Microbial diversity in individuals and their household contacts following typical antibiotic courses

Abstract

Background: Antibiotics are a mainstay of treatment for bacterial infections worldwide, yet the effects of typical antibiotic prescriptions on human indigenous microbiota have not been thoroughly evaluated. We examined the effects of the two most commonly prescribed antibiotics (amoxicillin and azithromycin) in the USA to discern whether short-term antibiotic courses may have prolonged effects on human microbiota.

Results: We sampled the feces, saliva, and skin specimens from a cohort of unrelated, cohabitating individuals over 6 months. An individual in each household was given an antibiotic, and the other a placebo to discern antibiotic impacts on microbiota, as well as determine whether antibiotic use might reshape the microbiota of each household. We observed household-specific patterns of microbiota on each body surface, which persevered despite antibiotic perturbations. While the gut microbiota within an individual became more dissimilar over time, there was no evidence that the use of antibiotics accelerated this process when compared to household members. There was a significant change in microbiota diversity in the gut and mouth in response to antibiotics, but analogous patterns were not observed on the skin. Those who received 7 days of amoxicillin generally had greater reductions in diversity compared to those who received 3 days, in contrast to those who received azithromycin.

Conclusions: As few as 3 days of treatment with the most commonly prescribed antibiotics can result in sustained reductions in microbiota diversity, which could have implications for the maintenance of human health and resilience to disease.

Keywords: 16S rRNA; Antibiotic courses; Antibiotic perturbations; Antibiotics; Gut; Microbiome; Saliva; Skin.

Fig. 1

Flowchart of study design

Fig. 2

Bar graph representing mean weighted UniFrac distances (±standard error) within a household (white bars) and between different households (gray bars) based on the antibiotic received in each household. The y-axis represents mean weighted UniFrac distances, while the body site sampled and antibiotic received within each household is represented on the x-axis. p values were determined using the Mann Whitney U test

Fig. 3

Bar graph (±standard error) representing the mean weighted UniFrac distances from day 0 in the feces of all subjects over time. The y-axis represents mean weighted UniFrac distances, and the x-axis represents the different subject groups over time based on the therapy they received. D3 represents day 3, D7 represents day 7, W8 represents week 8, and M6 represents month 6. p values were determined using the Kruskal Wallis test

Fig. 4

Principal coordinates analysis of beta diversity present in all subjects, time points, and sample types based on whether they received antibiotics (amoxicillin or azithromycin) or placebo

Fig. 5

Heatmaps representing the relative abundances of taxa in individuals taking antibiotics that were significantly different when compared to their housemates receiving placebo. Each individual taking an antibiotic is shown next to their housemate taking a placebo. Each household consisting of two subjects is separated by gray vertical boxes. a Feces, b saliva, and c skin. The family or order for each OTU shown on the heatmaps is shown to the right of each heatmap, and the antibiotic received is shown to the left of each heatmap. The index color scale is shown below

Fig. 6

Bar graphs (±standard error) representing the normalized difference in Shannon diversity in the gut between individuals taking antibiotics and their housemates taking placebo at each time point studied. a Households that took amoxicillin and placebo and b households that took azithromycin and placebo. All households collectively are represented by the blue bars, households that took 3 days of an antibiotic are represented by red bars, households that took 7 days of an antibiotic are represented by green bars, and control subjects who were not enrolled with housemates and are represented by purple bars. The x-axis represents the time point and the y-axis represents the change in normalized change in Shannon diversity since the prior time point. Negative results indicate lower diversity in the subjects taking antibiotics compared to their housemates taking placebo, and positive results indicate greater diversity in the housemates taking placebo compared to the individuals taking antibiotics. For the control subjects, the bars represent the mean change in diversity among all control subjects. p values were determined using the Mann Whitney U test, and *p values ≤0.05

Fig. 7

Bar graphs (±standard error) representing the change in Shannon diversity from day 0 across time in each subject group by body site tested. a Feces, b saliva, and c skin. Groups that received antibiotics, placebo, or no therapy (controls) are labeled across the x-axis, and the y-axis represents the change in Shannon diversity. *p values <0.05 using the Mann Whitney U test comparing subject groups at specified time points with the controls

Fig. 8

Bar graphs (±standard error) representing the normalized difference in Shannon diversity in the saliva between individuals taking antibiotics and their housemates taking placebo at each time point studied. a Households that took amoxicillin and placebo and b households that took azithromycin and placebo. All households collectively are represented by the blue bars, households that took 3 days of an antibiotic are represented by red bars, households that took 7 days of an antibiotic are represented by green bars, and control subjects who were not enrolled with housemates and are represented by purple bars. The x-axis represents the time point and the y-axis represents the change in normalized change in Shannon diversity since the prior time point. Negative results indicate lower diversity in the subjects taking antibiotics compared to their housemates taking placebo, and positive results indicate greater diversity in the housemates taking placebo compared to the individuals taking antibiotics. For the control subjects, the bars represent the mean change in diversity among all control subjects. p values were determined using the Mann Whitney U test, and *p values ≤0.05