The Household Resistome: Frequency of β-Lactamases, Class 1 Integrons, and Antibiotic-Resistant Bacteria in the Domestic Environment and Their Reduction during Automated Dishwashing and Laundering

Abstract

Households provide a habitat for bacteria originating from humans, animals, foods, contaminated clothes, or other sources. Thus, bacteria carrying antibiotic resistance genes (ARGs) may be introduced via household members, animals, or the water supply from external habitats into private households and vice versa. Since data on antibiotic resistance (ABR) in the domestic environment are limited, this study aimed to determine the abundance of β-lactamase, mobile colistin resistance, and class 1 integron genes and the correlation of their presence and to characterize phenotypically resistant strains in 54 private households in Germany. Additionally, the persistence of antibiotic-resistant bacteria during automated dishwashing compared to that during laundering was assessed. Shower drains, washing machines, and dishwashers were sampled and analyzed using quantitative real-time PCR. Resistant strains were isolated, followed by identification and antibiotic susceptibility testing using a Vitek 2 system. The results showed a significantly higher relative ARG abundance of 0.2367 ARG copies/16S rRNA gene copies in shower drains than in dishwashers (0.1329 ARG copies/16S rRNA gene copies) and washing machines (0.0006 ARG copies/16S rRNA gene copies). bla_CMY-2, bla_ACT/MIR, and bla_OXA-48 were the most prevalent ARG, and intI1 occurred in 96.3% of the households, while no mcr genes were detected. Several β-lactamase genes co-occurred, and the resistance of bacterial isolates correlated positively with genotypic resistance, with carbapenemase genes dominating across isolates. Antibiotic-resistant bacteria were significantly reduced during automated dishwashing as well as laundering tests and did not differ from susceptible strains. Overall, the domestic environment may represent a potential reservoir of β-lactamase genes and β-lactam-resistant bacteria, with shower drains being the dominant source of ABR.IMPORTANCE The abundance of antibiotic-resistant bacteria and ARGs is steadily increasing and has been comprehensively analyzed in natural environments, animals, foods, and wastewater treatment plants. In this respect, β-lactams and colistin are of particular interest due to the emergence of multidrug-resistant Gram-negative bacteria. Despite the connection of private households to these environments, only a few studies have focused on the domestic environment so far. Therefore, the present study further investigated the occurrence of ARGs and antibiotic-resistant bacteria in shower drains, washing machines, and dishwashers. The analysis of the domestic environment as a potential reservoir of resistant bacteria is crucial to determine whether households contribute to the spread of ABR or may be a habitat where resistant bacteria from the natural environment, humans, food, or water are selected due to the use of detergents, antimicrobial products, and antibiotics. Furthermore, ABR could limit the options for the treatment of infections arising in the domestic environment.

Keywords: antibiotic resistance; class 1 integron; domestic environment; household; multiresistance; β-lactamases.

FIG 1

Violin plots showing the distribution and variation of the absolute (A) and relative (B) abundances of intI1 and bla genes in washing machine (WM), shower drain (SD), and dishwasher (DW) samples. The total for each gene group is presented: OXA genes (bla_OXA-58 and bla_OXA-23), carbapenemase genes (Carba; bla_OXA-48, bla_GES, bla_KPC, and bla_VIM), AmpC genes (bla_CMY-2, bla_FOX, bla_ACT/MIR, and bla_DHA), and CTX-M genes (bla_CTX-M-1 and bla_CTX-M-9). In panel B, the values for samples without an ARG (relative abundance = 0) are presented as −10. Different letters indicate significant differences at a P value of ≤0.05 between SD, WM, and DW samples in each group. Where no letters are shown, no significant differences were detected.

FIG 2

Correlation matrix showing the co-occurrence of bla genes in shower drains (A) and dishwashers (B). Red indicates a strong positive correlation, while blue indicates a strong negative correlation, and a correlation with an r value of ≥0.3 is statistically significant. bla genes in washing machine samples did not correlate significantly.

FIG 3

Percentage of β-lactam-resistant bacterial species (A) and bla genes detected in screened isolates (B) from the shower drains (SD), washing machines (WM), and dishwashers (DW) of 54 different households. Resistances and the ESBL phenotype were determined using a Vitek 2 system (bioMérieux), and MDR bacteria were classified according to international recommendations. Different letters indicate significant differences at a P value of ≤0.05 between SD, WM, and DW (A) and carbapenemase, CTX-M, and ampC genes (B). Where no letters are shown, no significant differences were detected.

FIG 4

Spearman correlation (P ≤ 0.05) of genotype (carbapenemase genes bla_OXA-58, bla_OXA-23, bla_OXA-48, bla_GES, bla_KPC, and bla_VIM; ampC genes bla_CMY-2, bla_FOX, bla_ACT/MIR, and bla_DHA; and CTX-M genes bla_CTX-M-1 and bla_CTX-M-9) with phenotypic resistance (to piperacillin-tazobactam [PIP/TAZ], carbapenems, and ceftazidime [CAZ] and/or cefotaxime [CTX]) in screened isolates from households.

FIG 5

Impact of a main wash with AOB-free detergent on antibiotic-resistant strains (A) and a rinsing cycle with and without benzalkonium chloride (BAC) (B) in the laundering process, simulated using a lab-scale washing machine (Rotawash), on resistant and susceptible strains of E. coli, K. pneumoniae, and S. aureus. Violin plots show the distribution and variation, and different letters indicate significant differences at a P value of ≤0.05. CPE, carbapenemase-producing Enterobacteriaceae.

FIG 6

Impact of automated dishwashing on antibiotic-resistant strains of E. coli, K. pneumoniae, and S. aureus with detergent. The different values for the maximum LR (+, indicating a complete reduction of the microbial load in the case of 45°C for 5 or 90 min and 60°C for 15 min) were obtained due to different initial loads on the biomonitors. The violin plots show the distribution and variation, and different letters indicate significant differences at a P value of ≤0.05.