long-term microbiome effects after taking even occasional antibiotics

 a very recent article evaluated three Swedish databases to evaluate antibiotic use and the subsequent development of long-term profound gut microbiome dysbiosis, which can be associated with many human diseases: see antibiotic microbiome change 8 yrs NatureMed2026 in the dropbox, or https://doi.org/10.1038/s41591-026-04284-y

Details:

-- this multicenter confounder-adjusted Swedish study involved 14,979 adults from 3 data sets, with assessment by fecal deep-shotgun metagenomics:

    -- Swedish CArdiopulmonary bioImage Study (SCAPIS, n= 8488), a random sample of residents in areas adjacent to 6 academic hospitals, between 2013 and 2018

        -- mean age 57, 50% female, never smoker 53%/former smoker 34%/current smoker 12%

        -- education: compulsory 9%/upper secondary 44%/university 46%

        -- born in Scandinavia 90%

        -- BMI 26, Charlson index at least 2 (a measure of medical comorbidities) in approx 200 individuals (6%)

        -- medications: metformin 3%, beta-blocker 8%, SSRI 7%, statins 10%, antipsychotics 1%, PPI 11%, taking at least 5 medications 6%

        -- polypharmacy (taking at least 5 meds): 6%

 

    -- Swedish Infrastructure for Medical Population-based Life-course and Environmental Research (SIMPLER, n= 4784), 2 prospective cohorts, one of the Swedish Mammography Cohort initiated in 1987, and the Cohort of Swedish Men initiated in 1997

        -- mean age 72, 46 % female, never smoker 56%/former smoker 41%/current smoker 7%

        -- education: compulsory 40%/upper secondary 30%/university 30%

        -- born in Scandinavia 96%

        -- BMI 26, Charlson index at least 2 (a measure of comorbidities) in approx 200 individuals (12%)

        -- medications: metformin 5%, beta-blocker 23%, SSRI 5%, statins 30%, antipsychotics 0.6%, PPI 15%, taking at least 5 medications 10%

        -- polypharmacy (taking at least 5 meds): 10%

 

    -- Malmo Offspring Study (MOS, n= 1707), adult children and grandchildren of the population-based Malmo Diet and Cancer-Cardiovascular Cohort with recruitment between 2013 and 2021

        -- mean age 40, 50% female, never s mechanism

        -- BMI 25, Charlson index at least 2 (a measure of comorbidities) in approx 13 individuals (2%)

        -- medications: metformin 1%, beta-blocker 4%, SSRI 4%, statins 5%, antipsychotics 0.5%, PPI 8%, taking at least 5 medications 3%

        -- polypharmacy (taking at least 5 meds): 3%

 

-- exclusion criteria included a test center visit before July 2013 (to eliminate those with an incomplete history of antibiotic use in the databases in the prior 8 years), antibiotic prescription in the 30 days before the test center visit, the use of antibiotics to treat acne/rosacea or as prophylaxis for UTI at the time of the fecal sampling, diagnosis of COPD or inflammatory bowel disease

-- the full model of the analysis included BMI, Charlson Comorbidity Index (which includes such conditions as diabetes, cancer, cardiovascular disease, liver disease, and renal disease), polypharmacy (defined as at least five different medications at the time of fecal sampling), and the use of the following medications: proton pump inhibitors, metformin, SSRIs, statins, beta-blockers, and antipsychotics (these meds can affect the microbiome composition and diversity, as well as being associated with the comorbidities noted above)

 

-- participants answered an extensive questionnaire on lifestyle, diet, and health history (the details of these answers was not published in the paper). blood samples were collected and anthropometric measurements were conducted at the test centers. fecal sampling was performed to assess the composition and diversity of the gut microbiome

-- non-antibiotic medication use was derived from the National Prescribed Drug Register (NPDR), the Swedish database of all antibiotics, and prescription medications dispensed to outpatients 

    -- antibiotic exposure: all oral antibiotics in Sweden dispensed to outpatients require prescriptions, are registered through the NPDR, and were classified into groups: tetracyclines, extended spectrum penicillins (amoxicillin and pivmecillinam), beta-lactamase sensitive penicillins, beta-lactamase resistant penicillins (the only one used was flucloxacillin, a narrow-spectrum antibiotic used mostly for skin and soft tissue infections, similar to dicloxacillin), penicillins combined with beta-lactamase inhibitors, cephalosporins, sulfonamides, trimethoprim, macrolide, lincosamides (the only one used was clindamycin), fluoroquinolones, and nitrofurantoin

 

-- antibiotics were divided into 3 periods:

    -- antibiotic use <1 year before the fecal sampling

    -- antibiotic use 1 to 4 years before fecal sampling

    -- antibiotic use from 4 to 8 years before fecal sampling

-- the results were based on the estimated diversity of the gut microbiota (alpha diversity) for each of the additional antibiotic courses within the three periods: <1 year, between1-4 years, and at least 4 years

 

-- outcomes assessed: the effects of different antibiotics on the gut microbiome, both early (<1 yr) and up to 8 years after the antibiotic exposures

Results:

--the estimated decrease in microbiota and diversity was greater after the first two courses of antibiotics than after the third or fourth courses

-- of the 11 classes of antibiotics mentioned above, 6 were associated with lower microbial diversity <1 year before fecal sampling

--antibiotic use <1 year before fecal sampling was associated with the greatest reduction in microbial species diversity, though there were significant associations with antibiotic use after 1-4 and 4-8  years earlier

    -- a healthy microbiome has a significant diversity of largely helpful microbes, and one major characteristic of an unhealthy microbiome is losing microbial diversity: https://www.nature.com/articles/s41522-024-00580-y

    -- and, lower microbiome diversity is associated with several health conditions, including obesity, diabetes, inflammatory bowel disease, etc: see https://pubmed.ncbi.nlm.nih.gov/38486011/, more on this below

-- clindamycin, fluoroquinolones, and flucloxacillin (flucloxacillin is not available in the US, but is very similar to dicloxacillin in being a narrow-spectrum beta-lactamase-resistant penicillin used to treat staphylococcal infections) had the largest effects on microbial diversity

    -- for example, each course of clindamycin <1 yr before fecal sampling was associated with an average of 47 fewer microbial species detected; fluoroquinolones and flucloxacillin were associated respectively with 20 and 21 fewer species detected

-- fluoroquinolones, flucloxacillin, and tetracyclines used 1-4 years and 4-8 years before fecal sampling were associated with this longer term decreased microbial diversity

-- clindamycin and macrolides were associated with decreased diversity 1-4 years but not 4-8 years before fecal sampling

-- but nitrofurantoin had a positive effect on species diversity <1yr before fecal sampling, though these researchers argue that this could be spurious since subsequent analyses lacked confirmation of this finding (not spelled out in detail in the paper)

-- and there were no associations with extended-spectrum penicillins (eg pivmecillinam and amoxicillin), and amoxicillin-clavulanic acid (a broad-spectrum antibiotic with a markedly short-term effect on the gut microbiome), and sulfamethoxazole-trimethoprim.

    -- however, there was some evidence that women had a more negative effect on microbial diversity for amoxicillin-clavulanic acid at all 3 time periods before fecal sampling, with P=0.01

    -- also, there was a difference in those <55yo vs >55yo: for example, clindamycin affected 31 microbial species in the younger group vs 22 in the older one, as well as more profound effects by fluoroquinolone and flucloxacillin

-- in terms of recent antibiotic use, excluding patients taking the antibiotics within the last 30 days, 6 months, or 12 months, before fecal sampling:

    -- no difference for exposures occurring 4-8 and 1-4 years before fecal sampling, though for clindamycin use <1 yr before sampling there was less risk; no difference for the other antibiotics

-- microbiome diversity recovered most rapidly in the first 2 years after antibiotic exposure, though the recovery rate was proportional to the magnitude of the initial reduction in diversity 

-- results for a single course of antibiotics in the past 8 years, a sample size of 3,356 participants:

    -- a single course of tetracyclines, flucloxacillin, fluoroquinolones, clindamycin, sulfamethoxazole-trimethoprim, cephalosporins, or macrolides <4 years or 4-8 years before sampling was associated with lower microbiome species diversity

    -- clindamycin, flucloxacillin and fluroquinolones were the worst <1 year before fecal sampling, which of the 1,340 microbial species assessed, there were respectively 249, 203, and 172 microbial species depleted: very similar numbers when a single course of antibiotics 4-8 years before sampling

        -- as a comparison, penicillin V, the most prescribed antibiotic, had depletion of 29 species

    -- clindamycin and fluoroquinolones were associated with the highest risk of C. difficile infection

-- potential associations with clinical disease (in particular, cardiometabolic diseases), from links with microbial species known to be associated with clinical outcomes (though causality is unclear):

    -- clindamycin, fluoroquinolones, and flucloxacillin were associated with greater abundance of Enterocloster bolteae, E. citroniae (previously named Clostridium bolteae and C. citroniae), Flavonifractor plautii, sRuminococcus B gnavus and Eggerthella lenta

    -- these species in the microbiome have been associated with a higher BMI, serum triglycerides (TG) and risk of type 2 diabetes

    -- assessing the microbial disruptions associated with other conditions:

        -- BMI, waist-to-hip-ratio, serum triglycerides, and CRP: positive association with antibiotic use

        -- colorectal cancer (CRC): species enriched in the microbiome had both positive and negative microbial changes, though one of the major ones associated with CRC, Fusobacterium nucleatum (strongly linked to colorectal cancer) was underrepresented in the microbiomes assessed (<2%), so the conclusions are not so clear

        -- inflammatory bowel disease (IBD): microbial species depleted with IBD were negatively associated with antibiotic use, especially with clindamycin

Commentary:

-- several observational studies have found that recurrent and long-term use of antibiotics is associated with increased risk of obesity, type II diabetes, cardiovascular disease, and colorectal polyps and cancer

    -- the likely mechanism involves antibiotic-related gut microbiome disruptions.

    -- support for this mechanism is that the gut microbiome itself has been found to be associated with many aspects of human health including obesity, cardiometabolic disorders, autoimmune conditions, and colorectal cancer.

        -- a pretty recent blog highlighted several different ways that the microbiome can be related to significant cardiovascular diseases: https://gmodestmedblogs.blogspot.com/2024/07/microbiome-in-cardiovascular-disease.html

    -- an association between antibiotic use and cardiovascular disease in women was found an analysis of the Nurses' Health Study: https://pmc.ncbi.nlm.nih.gov/articles/PMC6911167/pdf/ehz231.pdf, finding that of the 36,429 women studied who were free from cardiovascular disease initially, there was a 32% increased likelihood of CVD after adjusting for multiple covariates if they used antibiotics for at least 2 months, and that the longer the exposure, the higher the CVD risk

    -- overall, the main concern here is that the gut microbiome is remarkably important in regulating many human physiological functions, including brain function (the gut-brain axis: https://gmodestmedblogs.blogspot.com/2017/01/microbiome2-overview.html), the gut-liver axis: https://pubmed.ncbi.nlm.nih.gov/31622696/, a slew of human diseases: https://pmc.ncbi.nlm.nih.gov/articles/PMC4838534/, and, of particular significance in the current study, antibiotic-resistant genes: https://www.nature.com/articles/s41467-023-42998-6

-- healthy human volunteers have drastic alterations in their gut microbiota a few days after course of oral antibiotics, particularly in decreases in species diversity, but as well with microbial gene richness

    -- studies have found that short-term microbiome alterations are associated with increases in potential pathogens such as Escherichia coli; reduces in healthy genera Dialister, Veillonella and Eubacterium; enrichment of antimicrobial-resistance genes; and increased risk of Clostridium difficile infection

    -- this current study was designed to look at long-term effects of antibiotics on gut microbiome alterations

-- lots of numbers above. the overall synthesis of their results:

    -- the strongest association between antibiotic use and the adverse gut microbiome changes was in those taking antibiotics <1y before fecal sampling

        -- but antibiotics used 1-4 years and 4-8 years before sampling had lower microbiome diversity and abundance of species

        -- the associations were most powerfully found in those on clindamycin, flucloxacillin, and fluoroquinolones; some antibiotics had no detectable effect (amoxacillin, perhaps nitrofurantoin)

            -- nitrofurantoin use in the first year before fecal sampling was associated with a positive association with species diversity. the researchers note that there was a lack of confirmation in subsequent analyses on this study is possibly due to chance; however other studies have found minimal gut microbiome effects https://pmc.ncbi.nlm.nih.gov/articles/PMC7507860/. this may be related to nitrofurantoin's rapid intestinal absorption and rapid direct effect on the urinary tract

-- there were no associations with microbial species diversity with extended-spectrum penicillins (pivmecillinam and amoxicillin), amoxicillin-clavulanic acid (the latter being a broad-spectrum antibiotic but has a marked short-term effect on the gut microbiome) or sulfamethoxazole- trimethoprim

        -- there were some differences by sex and age: women and younger people had more profound microbiome changes

    -- a single course of antibiotics taken 4-8 years before sampling was associated with effects of gut microbiome diversity and abundance of certain microbial species

-- this all means that the effect of antibiotics on the gut microbiome may persist for many years (a partial recovery often occurs within weeks, but a full recovery may take at least 8 years, the end of this study)

-- it should be noted that there are several nonantibiotic meds that affect the microbiome (see microbiome effect of nonantibiotic drugs Nature2018 in dropbox, or doi:10.1038/nature25979):

    -- in this study, the researchers screened >1000 meds against 40 representative ubiquitous bacterial strains in the gut microbiome, finding that 24% of the drugs (all with dosages in the range of what would be expected in the human gut), including members of all therapeutic classes, inhibited the growth of at least one bacterial strain in vitro; most drugs affected a few bacterial strains, but 40 drugs affected at least 10 strains

        -- more than ½ of antivirals had antibacterial activity, as do many antifungal imidazoles

        -- others meds that affect the microbiome included:

            -- the ovulation stimulant clomiphene

            -- several proton-pump inhibitors

            -- atypical anti-psychotics (these decreased Akkermansia species; notably metformin stimulates the growth of Akkermansia (and that stimulation seems to be a major mechanism in improving diabetes control:  (see dm metformin microbiome GUT 2014 in dropbox, or doi:10.1136/gutjnl-2012-303839; also summarized in https://gmodestmedblogs.blogspot.com/2024/07/microbiome-in-cardiovascular-disease.html ,which also summarizes a study finding that red meat (or the amino acid carnitine) elaborate TMAO, a powerful atherogen). perhaps some of the adverse effects of the atypical antipsychotics, including weight gain, insulin resistance, metabolic syndrome, and diabetes are related to this Akkermansia depletion

            --immunosuppressives (eg azathioprine), antidepressants (venlafaxine), anti-inflammatories (mesalazine, aminosalicylate), progesterone, estrogen, phenothiazines, calcium-channel blockers

        --of note, they did check on reported adverse effects of the meds and found that many had antibiotic-like adverse effects

 limitations:

-- though the researchers did exclude patients who had received antibiotic prescriptions in the 30 days before the start of the study, the 30-day mark seems arbitrary since there were much more prolonged effects on the microbiome with earlier antibiotic use. it would have been interesting to document this microbiome effect sooner after the antibiotics

-- the use of data from Sweden may limit generalization to other countries:

    -- there are some different antibiotics used there than in the US and likely other countries

    -- they have a restrictive use of antibiotics, unlike most other countries

    -- Sweden has an overall low incidence of antimicrobial resistance vs most other countries (likely related to their restricted use)

    -- the patients in these databases overall were not very sick (low percentages had Charlson Index of at least 2, meaning not so many medical comorbidities), few were on common medications, few taking more than 5 meds, and a high level of university education (?perhaps more medically literate than in less educated populations, and perhaps more attuned to healthy behaviors, eg better diets, more exercise, less alcohol, all or which decrease microbiome dysbiosis; and these issues were not elaborated in the article)

-- we have no information on the antibiotic doses and durations of treatment; also the study inclusion was limited to the 12 months preceding fecal sampling

-- the Charlson Comorbidity Index is derived from hospital records and may not reflect outpatients

-- we have no information on the large array of non-antibiotic meds that can affect the microbiomes, as noted above

-- there are major ways that antibiotics can enter human systems besides taking pills: antibiotic exposures could be related to foods we eat (meats/veges/fruits exposed to and accumulating antibiotics. perhaps the role of pesticides?? perhaps the role of high usage of antibiotics in livestock to increase their weight/value. Sweden does also have very strict policies on antibiotic use in livestock, for example, that might well limit generalization of their results to other areas

so, the importance of this study includes:

    -- reinforcing the essential role of the gut microbiome in clinical diseases, including many of the most common diseases associated with significant morbidity and mortality

    -- reinforcing the importance of minimizing antibiotic use in humans and animals, both of which lead to adverse human microbiome disruptions

        -- this study broadens the issue that the use of antibiotics goes beyond the well-publicized risk of developing antibiotic resistance and highlights the effects of even remote antibiotic usage on their long-term effects on microbiome function related to abnormal microbe diversity and the likely associated adverse multiple clinical effects

    -- reinforcing the use of the least microbiome-toxic meds and for the shortest time possible

        -- for example, there is a strong argument that we can use shorter antibiotic courses in many situations where antibiotics are appropriate: https://gmodestmedblogs.blogspot.com/2017/07/take-full-course-of-antibiotics.html

        -- so, this study and others noted above reinforce the use of nitrofurantoin for simple urinary tract infections over using trimethoprim/sulfamethoxazole (though this is better than many of the other antibiotics) or fluroquinolones (though upper tract disease and urosepsis does require these latter two antibiotics since there is low concentration of nitrofurantoin in the kidney or blood stream)

    -- the importance of reinforcing activities that are restorative to the microbiome:

        -- healthy diet and exercise are pivotal. also perhaps probiotics. other interventions to be determined (see prior blogs https://gmodestmedblogs.blogspot.com/2026/05/antibiotic-resistance-genes-in-newborns.html, https://gmodestmedblogs.blogspot.com/2019/01/antibiotic-overprescribing-2-more.html , https://gmodestmedblogs.blogspot.com/2023/01/antibiotic-resistant-gonorrhea-in-us.html for more info/references on antibiotic overprescribing, microbiome effects, effective antibiotic stewardship programs to decrease unnecessary antibiic prescribing)

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

to get access to all of the blogs:  go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

to get access to the dropbox: https://www.dropbox.com/scl/fo/pcf4vymtpo77f49wd2ljr/ANyJqB40CuBd-DMEr67XDoI?rlkey=5dsalmuhf8ekd0svsgvn93xm1&dl=0

or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them