Antibiotic use and gut microbiome composition links from individual-level prescription data of 14,979 individuals

Gabriel Baldanzi; Anna Larsson; Sergi Sayols-Baixeras; Koen F. Dekkers; Ulf Hammar; Diem Nguyen; Tíscar Graells; Shafqat Ahmad; Camila Gazolla Volpiano; Guillaume Meric; Josef D. Järhult; Thomas Tängdén; Jonas F. Ludvigsson; Lars Lind; Johan Sundström; Karl Michaëlsson; Johan Ärnlöv; Beatrice Kennedy; Marju Orho-Melander; Tove Fall

doi:10.1038/s41591-026-04284-y

Antibiotic use and gut microbiome composition links from individual-level prescription data of 14,979 individuals

Gabriel Baldanzi, Anna Larsson, Sergi Sayols-Baixeras, Koen F. Dekkers, Ulf Hammar, Diem Nguyen, Tíscar Graells, Shafqat Ahmad, Camila Gazolla Volpiano, Guillaume Meric, Josef D. Järhult, Thomas Tängdén, Jonas F. Ludvigsson, Lars Lind, Johan Sundström, Karl Michaëlsson, Johan Ärnlöv, Beatrice Kennedy, Marju Orho-Melander, Tove Fall

Research output: Contribution to journal › Article › peer-review

4 Citations (SciVal)

Abstract

Disruptions in gut microbiome are implicated in cardiometabolic disorders and other health outcomes. Antibiotics are known gut microbiome disruptors, but their long-term consequences remain underexplored. Here we combined individual-level data from the Swedish Prescribed Drug Register with fecal metagenomes of 14,979 adults to examine the association between oral antibiotic use over 8 years and gut microbiome. In multivariable confounder-adjusted regression models, antibiotic use <1 year before fecal sampling was associated with the greatest reduction in species diversity, but significant associations were also observed for use 1–4 and 4–8 years earlier. Clindamycin, fluoroquinolones and flucloxacillin accounted for most of the associations with the abundance of individual species. Use of these antibiotics 4–8 years earlier was associated with altered abundance of 10–15% of the species studied; penicillin V, extended-spectrum penicillins and nitrofurantoin were associated with only a few species. Similar results were found comparing one antibiotic course 4–8 years before sampling versus none in the past 8 years. These findings indicate that antibiotics may have long-lasting consequences for the gut microbiome.

Original language	English
Pages (from-to)	1351-1361
Number of pages	11
Journal	Nature Medicine
Volume	32
Issue number	4
Early online date	11 Mar 2026
DOIs	https://doi.org/10.1038/s41591-026-04284-y
Publication status	Published - 11 Mar 2026

Data Availability Statement

The data supporting the conclusions of this article were provided by the SCAPIS, SIMPLER and MOS data offices and contain sensitive personal information protected under privacy laws; therefore, they are not publicly available. Requests for access to data to verify the analyses and findings of this study should be directed to the corresponding author. An initial response to the request will be provided within two weeks. Data will be shared once a data-sharing agreement has been signed between Uppsala University and the requestor’s institution, and following approval from the Swedish Ethical Review Authority (https://etikprovningsmyndigheten.se) and the boards of SCAPIS, SIMPLER and MOS. Requests for data access for additional research purposes should be directed to the respective cohorts: SCAPIS (https://www.scapis.org/data-access/), SIMPLER (https://www.simpler4health.se/w/sh/en/researchers/data-access) and MOS (https://www.malmo-cohorts.lu.se/application-data-and-samples/applying-samples-mdc-and-mpp). Dehosted anonymized metagenomic sequencing data from SCAPIS are available in the European Nucleotide Archive under accession number PRJEB51353.

Acknowledgements

We would like to acknowledge the help of Biobank Sweden and the local biobank facilities for their services in handling biological samples and biobanking. We acknowledge SIMPLER for the data access, and we would like to thank A.-K. Kolseth and N. Håkansson for assistance. SIMPLER receives funding through the Swedish Research Council under grant nos. 2017-00644, 2017-06100, 2021-00160 and Stiftelsen Olle Engkvist Byggmästare. The computations and data handling were enabled by resources in project sens2019512 and simp2023007 provided by the National Academic Infrastructure for Supercomputing in Sweden (NAISS) at Uppsala Multidisciplinary Center for Advanced Computational Science (UPPMAX), funded by the Swedish Research Council through grant no. 2022-06725.

Funding

Financial support was obtained in the form of grants from the European Research Council (grant no. ERC-STG-2018-801965 (T.F.); grant no. ERC-CoG-2014-649021 (M.O.-M.)), the Swedish Heart-Lung Foundation (Hjärt-Lungfonden, grant no. 2023-0687 (T.F.); grant no. 2018-0343 (J.Ä.); grant no. 2023-0380 (M.O.-M.)), the Swedish Research Council (VR, grant no. 2019-01471 (T.F.), grant no. 2025-02673 (T.F.), grant no. 2021-03291 (M.O.-M.), grant no. 2019-01015 (J.Ä.), grant no. 2020-00243 (J.Ä.), grant no. 2022-01460 (S.A.) and Strategic Research Area Exodiab grant no. 2009-1039 (supporting M.O.-M.)), the Swedish Foundation for Strategic Research, LUDC-IRS, grant no. IRC-0067 (supporting M.O.-M), the Swedish Research Council for Sustainable Development (FORMAS, grant no. 2020-00989 (S.A.)), the A.L.F. governmental grant (grant no. 2018-0148 (M.O.-M.), grant no. 2022-0258 (M.O.-M.)), the Novo Nordisk Foundation (grant no. NNF20OC0063886 (M.O.-M.), grant no. NNF23OC0084419 (M.O.-M.)), the Swedish Diabetes Foundation (grant no. DIA 2018-375 (M.O.-M.), grant no. DIA 2024-928 (M.O.-M.)), Center of Clinical Research in Region Dalarna (J.Ä.), Epihealth (S.A.), the Göran Gustafsson Foundation for Research in Natural Sciences and Medicine (T.F.), the Elsa Lundberg and Greta Fleron Foundation (A.L.) and Ernhold Lundström Foundation (A.L.). B.K. is supported by a Gullstrand fellow grant from the Uppsala University Hospital. This research has been conducted using the SCAPIS Resource, under Petition Number 514. The main funding body of the Swedish CArdioPulmonary bioImage Study (SCAPIS) is the Swedish Heart and Lung Foundation. SCAPIS is also funded by the Knut and Alice Wallenberg Foundation, the Swedish Research Council, VINNOVA (Sweden’s Innovation agency), the University of Gothenburg and Sahlgrenska University Hospital, Karolinska Institutet and Region Stockholm, Linköping University and University Hospital, Lund University and Skåne University Hospital, Umeå University and University Hospital, Uppsala University and University Hospital. We would like to acknowledge the help of Biobank Sweden and the local biobank facilities for their services in handling biological samples and biobanking. We acknowledge SIMPLER for the data access, and we would like to thank A.-K. Kolseth and N. Håkansson for assistance. SIMPLER receives funding through the Swedish Research Council under grant nos. 2017-00644, 2017-06100, 2021-00160 and Stiftelsen Olle Engkvist Byggmästare. The computations and data handling were enabled by resources in project sens2019512 and simp2023007 provided by the National Academic Infrastructure for Supercomputing in Sweden (NAISS) at Uppsala Multidisciplinary Center for Advanced Computational Science (UPPMAX), funded by the Swedish Research Council through grant no. 2022-06725.

Funders	Funder number
Linköpings Universitet
Stiftelsen för Strategisk Forskning
Sahlgrenska Universitetssjukhuset
Knut och Alice Wallenbergs Stiftelse
Akademiska Sjukhuset
Karolinska Institutet
Region Stockholm
Lunds Universitet
VINNOVA
Skånes universitetssjukhus
Göteborgs Universitet
Umeå Universitet
Svenska Forskningsrådet Formas	2022-0258, 2018-0148, 2020-00989
Uppsala Universitet	2021-00160, 2017-06100, 2017-00644
Strategic Research Area Exodiab	2009-1039
Diabetesförbundet	DIA 2024-928, DIA 2018-375
European Research Council	ERC-STG-2018-801965, ERC-CoG-2014-649021
Novo Nordisk Fonden	NNF20OC0063886, NNF23OC0084419
Olle Engkvists Stiftelse	2022-06725, sens2019512
Hjärt-Lungfonden	2023-0687, 2023-0380, 2018-0343
Vetenskapsrådet	2025-02673, 2019-01015, 2020-00243, 2021-03291, 2019-01471, 2022-01460
LUDC-IRS	IRC-0067

ASJC Scopus subject areas

General Medicine
General Biochemistry,Genetics and Molecular Biology

Access to Document

10.1038/s41591-026-04284-yLicence: CC BY

https://www.nature.com/articles/s41591-026-04284-yLicence: CC BY

Cite this

Baldanzi, G., Larsson, A., Sayols-Baixeras, S., Dekkers, K. F., Hammar, U., Nguyen, D., Graells, T., Ahmad, S., Gazolla Volpiano, C., Meric, G., Järhult, J. D., Tängdén, T., Ludvigsson, J. F., Lind, L., Sundström, J., Michaëlsson, K., Ärnlöv, J., Kennedy, B., Orho-Melander, M., & Fall, T. (2026). Antibiotic use and gut microbiome composition links from individual-level prescription data of 14,979 individuals. Nature Medicine, 32(4), 1351-1361. https://doi.org/10.1038/s41591-026-04284-y

Baldanzi, G, Larsson, A, Sayols-Baixeras, S, Dekkers, KF, Hammar, U, Nguyen, D, Graells, T, Ahmad, S, Gazolla Volpiano, C, Meric, G, Järhult, JD, Tängdén, T, Ludvigsson, JF, Lind, L, Sundström, J, Michaëlsson, K, Ärnlöv, J, Kennedy, B, Orho-Melander, M & Fall, T 2026, 'Antibiotic use and gut microbiome composition links from individual-level prescription data of 14,979 individuals', Nature Medicine, vol. 32, no. 4, pp. 1351-1361. https://doi.org/10.1038/s41591-026-04284-y

@article{c168d9e8ecae4db2a4eefb3643ca9621,

title = "Antibiotic use and gut microbiome composition links from individual-level prescription data of 14,979 individuals",

abstract = "Disruptions in gut microbiome are implicated in cardiometabolic disorders and other health outcomes. Antibiotics are known gut microbiome disruptors, but their long-term consequences remain underexplored. Here we combined individual-level data from the Swedish Prescribed Drug Register with fecal metagenomes of 14,979 adults to examine the association between oral antibiotic use over 8 years and gut microbiome. In multivariable confounder-adjusted regression models, antibiotic use <1 year before fecal sampling was associated with the greatest reduction in species diversity, but significant associations were also observed for use 1–4 and 4–8 years earlier. Clindamycin, fluoroquinolones and flucloxacillin accounted for most of the associations with the abundance of individual species. Use of these antibiotics 4–8 years earlier was associated with altered abundance of 10–15\% of the species studied; penicillin V, extended-spectrum penicillins and nitrofurantoin were associated with only a few species. Similar results were found comparing one antibiotic course 4–8 years before sampling versus none in the past 8 years. These findings indicate that antibiotics may have long-lasting consequences for the gut microbiome.",

author = "Gabriel Baldanzi and Anna Larsson and Sergi Sayols-Baixeras and Dekkers, \{Koen F.\} and Ulf Hammar and Diem Nguyen and T{\'i}scar Graells and Shafqat Ahmad and \{Gazolla Volpiano\}, Camila and Guillaume Meric and J{\"a}rhult, \{Josef D.\} and Thomas T{\"a}ngd{\'e}n and Ludvigsson, \{Jonas F.\} and Lars Lind and Johan Sundstr{\"o}m and Karl Micha{\"e}lsson and Johan {\"A}rnl{\"o}v and Beatrice Kennedy and Marju Orho-Melander and Tove Fall",

year = "2026",

month = mar,

day = "11",

doi = "10.1038/s41591-026-04284-y",

language = "English",

volume = "32",

pages = "1351--1361",

journal = "Nature Medicine",

issn = "1078-8956",

publisher = "Nature Research",

number = "4",

}

TY - JOUR

T1 - Antibiotic use and gut microbiome composition links from individual-level prescription data of 14,979 individuals

AU - Baldanzi, Gabriel

AU - Larsson, Anna

AU - Sayols-Baixeras, Sergi

AU - Dekkers, Koen F.

AU - Hammar, Ulf

AU - Nguyen, Diem

AU - Graells, Tíscar

AU - Ahmad, Shafqat

AU - Gazolla Volpiano, Camila

AU - Meric, Guillaume

AU - Järhult, Josef D.

AU - Tängdén, Thomas

AU - Ludvigsson, Jonas F.

AU - Lind, Lars

AU - Sundström, Johan

AU - Michaëlsson, Karl

AU - Ärnlöv, Johan

AU - Kennedy, Beatrice

AU - Orho-Melander, Marju

AU - Fall, Tove

PY - 2026/3/11

Y1 - 2026/3/11

N2 - Disruptions in gut microbiome are implicated in cardiometabolic disorders and other health outcomes. Antibiotics are known gut microbiome disruptors, but their long-term consequences remain underexplored. Here we combined individual-level data from the Swedish Prescribed Drug Register with fecal metagenomes of 14,979 adults to examine the association between oral antibiotic use over 8 years and gut microbiome. In multivariable confounder-adjusted regression models, antibiotic use <1 year before fecal sampling was associated with the greatest reduction in species diversity, but significant associations were also observed for use 1–4 and 4–8 years earlier. Clindamycin, fluoroquinolones and flucloxacillin accounted for most of the associations with the abundance of individual species. Use of these antibiotics 4–8 years earlier was associated with altered abundance of 10–15% of the species studied; penicillin V, extended-spectrum penicillins and nitrofurantoin were associated with only a few species. Similar results were found comparing one antibiotic course 4–8 years before sampling versus none in the past 8 years. These findings indicate that antibiotics may have long-lasting consequences for the gut microbiome.

AB - Disruptions in gut microbiome are implicated in cardiometabolic disorders and other health outcomes. Antibiotics are known gut microbiome disruptors, but their long-term consequences remain underexplored. Here we combined individual-level data from the Swedish Prescribed Drug Register with fecal metagenomes of 14,979 adults to examine the association between oral antibiotic use over 8 years and gut microbiome. In multivariable confounder-adjusted regression models, antibiotic use <1 year before fecal sampling was associated with the greatest reduction in species diversity, but significant associations were also observed for use 1–4 and 4–8 years earlier. Clindamycin, fluoroquinolones and flucloxacillin accounted for most of the associations with the abundance of individual species. Use of these antibiotics 4–8 years earlier was associated with altered abundance of 10–15% of the species studied; penicillin V, extended-spectrum penicillins and nitrofurantoin were associated with only a few species. Similar results were found comparing one antibiotic course 4–8 years before sampling versus none in the past 8 years. These findings indicate that antibiotics may have long-lasting consequences for the gut microbiome.

UR - https://www.scopus.com/pages/publications/105033295683

U2 - 10.1038/s41591-026-04284-y

DO - 10.1038/s41591-026-04284-y

M3 - Article

SN - 1078-8956

VL - 32

SP - 1351

EP - 1361

JO - Nature Medicine

JF - Nature Medicine

IS - 4

ER -

Antibiotic use and gut microbiome composition links from individual-level prescription data of 14,979 individuals

Abstract

Data Availability Statement

Acknowledgements

Funding

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this