: 3 Points to Consider. Glob. Spine J.12, 1624–1626 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Melin, E. O. et al. Depression was associated with younger age, female sex, obesity, smoking, and physical inactivity, in 1027 patients with newly diagnosed type 2 diabetes: a Swedish multicentre cross-sectional study. BMC Endocr. Disord.22, 273 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Smarr, K. L. & Keefer, A. L. Measures of depression and depressive symptoms: Beck Depression Inventory-II (BDI-II), Center for Epidemiologic Studies Depression Scale (CES-D), Geriatric Depression Scale (GDS), Hospital Anxiety and Depression Scale (HADS), and Patient Health Questionnaire-9 (PHQ-9). Arthritis Care Res. 63, (2011). [DOI] [PubMed]
  • 45.Van Eeden, W. A., Van Hemert, A. M., Carlier, I. V. E., Penninx, B. W. & Giltay, E. J. Severity, course trajectory, and within-person variability of individual symptoms in patients with major depressive disorder. Acta Psychiatr. Scand.139, 194–205 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Williams, J. B. W. A structured interview guide for the hamilton depression rating scale. Arch. Gen. Psychiatry45, 742 (1988). [DOI] [PubMed] [Google Scholar]
  • 47.Nöthlings, U., Hoffmann, K., Bergmann, M. M. & Boeing, H. Fitting portion sizes in a self-administered food frequency questionnaire. J. Nutr.137, 2781–2786 (2007). [DOI] [PubMed] [Google Scholar]
  • 48.Edwin Thanarajah, S. et al. Soft drink consumption and depression mediated by gut microbiome alterations. JAMA Psychiatry82, 1095 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Radjabzadeh, D. et al. Gut microbiome-wide association study of depressive symptoms. Nat. Commun.13, 7128 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Lin, S.-K. K. et al. Exploring the human gut microbiota targets in relation to the use of contemporary antidepressants. J. Affect. Disord.344, 473–484 (2024). [DOI] [PubMed] [Google Scholar]
  • 51.Tomizawa, Y. et al. Effects of psychotropics on the microbiome in patients with depression and anxiety: considerations in a naturalistic clinical setting. Int. J. Neuropsychopharmacol.24, 97–107 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Oliva, V. et al. Gastrointestinal side effects associated with antidepressant treatments in patients with major depressive disorder: A systematic review and meta-analysis. Prog. Neuropsychopharmacol. Biol. Psychiatry109, 110266 (2021). [DOI] [PubMed] [Google Scholar]
  • 53.Bohnert, J. A., Szymaniak-Vits, M., Schuster, S. & Kern, W. V. Efflux inhibition by selective serotonin reuptake inhibitors in Escherichia coli. J. Antimicrob. Chemother.66, 2057–2060 (2011). [DOI] [PubMed] [Google Scholar]
  • 54.Klünemann, M. et al. Bioaccumulation of therapeutic drugs by human gut bacteria. Nature597, 533–538 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Wang, Y. et al. Antidepressants can induce mutation and enhance persistence toward multiple antibiotics. Proc. Natl. Acad. Sci.120, e2208344120 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Ayaz, M. et al. Sertraline enhances the activity of antimicrobial agents against pathogens of clinical relevance. J. Biol. Res. -Thessalon.22, 4 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Fung, T. C. et al. Intestinal serotonin and fluoxetine exposure modulate bacterial colonization in the gut. Nat. Microbiol.4, 2064–2073 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Yano, J. M. et al. Indigenous bacteria from the gut microbiota regulate host serotonin biosynthesis. Cell161, 264–276 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Bhattacharyya, S. et al. Metabolomic signature of exposure and response to citalopram/escitalopram in depressed outpatients. Transl. Psychiatry9, 173 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Miyaoka, T. et al. Clostridium butyricum MIYAIRI 588 as adjunctive therapy for treatment-resistant major depressive disorder: a prospective open-label trial. Clin. Neuropharmacol.41, 151–155 (2018). [DOI] [PubMed] [Google Scholar]
  • 61.Gill, R. K. et al. Function, expression, and characterization of the serotonin transporter in the native human intestine. Am. J. Physiol. -Gastrointest. Liver Physiol.294, G254–G262 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Transparent Peer Review file (477.8KB, pdf)
Supplemental Information (1.4MB, pdf)
43856_2026_1782_MOESM3_ESM.pdf (110.5KB, pdf)

Description of Additional Supplementary Files

Supplementary Data 1-31 (721.6KB, xlsx)

Data Availability Statement

The code supporting the findings of this study, along with software versions, is available at 10.5281/zenodo.20394496.

The raw clinical metadata, and microbial community datasets generated and analyzed during this study are available from Tim Hahn (hahnt@uni-muenster.de) upon reasonable request. Anonymized data can be accessed directly, provided that the intended use of the data is clearly specified in writing at the time of request. Requests will be reviewed within a timeframe of up to four weeks. All summary statistics produced in this study are included with the manuscript as Supplementary Data 16–31. The source data for Fig. 2 are provided in Supplementary Data 14, and the source data for Fig. 3 are provided in Supplementary Data 11.

Articles from Communications Medicine are provided here courtesy of Nature Publishing Group

RESOURCES