CoChat

A hidden confounder for microbiome studies: medications used years before sample collection

Open access

The DOI matches its Crossref record; 35 of the 35 references it cites were found in a registry or an index.

Abstract

Medication usage is a known contributor to the inter-individual variability of the gut microbiome. However, medications are often used repeatedly and for long periods, a notion yet unaccounted for in microbiome studies. Recently, we and others showed that not only the usage of antibiotics and antidepressants at sampling, but also past consumption, is associated with the gut microbiome. This effect can be “additive”—the more a medication is used, the stronger the impact on the microbiome. Here, by utilizing retrospective medication usage data from the electronic health records and the observational Estonian microbiome cohort shotgun metagenomics data set (n = 2,509), we systematically evaluate the long-term effects of antibiotics and human-targeted medications on the gut microbiome. We show that past usage of medications is associated with the gut microbiome. For example, the effects of antibiotics, psycholeptics, antidepressants, proton pump inhibitors, and beta-blockers are detectable several years after use. Furthermore, by analyzing a subcohort (n = 328) with a second microbiome characterization, we show that similar changes in the gut microbiome occur after treatment initiation or discontinuation, possibly indicating causal effects. IMPORTANCE This is the first study using detailed retrospective medication usage data from electronic health records to systematically assess the long-term effects of medication usage on the gut microbiome. We identified carryover and additive…

Overview

Integrity signals

Registry

Crossref · 10.1128/msystems.00541-25

Cites

35 of 35 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Figures

Fig 1 Drug usage in the EstMB cohort. (a) Overview of the EstMB cohort and phenotype data available through the electronic health records (EHR). We aimed to analyze the direct effects of drug usage at the time of sampling (Q1), carryover effects induced by past drug usage (Q2), and additive effects arising from the differing amounts of drugs used in the past (Q3). In a subcohort of the EstMB (n = 328) that provided an additional stool sample (T2), we analyzed the alterations in the microbiome attributable to treatment initiation (Q4). (b) Total number of different drugs used in the EstMB at different ATC levels. (c) Distribution of the number of drugs used by the participants at T1 across different ATC levels. (d) The number of drug users at T1 on ATC4 level; drugs with at least 50 users at T1 are shown. Drugs are ordered according to the number of active drug users. Antibiotics are highlighted in bold. ATC, Anatomical Therapeutic Chemical classification; Q, study question.
Fig 1 Drug usage in the EstMB cohort. (a) Overview of the EstMB cohort and phenotype data available through the electronic health records (EHR). We aimed to analyze the direct effects of drug usage at the time of sampling (Q1), carryover effects induced by past drug usage (Q2), and additive effects arising from the differing amounts of drugs used in the past (Q3). In a subcohort of the EstMB (n = 328) that provided an additional stool sample (T2), we analyzed the alterations in the microbiome attributable to treatment initiation (Q4). (b) Total number of different drugs used in the EstMB at different ATC levels. (c) Distribution of the number of drugs used by the participants at T1 across different ATC levels. (d) The number of drug users at T1 on ATC4 level; drugs with at least 50 users at T1 are shown. Drugs are ordered according to the number of active drug users. Antibiotics are highlighted in bold. ATC, Anatomical Therapeutic Chemical classification; Q, study question.CC BY · PubMed Central
Fig 2 Active drug usage effects on the microbiome. (a) The inter-individual variance of the microbiome explained by individual drug classes. (b) Partial Pearson correlation between drug classes at the ATC4 level and observed richness and Shannon diversity index. (c) Partial Pearson correlation between drug classes at the ATC4 level and bacterial species analyzed as presence-absence and as abundance (centered log-ratio transformed) data. The results of both analyses refer to the same bacteria, i.e., the same column refers to the same bacterial species in presence-absence and in abundance analyses. Further, only drugs and microbes with at least 10 confidently deconfounded nominally significant associations (P < 0.05) in two analyses combined are shown, while drugs are presented in descending order based on the number of associations detected (top 20 drugs shown). The colored cells indicate nominally significant and confidently deconfounded associations (P < 0.05). *FDR ≤ 0.1 and confidently deconfounded. Antibiotics are highlighted in bold.
Fig 2 Active drug usage effects on the microbiome. (a) The inter-individual variance of the microbiome explained by individual drug classes. (b) Partial Pearson correlation between drug classes at the ATC4 level and observed richness and Shannon diversity index. (c) Partial Pearson correlation between drug classes at the ATC4 level and bacterial species analyzed as presence-absence and as abundance (centered log-ratio transformed) data. The results of both analyses refer to the same bacteria, i.e., the same column refers to the same bacterial species in presence-absence and in abundance analyses. Further, only drugs and microbes with at least 10 confidently deconfounded nominally significant associations (P < 0.05) in two analyses combined are shown, while drugs are presented in descending order based on the number of associations detected (top 20 drugs shown). The colored cells indicate nominally significant and confidently deconfounded associations (P < 0.05). *FDR ≤ 0.1 and confidently deconfounded. Antibiotics are highlighted in bold.CC BY · PubMed Central
Fig 3 Long-term drug effects on the gut microbiome. (a) Drug carryover effects. The number of univariate associations between subjects not taking the drug and subjects having taken the drug more than 1 to 4 years prior to the microbiome sample collection (CLR-transformed abundance). (b) Long-term effects of antibiotic use on the microbiome. (c) Additive drug effects. The proportion of the univariate associations identified with active drug usage according to the model that best describes the association (CLR-transformed abundance). In addition to active drug usage as a binary trait, active drug usage together with past drug usage as a binary trait or the amount of drug usage as a continuous trait is considered. (d) Variance explained in the gut microbiome composition according to different factor groups. The bars indicating the variance explained by factor groups are further divided into individual factors according to their contribution (individual factors are listed in Table S9). Past drug usage indicates drugs used during the 5 years preceding sample collection at T1. Antibiotics are highlighted in bold.
Fig 3 Long-term drug effects on the gut microbiome. (a) Drug carryover effects. The number of univariate associations between subjects not taking the drug and subjects having taken the drug more than 1 to 4 years prior to the microbiome sample collection (CLR-transformed abundance). (b) Long-term effects of antibiotic use on the microbiome. (c) Additive drug effects. The proportion of the univariate associations identified with active drug usage according to the model that best describes the association (CLR-transformed abundance). In addition to active drug usage as a binary trait, active drug usage together with past drug usage as a binary trait or the amount of drug usage as a continuous trait is considered. (d) Variance explained in the gut microbiome composition according to different factor groups. The bars indicating the variance explained by factor groups are further divided into individual factors according to their contribution (individual factors are listed in Table S9). Past drug usage indicates drugs used during the 5 years preceding sample collection at T1. Antibiotics are highlighted in bold.CC BY · PubMed Central
Fig 4 Drug initiation and discontinuation effects. Associations between drug initiation, discontinuation, and changes in the microbiome between T2 and T1 are shown for macrolides, penicillins, and PPIs. Dark blue refers to associations with active drug usage, light blue refers to associations with drug usage that happened more than 1 year before T2, and green refers to associations with discontinued drug use. Effect size refers to partial Pearson correlation adjusted for BMI, gender, and age.
Fig 4 Drug initiation and discontinuation effects. Associations between drug initiation, discontinuation, and changes in the microbiome between T2 and T1 are shown for macrolides, penicillins, and PPIs. Dark blue refers to associations with active drug usage, light blue refers to associations with drug usage that happened more than 1 year before T2, and green refers to associations with discontinued drug use. Effect size refers to partial Pearson correlation adjusted for BMI, gender, and age.CC BY · PubMed Central

Verification

Claims the paper makes

We couldn’t retrieve this paper’s full text, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Research works printed but not linked: 1

  1. 10.1046/j.1442-9993.2001.01070.x · DOI · Paper page

Cited by

See all 30

Metadata

Authors
Aasmets, O., Taba, N., Krigul, K. L., Andreson, R., Estonian Biobank Research Team, Org, E., Metspalu, M., Metspalu, A., Milani, L., & Esko, T.
Affiliations
University of Tartu (EE)
Corresponding author
  • Aasmets, Oliver
  • Org, Elin
Venue
mSystems
Published
2025-10-22
Volume
10
Issue
10
Pages
e0054125-e0054125
ISSN
2379-5077 (ISSN-L)
OpenAlex
W4414017107
Cited by
30
References
35
Citations by year
2026: 27 · 2025: 3
FWCI
8.96
Type
journal article
Language
en
Open access
Licence
Locations
  • mSystems (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • Eesti Teadusagentuur (PRG1414)
  • European Molecular Biology Organization (Installation grant 3573)
  • Biocodex Microbiota Foundation
  • Estonian Center of Genomics/Roadmap II (16-0125)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gut microbiota and health
  • Health Sciences › Medicine › Gastroenterology › Gastrointestinal motility and disorders
  • Health Sciences › Dentistry › Periodontics › Oral microbiology and periodontitis research
Keywords
gut microbiota, chronic drug administration, confounding, residual effects, additive effects, antibiotics, antidepressants, psycholeptics, benzodiazepines, glucocorticoids, proton pump inhibitors, beta-blockers, electronic health records
MeSH terms
Gastrointestinal Microbiome, Adult, Aged, Anti-Bacterial Agents, Antidepressive Agents, Female, Humans, Male, Middle Aged, Retrospective Studies, Metagenomics, Electronic Health Records, Microbiota

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.