Short term antibiotic effects on gut microbiome in Indian preschoolers: A 16S rRNA analysis

INTRODUCTION

The oral cavity, skin, and intestine of humans are inhabited by 1000 trillion microbiomes.^[1] The collection of microbiomes colonizing the gut is known as “gut microbiota”^[2] that forms a densely populated “mini ecosystem.”^[3] It is a microbial ecosystem where a diverse group of organisms live in close proximity to each other, interacting and influencing complex changes. Out of 55 phyla in the human body, the majority of the gut bacteria belong to the Bacteroidetes, Firmicutes, Proteobacteria, Actinobacteria, Fusobacteria, and Verrucomicrobia.^[4] An important characteristic feature of the intestinal microbiome is the composition of microbiota and the metagenome that remains largely unaffected irrespective of the type of food intake.^[3] Gut microbes are gaining importance due to their functions of stimulating host immune development, nutrient metabolism, differentiation of mucosal structure,^[3] maintaining intestinal mucosal barrier by enhancing gut integrity, formation of the intestinal epithelium, and protection against pathogens, providing anti-inflammatory signals to the host,^[2] and are necessary for the maintenance of intestinal homeostasis.^[5]

The role of microbiomes is crucial during early life as the changes in relation to the composition and the abundance of microbiomes become more or less stable, remain the same throughout life, and dictate the health of the host.^[5] Various internal and external factors influence the gut microbiota, starting from birth to adulthood.^[3] The human fecal microbiota consists of four main groups of bacteria (phyla), that are Firmicutes, Bacteroidetes, Proteobacteria, and Actinobacteria, of which the first two bacteria account for more than 80% of the microbiota. Firmicutes comprise mostly of Gram-positive bacteria, while Bacteroidetes include Gram-negative bacteria; Proteobacteria consist of Gram-negative bacteria and includes a wide variety of well-studied pathogens. Actinobacteria are a group of Gram-positive bacteria.^[6] The Firmicutes phylum is composed of Lactobacillus, Bacillus, Clostridium, Enterococcus, and Ruminococcus. Bacteroidetes phylum comprises four classes: Bacteroidia, Flavobacteria, Sphingobacteria, and Cytophagia.^[7]

There has been a rapid rise in the use of antibiotics for various diseases among adults and children.^[8] Infants, toddlers, and preschool children are usually treated with oral/intravenous antibiotics (e.g., penicillins, cephalosporins, and macrolides), due to their health conditions.^[9] Antibiotics disrupt the composition of the gut microbiota,^[10] affect the normal maturation of the microbiome, destabilizes it, and alter the basic physiological equilibria.^[11] Studies on the effect of exposure to antibiotics show that the abundance and diversity of intestinal microbiota are affected irrespective of which type of antibiotics are used.^[12] Early antibiotic exposure reduces microbiome composition and diversity with a marked reduction in Bifidobacterium and increases in Proteobacterium levels. Over time, when antibiotic treatment is stopped, the intestinal microbiota shows great resilience and returns to a composition like the original one, but it’s not fully recovered in most cases.^[13]

Several studies have demonstrated the impact of antibiotic usage on gut dysbiosis in adults, neonates, and infants. However, the short-term effect of antibiotics on gut dysbiosis in children is poorly understood.^[6] Therefore, this study aims to investigate the effect of antibiotics on gut composition in children aged 3–5 years receiving antibiotics compared to children who did not receive antibiotics.

MATERIALS AND METHODS

RESULTS

A total of 54 participants aged 3–5 years were included in the study, with 27 participants in Group 1 (treatment-experienced) and 27 participants in Group 2 (treatment-not-experienced). A total of 108 stool samples were collected from both groups on Day 0 (54 samples) and Day 5 (54 samples). All these samples were analyzed using 16S RNA gene amplicon sequencing.

Baseline demographics: In Group 1, 15 male and 12 female participants were included with a mean age of 46.4 months, whereas in Group 2, 16 male and 11 female participants were included with a mean age of 46.4 months.

The minimum and maximum gestational ages at birth between the two groups are between 37 and 40 weeks. About 40.7% of participants in Group I and 44.4% of participants in Group II had received breastfeeding. Participants having a normal diet in Group 1 and Group 2 are 33.3% and 30%, respectively. Normal vaginal delivery birth rates in Group 1 and II were 52% and 63%, respectively. In the 1^st year of life, only 3.7% of Group 1 participants and 14.8% of Group 2 participants had received antibiotics. In the 2^nd year, participants who received antibiotics in Group 1 and Group 2 were 48.1% and 40.7%, respectively. About 48.1% participants in Group 1 and 44.4% participants in Group 2 had received antibiotics at ages > 2 years.

The data have been categorized and analyzed at day 0 and day 5 for both males and females separately between the two groups. At the phyla level, the mean relative abundance of Bacteroidota and Firmicutes in both breast-feed and nonbreastfeed participants was P = 0.1728, P = 0.1292, P = 0.1212, and P = 0.9663, respectively, which was statistically less significant [Table 1]. A statistically significant difference was observed between the treatment-experienced and treatment-not-experienced groups on the mean relative abundance of Bacteroidota (P = 0.0496) and Verrucomicrobiota (P = 0.0291), while a less significant difference was seen in Firmicutes (P = 0.2795) [Table 2].

It was evident from the plot Figure 1 that there was a non-significant increase in the mean abundance of Bacteroidota, Proteobacteria, Actinobacteria, and Verrucomicrobiota in both groups from day 0 to day 5.

Close

Figure 1:

Relative abundance of phyla assigned to fecal microbiome of treatment-experienced and treatment-not-experienced participants on day 0 and day 5. Group 1 comprises treatment-experienced participants and group 2 comprises treatment-not-experienced participants. (1a) Subgroup treatment-experienced at day 0, (1b) subgroup experienced at day 5, (2a) subgroup treatment-not-experienced at day 0, and (2b) subgroup treatment-not-experienced at day 5. (F_1a) Treatment-experienced females at day 0, (F_1b) treatment-experienced females at day 5, (F_2a) treatment-not-experienced females at day 0, (F_2b) treatment-not-experienced females at day 5, (M_1a) treatment-experienced males at day 0, (M_1b) treatment-experienced males at day 5, (M_2a) treatment-not-experienced males at day 0, and (M_2b) treatment-not-experienced males at day 5 (M_2b).

Export to PPT

A mean relative increase in Firmicutes abundance was observed in both groups, more pronounced in Group 2 compared to Group 1.

At the phylum level, the mean relative abundance of Firmicutes was 1a < 1b, whereas that of Bacteroidota was 1b < 1a [Table 3] and sub categorization [Table 4].

At the family level, the mean relative abundance of Enterobacteriaceae in breast-feed participants was P = 0.401, which was statistically not significant; whereas non-breastfeed participants, it was P = 0.0016, which was statistically significant. The mean relative abundance of Enterococcaceae in both breast-feed and non-breastfeed participants was P = 0.0068, and P = 0.0162, respectively, which was statistically significant [Table 1].

Figure 2 and Table 5 revealed an increase in the relative abundance of Enterobacteriaceae, Enterococcaceae, and Peptostreptococcaceae on day 5 over day 0 of sample collection in participants with antibiotic treatment. A significant decrease in the relative abundance of Lactobacillaceae was observed in Group 1 from day 0 to day 5.

Close

Figure 2:

Relative abundance of family level assignments of the fecal microbiome of treatment-experienced and treatment -not-experienced female and male participants on zero and fifth day of sampling. Group 1 comprises treatment-experienced participants and group 2 comprises treatment-not-experienced participants. (1a) Subgroup treatment-experienced at day 0, (1b) subgroup treatment-experienced at day 5, (2a) subgroup treatment-not-experienced at day 0, and (2b) subgroup treatment-not-experienced at day 5. (F_1a) Treatment-experienced females at day 0, (F_1b) treatment-experienced females at day 5, (F_2a) treatment-not-experienced females at day 0, (F_2b) treatment-not-experienced females at day 5, (M_1a) treatment-experienced males at day 0, (M_1b) treatment-experienced males at day 5, (M_2a) treatment-not-experienced males at day 0, and (M_2b) treatment-not-experienced males at day 5.

Export to PPT

A statistically significant (P < 0.05) increase in the relative abundance of Enterococcaceae and Peptostreptococcaceae is observed in Group 1 from day 0 to day 5. A similar increase was observed in the relative abundance in Group 2. However, the increase in Enterococcaceae was less in participants without antibiotics when compared to participants on antibiotics.

At the family level, the mean relative abundance of Enterococcaceae was 1a < 1b, whereas that of Lactobacillaceae was 1b < 1a [Table 6] and subcategorization [Table 7].

The mean relative abundance of Escherichia-Shigella at genus level in breast-feed participants was P = 0.6228, which was statistically non-significant; whereas non-breastfeed participants, it was P = 0.0011, which was statistically significant. The mean relative abundance of Enterococcaceae in both breast-feed and non-breastfeed participants was P = 0.0068, and P = 0.0141, respectively, which was statistically significant [Table 1].

A significant decline was observed in the relative abundance of Lactobacillus from day 0 to day 5 in Group 1, while it increased from day 0 to day 5 in Group 2. The relative abundance of Enterococcus increased significantly from day 0 to day 5 in Group 1.

The relative abundance of Bifidobacterium increased from day 0 to day 5 in Group 1 and Group 2 participants. However, these changes are not statistically significant.

At the Genus level, the mean relative abundance of Enterococcus was 1a < 1b, whereas Lactobacillus was 1b < 1a [Figure 3 and Tables 8-10].

Close

Figure 3:

Relative abundance of genus level assignments of the fecal microbiome of treatment-experienced and treatment-not -experienced female and male participants on zero and fifth day of sampling. Group 1 comprises treatment-experienced participants and group 2 comprises treatment-not-experienced participants. (1a) Subgroup treatment-experienced at day 0, (1b) subgroup treatment-experienced at day 5, (2a) subgroup treatment-not-experienced at day 0, and (2b) subgroup treatment-not-experienced at day 5. (F_1a) Treatment-experienced females at day 0, (F_1b) treatment-experienced females at day 5, (F_2a) treatment-not-experienced females at day 0, (F_2b) treatment-not-experienced females at day 5, (M_1a) treatment-experienced males at day 0, (M_1b) treatment-experienced males at day 5, (M_2a) treatment-not-experienced males at day 0, and (M_2b) treatment-not-experienced males at day 5.

Export to PPT

In Figure 4, beta-diversity indices revealed significant differences at group and subgroup levels regarding the bacterial counts. The biggest shift in the microbial community was observed in the antibiotic group for female participants from day 0 (black eclipse) to day 5 (gray eclipse); whereas a similar but smaller shift was observed for the male participant group with antibiotic treatment from day 0 (yellow eclipse) to day 5 (dark green eclipse). Among the antibiotic treatment groups for both male and female participants, the microbial communities were more heterogeneous after 5 days of antibiotic treatment. On the other hand, the microbial communities either remained similar or became more homogeneous on the 0 day and 5^th day of sample collection for both participant groups without antibiotic treatment. However, a slight shift was noticed from the 0-day (red eclipse) to 5^th day (blue eclipse) in samples collected from the female participants without antibiotic treatment. In the case of male participants without antibiotic treatment, the microbial communities become more homogenous on the 5^th day (light green eclipse) than on the 0-day (brown eclipse).

Close

Figure 4:

Principal coordinate analysis (PCoA) plots depicting beta-diversity calculated using the Bray-Curtis similarity distance among the prokaryotic communities in the fecal samples collected from the female participants at day 0 and day 5. (F_1a) treatment-experienced females at day 0, (F_1b) treatment-experienced females at day 5, (F_2a) treatment -not-experienced females at day 0, (F_2b) treatment not experienced females at day 5, (M_1a) treatment-experienced males at day 0, (M_1b) treatment-experienced males at day 5, (M_2a) treatment-not-experienced males at day 0, and (M_2b) treatment-not-experienced males at day 5.

Export to PPT

The Firmicutes/Bacteroidetes (F/B) ratio is widely accepted to have an important influence on maintaining normal intestinal homeostasis. An increased or decreased F/B ratio is considered as dysbiosis.^[23] The presented study shows that the F/B ratios of Group 1 and Group 2 were 1.535 and 1.006, respectively [Tables 11-13].

DISCUSSION

Antibiotics are frequently used in children to treat common infections and diseases. However, little is known about the effects of antibiotics on the composition and load of the gut microbiota immediately after treatment.^[2,6] The present study was, thus, undertaken to study how short-term antibiotic usage is associated with altered microbial abundance and diversity.

Furthermore, this study reasserts that there is an increasing need for global awareness and a detailed understanding of the relationship between antibiotic use and gut dysbiosis. The need of the hour is that healthcare practitioners consider the damage to the gut microbiome while prescribing antibiotics for children and limit their systematic use as they can reshape the microbiota in favor of resistant bacterial strains in the long term.

In the present study, Firmicutes and Bacteroidota represent a large majority of the prokaryotic communities in both groups that were exposed and not exposed to antibiotics for 5 days. This is in concordance with the findings of Wei et al. who observed that the most abundant phyla reported in the fecal samples of children were Bacteroidetes and Firmicutes, followed by the Proteobacteria, Actinobacteria, and Verrucomicrobia, 14 days after treatment with azithromycin in children aged 12–36 months.^[8] On comparing groups with azithromycin and placebo, Parker et al. stated that the relative abundance of Proteobacteria, and Verrucomicrobia decreased on day 14. Study findings revealed a modest effect of azithromycin on the composition of the bacterial microbiota among 6–11-month-old infants.^[24] The present study also showed a significant decrease in Verrucomicrobia mean relative abundance in the antibiotic-treated group but an increase in the Proteobacteria group, which could Possibly be due to the brief period of observation and short-term antibiotic usage in our study.

Ma et al. and Li et al. stated that breast milk is the main influence of gut microbiota, with differences observed among infants fed exclusively or with formula. Firmicutes on days 0 and 30 dominated the breast milk gut microbiota.^[25,26] Similarly, the present study reported comparable results with gut microbiota composition. It was also observed in the present study that the mean relative abundance of Bacteroidetes significantly decreased, and there was a non-significant increase in the mean relative abundance of Firmicutes in both groups from day 0 to day 5. This was in contrast to the findings of Kwon et al., who demonstrated a significant decrease in Firmicutes and Bacteroidetes phyla in the antibiotic group as compared to the control group in infants under 3 months of age.^[2] Firmicutes count also decreased from 36% to 4% after 5 days of antibiotic treatment in a child with otitis media, as reported by Sturød et al.^[27]

The present study demonstrated a significant increase in the Escherichia-Shigella groups at the genus level in the antibiotic group as compared to the non-antibiotic group. This was in concordance with the study of Kwon et al., who reported a significant increase in the Escherichia/Shigella, and Bifidobacterium groups at the genus level in the antibiotic group as compared to the control group.^[2] Wei et al. also found that the Bifidobacterium count significantly reduced in the azithromycin group at day 14 of fecal sample analysis.^[8] While Mangin et al. found no significant differences in total Bifidobacteria concentrations after amoxicillin treatment for 7 days, the average number of Bifidobacterium species per microbiota was significantly lower for treated infants compared to the healthy group at day 7.^[28]

Alpha diversity indices in the present study suggested no significant differences between the two groups at baseline or post-treatment [Supplementary Figure 1]. This was unlike the study of Oldenburg et al., who found non-significant differences at baseline and significant differences in Simpson’s (0.003) and Shannon’s (0.0001) a-diversity indices in all four treatment arms on the 5^th day post-treatment.^[10] Furthermore, in a study conducted by Kwon et al., similarly, significant differences in Chao1 (0.033) and Shannon index (0.009) between the control and the antibiotic groups at 4 weeks of sample collection were observed.^[2] Similar to our study, Doan et al. found that alpha diversity indices at baseline were non-significant across the two groups, but analysis of stool samples after 5 days of antibiotic treatment revealed significant changes in Inverse Simpsons’ a-diversity indices, with the antibiotic-treated group showing decreased microbial count.^[29] Wei et al. also demonstrated that the Shannon diversity index showed statistically lower results for the azithromycin group as compared to the placebo on 14 days of antibiotic use.^[8]

In the present study, the b-diversity indices showed that the microbial communities in the antibiotic-treated group were more heterogeneous on day 5 as compared to the non-antibiotic treated group, which showed a more homogenous composition on both days. This finding harmonizes with the study of Bokulich et al., who in their analysis, found a significant relation in the b-diversity index of stool samples of children collected over 2 years. In their study, antibiotic exposure was associated with deficits in Clostridiales and Ruminococcus from 3 to 9 months of life but with no consistent changes in other taxa.^[30] On the other hand, Doan et al. reported that b-diversity indices did not show any significance in the azithromycin versus placebo group 5 days post-treatment.^[29]

The present study also revealed that a higher abundance of Erysipelatoclostridium, Clostridium species, Ruminococcus, and Eschirichia-Shigella was seen on day 5 of antibiotic treatment [Supplementary Figure 2]. This is in contrast to a study by Abeles et al. (2016), who observed depletion of the Erysipelotrichaceae, Veillonellaceae, and Clostridiales in the gut flora of children following a 3 or 7 day antibiotic course.^[31] Doan et al. stated in their study that Faecalibacterium, Blautia, Bifidobacterium, Succinivibrio, Ruminococcus, Roseburia, Escherichia, and Clostridium, account for 61% (higher abundance) of the filtered reads on days 0 and 5 for Antibiotic-treated group.^[29]

The present study has shown that the relative abundance of Enterococcus, Lactobacillus, Sellimonas, and Eggerthella was significantly higher in participants receiving antibiotic treatment. On the contrary, the relative abundance of Prevotella, Dialister, and Senegalimassilia was lower in participants with antibiotic treatment in comparison to participants without antibiotic treatment. Prediction of the functions of prokaryotic communities thereby revealed a higher abundance of genes associated with antibiotic resistance in the samples from the participants with antibiotic treatment in comparison to the non-antibiotic group.

In their study, Panda et al. (2014) reported that fluoroquinolones and b-lactams significantly decreased microbial diversity by 25% and reduced the core phylogenetic microbiota from 29 to 12 taxa. However, at the phylum level, these antibiotics increased the Bacteroidetes/Firmicutes ratio (B/F ratio) (P = 0.0007, FDR = 0.002).^[6] In contrast to this, the present study demonstrates that antibiotics increased the phylum Firmicutes from day 0 to day 5, while Bacteroidota decreased from day 0 to day 5, decreasing the B/F ratio. In our study, the F/B ratio of the antibiotic group was 1.535. In subgroup analysis, the F/B ratio of group 1 at day 0 and day 5 was 1.506 and 1.564, respectively.

The differences in these findings obtained in our study and previous studies could possibly be attributed to the age of the study participants, lifestyle-associated factors including diet, physical activity, food additives and contaminants, antibiotic consumption, physical activity, the study setting (different geographical areas have been shown to have different microbial compositions in the intestine), the study duration, and the time since the antibiotic exposure.

This study has few limitations considering the short-term duration of post-antibiotic exposure observation and the inclusion of participants from the same region. Furthermore, this study did not give importance to the class and type of antibiotics used by the study participants.

CONCLUSIONS

Several studies have demonstrated that the gut microbiome is sensitive to antibiotic treatment. However, there has been little to no information available on the short-term usage of antibiotics in children aged 3–5 years. This is the first Indian study conducted on children to determine the differences in the gut composition of participants receiving antibiotics compared to participants who did not receive antibiotics.

In our study, it was observed that there was a significant decrease in the gut flora of study participants who belonged to the treatment-experienced group from day 0 to day 5 at the phyla, family, and genus level, whereas in the treatment-not-experienced group, there were no significant changes in the gut flora from day 0 to day 5. Future studies involving specific classes of antibiotics to study their effects on the gut microbiome are warranted.