Effects of breastfeeding on children’s gut colonization with multidrug-resistant Enterobacterales in peri-urban Lima, Peru

Figures & data

Table 1. Characteristics of 112 children participating in a prospective cohort study of enteric infections in Villa El Salvador, Lima, Peru, 2016–2019.

	N = 112
Male (%)	55 (49)
Birth mode (%)
Vaginal	78 (70)
Cesarean section	33 (30)
Missing	1
Mean no. of antibiotic courses consumed/child-year (SD)	4.34 (2.8)
Mean no. of diarrhea episodes/100 child-month (SD)	6.71 (4.9)
Maternal education level (%)
Less than high school	38 (30)
At least high school	74 (70)
Household size (%)
7 or more	27 (24)
5 to 6	41 (36)
≤4	44 (40)
Owns a refrigerator (%)	83 (74)
Water source (%)
Public tap	20 (18)
Indoor piped water	87 (78)
Other	5 (5)
Number of times water cuts experienced in past 2 months (%)
None	48 (43)
At least twice	37 (33)
3 times or more	26 (23)
Toilet type (%)
Indoor flush toilet	91 (81)
Pit latrine	19 (17)
Other	2 (2)
Household owns poultry (%)
Neither at enrollment nor during Year 2 visit	88 (79)
Both visits	10 (9)
Household owns dogs (%)
Neither at enrollment nor during Year 2 visit	50 (45)
Both visits	45 (40)
Household owns cats (%)
Neither at enrollment nor during Year 2 visit	34 (30)
Both visits	53 (47)

Number and percent are presented for categorical variables. Mean and standard deviation (SD) are presented for continuous variables. Household characteristics were surveyed during an annual questionnaire, administered for the first time at enrollment. Data reported here were collected during the enrollment questionnaire unless otherwise noted.

Figure 1. Changes in children’s exposures to breastmilk over the first 16 months of life, Lima, Peru, 2016–2019. Children’s feeding patterns were derived from 52,816 daily survey visits.

Table 2. Early-life feeding patterns of 112 children participating in a prospective cohort study of enteric infections in Villa El Salvador, Lima, Peru, 2016–2019.

	N = 112
Exclusively breastfed for at least six months (%)	7 (6)
Median duration of exclusive breastfeeding in months (SD)	2.9 (2.1)
Consumed any formula (%)	63 (56)
Median age in months at introduction of formula (SD)	1.0 (1.7)
Number of children receiving any breastmilk at 16 months (%)	95 (85)
Median duration of any breastfeeding in months (SD)	16.4 (6.4)

Number and percent are presented for categorical variables. Median and standard deviation (SD) are presented for continuous variables.

Figure 2. Gut colonization with ESBL-producing Enterobacterales over the first 16 months of life, Lima, Peru, 2016–2019. Changes in the mean point prevalence and incidence of a) ESBL-producing Escherichia coli and b) ESBL-producing Klebsiella spp., Enterobacter spp., or Citrobacter spp. Are depicted across specific age categories. Gray bands represent 95% confidence intervals around the mean.

Note: age (in months) was included as a continuous variable in all models. Age categories are depicted here for visualization purposes only: 1: ≥0.5 to <2months; 3: ≥2 to <3.5months; 4: ≥3.5 to 4.5months; 5: ≥4.5 to 5.5months; 6: ≥5.5 to 6.5months; 7: ≥6.5 to 8months; 9: ≥8 to <10.5months; 12: ≥10.5 to <14.5months; 16: ≥14.5months. For some children, two stool samples belonged to the same age bin, in this case, the earliest sample was chosen for depiction here.

Table 3. Associations between breastfeeding over the first 16 months of life and children’s incident gut colonization with extended-spectrum β-lactamase-producing enterobacterales, Lima, Peru, 2017–2019.

		ESBL-Escherichia coli			ESBL-Klebsiella spp., Enterobacter spp., Citrobacter spp.
	Follow-up time (child-months)	Incident gut-colonization episodes per 100 child-months	Adj. Risk Ratio (95% CI)^a	Pr (>|W|)	Incident gut-colonization episodes per 100 child-months	Adj. Risk Ratio (95% CI)^a	Pr (>|W|)
1–6 months
Percent of past 30 days child received only human milk^b	471	23.1	0.98 (0.94, 1.02)	0.29	16.6	0.99 (0.95, 1.04)	0.78
>6–16 months
Percent of past 30 days child received any human milk^b	1037	7.7	0.94 (0.90, 0.98)	0.003	6.7	1.01 (0.95, 1.08)	0.74

Pr (>|W|) = Probability of the observed Wald test statistic if the null hypothesis were true.

a Adjusted for child age, vaginal birth (yes/no), number of diarrhea episodes in the past 30 days, number of antibiotic courses in the past 30 days, and household poultry ownership (yes/no).

b Modeled in increments of 10% (3 days).

Figure 3. Core genome phylogenetic tree depicting the evolutionary relationships of 78 ESBL-producing E. coli from 47 children in Lima across two time points (predominantly breastfeeding versus breastfeeding while also eating complementary foods). We observed a high diversity of ESBL-ec and few differences in the occurrence of specific clonal complexes (CCs), sequence types (STs), phylogroups, ESBL alleles, or other antibiotic resistance genes (ARGs) across time points, suggesting the same ESBL-ec sequence types that were circulating among the broader community, in animals, and in the environment were also circulating within the household environment. Phylogroup designations A, B1, B2, or F are listed after highlighted CCs and STs.