Worldwide, antibiotics continue to save millions of lives, relieve suffering, and prevent long-term illness (1). Large quantities of these drugs are involved: for example, over 250 million prescriptions of antibiotics were issued in the USA alone in 2016 (2). One can reasonably hypothesize that worldwide, most people (especially children) are treated with an antibiotic at least once a year (3–6). Antibiotics are extremely effective and generally lead to the eradication of the targeted pathogenic bacteria (7). By definition, antibiotics create dysbiosis [defined in several ways (8)] by killing significant components of the gut community (9–11) (Figure 1). Broadly, levels of Enterobacteriaceae, Bacteroidaceae, enterococci, and drug-resistant Escherichia coli rise following antibiotic treatment in adults, whereas levels of bifidobacteria, lactobacteria, actinobacteria and Lachnospiraceae decrease (11–13). Despite the emergence of microbial resistance as a long-term public health risk, antibiotics are still irreplaceable in the treatment of bacterial infections (14).

However, antibiotics also have negative effects on health. The acute (short-term) effects have been extensively studied; for example, between 5 and 20% of antibiotic users (depending on the population studied) will develop antibiotic-associated diarrhea with days or weeks of treatment initiation, and the incidence is greatest in frail, hospitalized patients and young children (11, 15–19). There is now a large body of epidemiological evidence to show that exposure to antibiotics has chronic (long-term) negative effects on human health—effects likely to be accentuated by poor antibiotic stewardship, inappropriate prescribing, excessive or chronic administration, and off-label use.

The present review covers the literature data on the long-term negative health effects of antibiotic exposure, the link to dysbiosis, and how the associated risks might be managed through the evidence-based administration of specific probiotics. Although this review was not systematic, we searched the PubMed database for recent publications (from January 1st, 2023, to June 15th, 2024) using logical combinations of the following keywords: antibio*, exposure, microbiot*, microb*, gut, intestine*, dysbiosis, DOHaD, thousand days, health, disease, and probiotic.

From the mouth to the anus, the adult human gastrointestinal tract has a surface area of around 30 m² and a luminal volume of around 3 L (20). After the skin, the gastrointestinal tract constitutes the body’s largest interface with the environment. The gut carries 10¹³–10¹⁴ microbes from thousands of species, which contain several million genes in total—far more than in the human genome. Here, we shall use the term “gut microbiota” to refer to the set of microbial (mainly bacterial) species contained in the environment of the human gastrointestinal tract. We prefer “microbiota” to the term “microbiome,” which we take to encompass not only the microorganisms (bacteria, archaea, and lower and higher eukaryotes) and their genomes but also the human cells and substances in the surrounding gut (21). Although there are huge inter- and intra-individual variations in the composition of the gut microbiota, over 99% of the gut microbiota is composed of species from the Firmicutes, Bacteroidetes, Proteobacteria, and Actinobacteria (22).

It is generally accepted that a human’s gut microbiota starts to form after birth (i.e., ex utero) and reaches maturity at around the age of 3 years (23, 24). Interesting, this period coincides largely with the “first 1,000 days” (from conception to the age of 2 years)? The “first 1,000 days” concept grew out of a body of research on the importance of the early life environment for the child’s current and future mental, physical and emotional health states (25–30). Awareness of this concept grew markedly after a keynote speech by the then US Vice President Hilary Clinton at the “1,000 Days: Change a Life, Change the Future” international conference on global child undernutrition in 2010. Furthermore, the “first 1,000 days” concept fits well with the “thrifty phenotype” hypothesis, the Barker hypothesis (linking adverse nutrition childhood to metabolic syndromes in adulthood) and the subsequent “developmental origins of health and disease” (DOHaD) concept; all hold that a poor fetal environment and perturbed early neonatal life prompt the development of disease (31–33). Of course, a focus on the first 1,000 days does not mean that day 1,001 is medically and scientifically unimportant. Nevertheless, the “first 1,000 days” concept has been useful for (i) focusing research efforts and public opinion on this critical period in the development of the child and, indeed, of the microbiota, and (ii) emphasizing the importance of the microbiota’s origin and development in the first hours, days, weeks, months and then years of the human host’s life.

As mentioned above, most experts consider that the human gut microbiota starts to form after birth. However, there is some debate as to the presence of a fetal microbiota, and it is not inconceivable that bacteria from the mother can cross the maternal-fetal barrier (34). However, the ethical and practical difficulties of collecting contamination-free samples of fetal tissue have complicated research efforts to settle this debate (35–39). Delivery by cesarean section (i.e., the avoidance of contact with the vaginal microbiota) is a major dysbiosis-promoting factor and is associated with a low-Bacteroides profile in the first 6 months postpartum (40). In children delivered by cesarean section, relatively high abundances of Burkholderiaceae, Bacteroidales, and Ruminococcaceae persist for at least 7 years (41, 42). It is also noteworthy that the gut microbiota is much less diverse in preterm infants (born after less than 34 weeks of gestation) than in term infants—even in the absence of antibiotic treatment, which most preterm infants nevertheless receive. Cetinbas et al. used 16S rDNA sequencing to evaluate long-term, antibiotic-induced dysbiosis on the basis of 363 stool samples collected between 26 and 48 weeks of adjusted age from 65 preterm infants treated with antibiotics in the NICU and 52 samples from 14 preterm counterparts not treated with antibiotics (43). Antibiotic-treated preterm infants were slower to return to a “normal” gut microbiota, in terms of Shannon diversity and species richness. The difference in the abundance of Paenibacillus amylolyticus was quickest to disappear, whereas changes in Veillonella, Shuttleworthia and Clostridioides persisted up to 40 weeks of age (43).

During the potentially unstable first years of life, the gut microbiota can be perturbed by pathogenic microbes and external (environmental) factors, such as diet, xenobiotics, and other exogenous compounds—notably antibiotics (44). In-depth molecular biology studies have shown very clearly that the systemic administration of an antibiotic to infants has a rapid but often long-lasting impact. For example, Yassour et al.’s study of 1,069 samples of feces from 39 children (around half of whom had been treated with antibiotics in the first year of life) collected over a period of 36 months showed that antibiotic-treated individuals had less stable gut microbiotas, with low diversity still visible at the age of 3 years (40). Kwon et al. used 16S rRNA sequencing and linear discriminant effect size analysis to evaluate alpha and beta diversities in fecal samples from (i) 20 infants under 3 months of age who had received antibiotics for at least 3 days and (ii) 34 age-matched, healthy controls not exposed to antibiotics (45). Relative to controls, the relative abundances of Escherichia, Shigella and Bifidobacterium were significantly greater in the antibiotic-treated group, whereas that of Bacteroides was significantly lower. The abundances of Firmicutes genera (Allobaculum, Enterococcus, and Candidatus arthromitus), Proteobacteria genera (Klebsiella), and Actinobacteria genera (Bifidobacterium) were three or four times higher in the treated group than in the control group. A phylogenetic investigation of communities by reconstruction of unobserved states revealed a significant difference in gut microbiome metabolic activity between the two groups; the antibiotic-treated group showed the expression of significantly more genes involved in naphthalene degradation, glycolysis gluconeogenesis, and lipoic acid metabolism and less expression of genes involved in porphyrin metabolism and fatty acid biosynthesis (45).

The effects of antibiotics on the gut microbiota of newborns and infants are accentuated by a lack of standardization in antibiotic treatment plans. For example, Schulman et al.’s study of 127 neonatal intensive care units (NICUs) in the USA evidenced up to 40-fold variations in dosing regimens (46). In France, Leroux et al. study of 44 NICUs found an average of nine different dosing regimens for each of the 41 antibiotics documented (47). Similarly, in 43 NICUs in the UK, Kadambari et al. identified 10 different dosing regimens for prescriptions of gentamicin (48).

Although many experts have sought to define dysbiosis, the topic is subject to debate and constitutes a field of research in its own right (49). As emphasized by Brüssow and by Drago et al., no consensus conference has yet worked out a definition of “dysbiosis” (50, 51). To the best of our knowledge, none of the learned societies or medical associations in this field have issued a single, unambiguous definition of “dysbiosis.” Given this lack, there is no “gold standard” approach to identifying a dysbiotic state (52). A conventional, broad definition published by Petersen and Round in 2014 is related primarily to the composition of the microbiota: “any change to the composition of resident commensal communities relative to the community found in healthy individuals” (13). These compositional changes usually encompass a decrease in bacterial diversity, the loss of beneficial microbes, and overgrowth by potential pathogens. More recent definitions have emphasized function as well as composition: for example, Levy et al. define dysbiosis as “a compositional and functional alteration in the microbiota that is driven by a set of environmental and host-related factors that perturb the microbial ecosystem to an extent that exceeds its resistance and resilience capabilities” (53). In this definition, gut microbiota can be viewed as a conceptual energy landscape in which both healthy and dysbiotic states can exist in energy minima. External factors (such as antibiotics, various xenobiotics, dietary components, and pathogens) that exceed a certain threshold cause transitions from a stable, healthy state to a metastable, dysbiotic state (54). Levy et al. acknowledged that given the high degree of interindividual variability, it is impossible to define a single “healthy” microbiota; however, they suggest that the main characteristic of a “healthy” microbiota in a given individual is richness and stability over time (53). Wilkins et al. have suggested that dysbiosis should be defined by reference to a particular disease state, i.e., a single, all-encompassing definition may not be feasible (49). Lastly, Malard et al. have defined dysbiosis as the disruption of host-microbe symbiosis and crosstalk, with auto-aggravating signals from both the host and microbes that maintain the metastable dysbiotic state (54). Hence, as detailed in the following section, dysbiosis of the gut microbiota can be viewed as the disruption of the positive actions of (i) short-chain fatty acids (SCFAs, produced by the fermentation of host-enzyme-resistant carbohydrates by obligate anaerobes) (55, 56), (ii) the microbial production of serotonin, catecholamines and other neurotransmitters (affecting gut peristaltic motility and the genesis of the enteric nervous system) (57–61), (iii) minimization of the “leaky gut” in which potentially pathogenic whole bacteria and bacterial wall components can enter the circulation (62–64), and (iv) the bacterial expression of enzymes that influence the levels of host metabolites (65, 66). In summary, dysbiosis does not have a consensus definition but can usefully be viewed as the disruption of a previously stable, symbiotic, functionally complete microbiota by dietary, xenobiotic or other factors.

It is now clear that a stable, functionally rich gut microbiota exerts complex direct and indirect effects on the human host’s health through multiple genetic, immune-mediated and metabolic factors. Conversely, dysbiosis of the gut microbiota (i.e., the disruption of a previously stable, functionally complete microbiota) has many harmful short- and long-term effects on health. Common signs of gut dysbiosis include diarrhea, gas, constipation, nausea, and even chest pain (67). Admittedly, in most settings, it is difficult to determine whether dysbiosis is the cause or result of disease. Determining the causal nature and direction of a relationship notably requires prospective, longitudinal studies to see whether the dysbiosis or the disease occurs first. However, this approach can be confounded if a patient remains free of symptoms for months or years. A causal role of antibiotic-associated gut dysbiosis can nevertheless be suspected when the disease is more intense after exposure to broad-spectrum antibiotics (compared with narrow-spectrum antibiotics) and shows an antibiotic dose dependence (68–70). Lastly, one cannot rule out direct, negative effects (i.e., not mediated by the gut microbiota) of antibiotics. Some classes of antibiotic affect myocytes and neurons directly; for example, aminoglycosides, capreomycin, and macrolides are ototoxic (71, 72). Below, we describe the main pathways through which the gut microbiota are known to influence human health.

Many chronic diseases are known to have both genetic factors and environmental susceptibility/triggering factors, of which antibiotic exposure is only one. Many of these environmental factors are linked to a “modern,” “post-industrial” or “Western” lifestyle, including the “Western diet” (high in sugar, fat, and processed food components, and low in fiber) and sedentary behavior (typically defined as spending time in a sitting, reclining or lying posture with an energy expenditure of 1.5 metabolic equivalents of task or less) (89–91). Sedentary behavior (including total sitting time and TV viewing time) is associated with an elevated risk of chronic diseases, including diabetes, cardiovascular disease, and (to a lesser extent) cancer—even after adjustment for the level of physical activity (92). Interestingly, lower sedentary behavior through increased physical activity is known to influence gut microbiota diversity in adults and children, with higher alpha-diversity (notably including SCFA-producing Lachnospiraceae and Erysipelotrichaceae families and Akkermansia, Roseburia, and Veillonella, genera) in physically active individuals (93–96). Indeed, SCFAs appear to be the main molecular link between physical activity and the gut microbiome, although few published studies of physically active vs. sedentary populations controlled for dietary confounders (i.e., fiber intake) (97). The high-sugar, high-fat “Western diet” is clearly associated with dysbiosis, as characterized by a decrease in the abundance of Bacteroidetes and bifidobacteria and an increase in organisms that can utilize excess monosaccharides, such as Enterobacteriaceae and Proteobacteria (98–103). Again, the main molecular link between diet, a normal microbiota and the associated health effects on the host is likely to be the SCFAs (104).

Antibiotic-induced perturbation of the gut microbiome nevertheless constitutes a key “environmental” factor in the development of chronic disease (Figure 2). The prevalent use of antibiotics in infants reveals concerning associations between antibiotic exposure and the onset of a number of distinct immunological, metabolic and neurobehavioral health conditions (whether isolated or combined) during childhood. For example, in a time-to-event analysis of medical records for 14,572 children born in Olmsted County (MI, United States) between January 1, 2003, and December 31, 2011, Aversa et al. found that antibiotic exposure in the first 2 years of life (i.e., during the first 1,000 days) was associated with asthma, allergic, rhinitis, atopic dermatitis, celiac disease, overweight, obesity, attention deficit hyperactivity disorder, and learning disability (105). These health risks were influenced by the number, type, and timing of the antibiotic prescriptions. Relationships between antibiotics, dysbiosis, and a number of chronic diseases are described below.

The quotation “all disease begins in the gut” is often attributed to Hippocrates circa 400 BC. The father of modern medicine was probably not fully correct but, as seen for the diseases reviewed above, the DOHaD concept can be logically extended to what we term the “gastrointestinal origins of health and disease” (“GOHaD”) or even the “microbiotic origins of health and disease” (“MOHaD”). It should nevertheless be borne in mind that the ORs for the associations between antibiotic exposure and the onset of chronic disease are generally quite low (i.e., between 1 and 1.5) and, despite the investigators’ best efforts in study design and data analysis, may be influenced by confounding factors.

More generally, we found that most of the clinical data on the chronic effects of antibiotics were generated in Europe and in North America. Further research in low- and middle-income countries is warranted because the latter are especially burdened by antibiotic resistance problems, vulnerability to infections by antibiotic-resistant pathogens, and the corresponding effects on the gut microbiota (158).

Probiotics have long been viewed as a means of treating the acute gastrointestinal disorders associated with antibiotic-associated gut dysbiosis (8, 159–161). Can probiotics be recommended as adjunct treatments to mitigate the negative effects of antibiotics? The European Society for Paediatric Gastroenterology Hepatology and Nutrition (ESPGHAN) recommends strain-specific probiotics for the prevention of antibiotic-associated diarrhea in children, whereas the American Gastroenterological Association (AGA) guidelines indicate that strain-specific probiotics may be used to prevent Clostridiodes difficile infections (162, 163). For adults and children with IBDs like CD or UC, the AGA guidelines only recommend the use of probiotics only in the context of a clinical trial (162). Similarly, the guidelines issued by the European Crohn’s and Colitis Organisation and the ESPGHAN state that in patients with CD, probiotics should not be used to induce or maintain remission (163–165). According to the guidelines issued by the World Gastroenterology Organisation, there is evidence of strain-specific efficacy of probiotics in the prevention of antibiotic-associated diarrhea in adults or children who are receiving antibiotic therapy (166).

Hence, in view of the links between gut dysbiosis and the chronic diseases described above, one can reasonably hypothesize that the administration of probiotics will provide a degree of disease modification or symptom prophylaxis and mitigate the long-term consequences of antibiotic exposure. The main species to have been tested are lactic acid bacteria (such as Lacticaseibacillus rhamnosus GG, Limosilactobacillus reuteri, Lacticaseibacillus paracasei, Lactiplantibacillus plantarum, Lactobacillus acidophilus, Lactobacillus helveticus, Bifidobacterium lactis, Bifidobacterium breve, and Streptococcus thermophilus) and the yeast Saccharomyces boulardii. The positive reported short-term effects of probiotics in various patient populations (IBD, celiac disease, cancer, obesity, types 1 and 2 diabetes mellitus, allergic disease, idiopathic nephrotic syndrome, KD; multiple sclerosis, rheumatoid arthritis, and systemic lupus erythematosus, atherosclerosis, and neurological, neurodevelopmental and neurodegenerative diseases) will not be described in detail here because (i) the topic falls outside the scope of this review and (ii) the results have been extensively reviewed elsewhere in the literature (103, 159–161, 163–165, 167–210). Certain studies showed a positive effect of certain probiotic strains in combination with antibiotics and have provided a rationale for using probiotics to protect the gut microbiota and intestinal barrier functions.

Probiotics have been tested in animal models of chronic disease and in observational or interventional clinical studies, with moderate, variable but generally positive results: evidenced significant differences or improvements in clinical disease scores, symptom scores, and disease marker levels, whereas other studies found no benefit. In the field of cognitive and psychiatric disorders and mental health problems, the term “psychobiotic” has been used to describe probiotics that act through the gut-brain axis. Following on from extensive preclinical data on an association between antibiotic-induced gut dysbiosis and psychopathologies in the rat (211), there is preliminary evidence to suggest that specific probiotics may improve cognitive function, particularly in people with age-related mild cognitive impairment (212, 213).

In the literature on probiotics, the most common design is the randomized, controlled trial of the efficacy and safety of a probiotic in patients. The second most common design is the pharmacokinetic study, which documents the recovery and/or clearance of an oral dose of probiotic or measures pre−/post differences in the abundances of probiotic strains. According to McFarland, three models of dysbiosis have been frequently evaluated (214). In model A (restoration), probiotic therapy is initiated and studied after the microbiota of initially healthy patients has become disrupted (e.g., by antibiotic exposure). In model B (alteration), patients with a pre-existing disruption of the microbiota are studied after probiotic therapy. In model C (no dysbiosis), volunteers with no disruptive events are studied before and after probiotic therapy. In McFarland’s systematic review of 63 trials (published in 2014), 83% of the probiotic products evaluated in model A restored the microbiota. The corresponding proportion was 56% in model B. Only 21% of the probiotics evaluated in model C had an effect on the microbiota. Clinical efficacy was more commonly observed for with strains capable of restoring the normal microbiota (214).

Furthermore, meta-analyses of trials in better-studied disease areas have sometimes failed to show a clear, beneficial effect of probiotics. However, the meta-analyses’ authors almost always highlight the degree of interstudy heterogeneity with regard to probiotic doses, strains, and treatment durations. Larger, multicenter, randomized, controlled trial of probiotics (possibly simultaneously investigating the gut microbiota and disease markers) are warranted.

The gut virome and gut mycobiome have attracted less attention than the gut’s bacterial communities but, as mentioned above, are known to be abnormal in people with IBD (107–109, 215). We recommend further investigation of the potential indirect effects of antibiotic treatment and probiotics on the gut virome and mycobiome. Specific strains of yeast probiotics are a topic of interest; as mentioned above, one of the most effective and frequently evaluated probiotics is a yeast (Saccharomyces boulardii CNCM I-745) and therefore is not directly affected by antibiotics (8, 216–218). Saccharomyces boulardii is not a natural member of the human gut microbiota and is eliminated rapidly after probiotic administration is discontinued. However, when present as a probiotic, certain strains of S. boulardii exert several beneficial actions (including protection of the mucus layer, the stimulation of SCFA production by Lachnospiraceae and Ruminococcaceae, and a reduction in local inflammation) that counter antibiotic-associated dysbiosis (8, 216–218). In terms of the composition of the microbiota, treatment with S. boulardii is associated with increased abundances of Bacteroidaceae and Prevotellaceae and the suppression of pioneer bacteria (218). Treatment with S. boulardii CNCM I-745 can mitigate antibiotic-associated dysbiosis and diarrhea (219). More studies are needed to explore the full potential of this versatile probiotic yeast (218).

The results of our review indicate that antibiotic exposure is associated with a number of negative long-term (i.e., chronic) effects on health. Gut dysbiosis might be the causal link between antibiotic exposure and these chronic negative effects, although the lack of a replicable consensus definition of dysbiosis can lead to ambiguity in the interpretation of the data. Given that certain well-studied probiotics (such as S. boulardii CNCM I-745 and L. rhamnosus GG) are inexpensive, safe and effective in preventing short-term negative consequences of antibiotic exposure (including dysbiosis), there is no reason to summarily rule out potential longer-term benefits in a particular chronic disease setting or patient population. We recommend that decisions to initiate probiotic treatment should be made on a case-by-case basis after informed, evidenced-based discussion between the patient and his/her physician.