Keywords
Escherichia coli
- fecal carriage - virulence factors - extended-spectrum β-lactamase - ESBL
Extended-spectrum β-lactamase (ESBL)-producing Escherichia coli is a recognized public health threat, where it can lead to potentially life-threatening
bloodstream infections. Escherichia coli bloodstream infections present a substantial challenge within neonatal care and are
acknowledged as an important global public health concern.[1] Maternal transmission has been recognized as a significant source of neonatal colonization
with E. coli, and extended stay in the intensive care unit increases the risk of nosocomial infections
due to gram-negative rods such as E. coli.[2] This study aimed to investigate the frequency of ESBL-producing E. coli in an obstetric ward in Changchun City, China, a country where a previously reported
high prevalence of ESBL-producing Enterobacteriaceae has reached up to 58%.[3] The occurrence of virulence factors linked to invasive disease was examined in all
ESBL-producing E. coli using a recently established virulence database.[4]
Materials and Methods
Patients
Between September 2013 and January 2014, ESBL-producing E. coli isolates from fecal samples acquired from pregnant women and their newborns at the
Children's Hospital in Changchun City, China, were consecutively collected. The hospital
is a referral and teaching hospital in Northeast China, serving approximately 20,000
inpatients annually from a catchment area of approximately 4,050,000 inhabitants.
After recording information on whether the isolate originated from a mother or a child,
the isolates were treated anonymously. The study was reviewed by the local ethical
review board prior to its start, and after obtaining patient consent, fecal samples
were collected from both mothers and their newborns. Maternal specimens were obtained
through rectal swabs immediately after defecation, while neonatal specimens were collected
from feces and transported in charcoal Amies media (Copan, Brescia, Italy) to the
laboratory. Maternal specimens were taken upon arrival at the obstetric department
before delivery, and one sample was collected from the newborns before they left the
maternity ward, approximately 1 week later. Neonates in need of neonatal inpatient
care were excluded from the study.
Culture-Based Methods
All swabs were plated on chromogenic urinary tract infection (UTI) agar plates (Oxoid,
United Kingdom) and incubated at 35 °C overnight. Subcultivation of presumptive coliform
bacteria on chromogenic UTI agar was performed on chromogenic ESBL agar plates (Oxoid,
United Kingdom). Presumptive E. coli (pink colonies) were susceptibility tested according to the European Committee for
Antimicrobial Susceptibility Testing (EUCAST, www.eucast.org) using a diffusion synergy test with clavulanic acid (10 μg) and the cephalosporins—cefotaxime
(5 μg), ceftazidime (10 μg), and cefepime (30 μg). Isolates were categorized as reduced
susceptibility when the zone diameter for cefotaxime, ceftazidime, or cefepime was
<17 mm. Isolates were categorized as classical ESBL phenotype when synergy between
clavulanic acid and the tested cephalosporins was observed. For isolates with a negative
synergy test but reduced susceptibility to cefotaxime or ceftazidime and cefoxitin
(30 μg; <19 mm), an AmpC-type was suspected. All E. coli isolates with suspected ESBL production were stored in glycerol stock at −80 °C.
Whole-Genome Sequence Analysis of Bacterial Isolates and Calculations
ESBL-producing E. coli isolates were analyzed by whole-genome sequencing. Extraction, library preparation,
and bioinformatics analysis were performed as described elsewhere, all sequence data
are available through EnteroBase (
https://enterobase.warwick.ac.uk/
).[4] ESBL-production was confirmed by BLASTn (nucleotide Basic Local Alignment Search
Tool) searches on draft genomes by using Comprehensive Antibiotic resistance database
(CARD) (
https://card.mcmaster.ca/
; October 2020) as reference database. Gene prediction for virulence factors was performed
on draft genomes by using the BLASTn algorithm with standard settings and a previously
established virulence database for ExPEC (Extraintestival pathogene E. coli).[4] Virulence factors were interpreted as predicted with a nucleotide coverage of ≥99%
and nucleotide identity of ≥98%. Phylotypes were determined according to the Clermont
scheme, and the phylogenetic relationship of all isolates was analyzed using hierarchical
clustering as implemented in EnteroBase (
https://enterobase.warwick.ac.uk
). The frequency of virulence factors was compared with those from invasive E. coli isolates derived from neonates. To do so, a total of 32 isolates published in a previous
study, isolated from blood and cerebral spine fluid cultures at Uppsala University
Hospital between 2005 and 2015, were previously analyzed accordingly.[4] Genetic determinants for virulence factors were treated as presence/absence data
and statistical computations were performed using the statistical software R (v4.3.0,
April 21, 2023, R Foundation for Statistical Computing, Vienna, Austria). Odds ratios
and statistical significance were calculated using Fisher's exact test for small sample
sizes. All sequence data are publicly available through EnteroBase, accession numbers
are listed in [Supplementary Material S1] (available in the online version).
Results
In the present investigation, a total of 347 samples were analyzed, and ESBL-producing
E. coli were detected in 18% (31/177) of the maternal samples and 5% (9/170) of the newborns.
Hierarchical clustering as implemented in EnteroBase defined three isolate pairs,
three isolates, and additional four isolates as indistinguishable, suggesting maternal
or health care-related transmission to the newborn. All phylogenetic lineages according
to Clermont were represented and distributed as follows: A (11/40, 28%), D (11/40,
28%), B2 (10/40, 25%), B1 (4/40, 10%), E (2/40, 5%), C (1/40, 3%), and F (1/40, 3%).
From our isolates, a representative isolate from each indistinguishable isolate pair
was included in further analysis, resulting in 32 fecal isolates included in the comparison.
Virulence factors from all virulence factor groups were found in both fecal and invasive
isolates, respectively: fimbriae and adhesins 25/67 (37%) versus 29/67 (43%), iron
metabolism 12/13 (92%) versus 13/13 (100%), exotoxins 7/19 (37%) versus 9/19 (47%),
immunomodulation 15/15 (100%) versus 15/15 (100%), bacteriocins 4/29 (14%) versus
14/29 (48%), and capsule types 2/2 (100%) each. Identical isolates accordingly shared
the same virulence factor pattern. No virulence factor was found to be significantly
more frequent in fecal isolates overall, with the most notable differences between
fecal and invasive isolates being observed in virulence factors targeting iron metabolism
and immunomodulation. Significantly less often found in fecal isolates were determinants
for the siderophores salmochelin and yersiniabactin, as well as determinants for the
iron acquisition proteins Fec, Fhu, and ChuA. Iron is regarded as essential to E. coli causing invasive disease, especially in neonatal meningitis. Likewise, genes encoding
proteins involved in translocation of the blood–brain barrier (IbeB), the iss determinant and capsule K1 involved in complement resistance, the plasminogen activator
OmpT, the cell membrane stabilizing Tol-Pal, and the genotoxin colibactin were statistically
significantly less often found in the fecal ESBL-producing isolates ([Table 1] and [Fig. 1]).
Table 1
Comparison of virulence factor frequency of ESBL-producing Escherichia coli isolates from feces with invasive E. coli isolates, only statistically significant results were listed
|
Virulence factor group
|
Genetic determinants for virulence factors found
|
ESBL-producing E. coli from feces
|
Invasive E. coli isolates
|
OR
|
p-Value
|
|
Fimbriae and adhesins
|
Type 1 fimbriae
|
24/32 (75%)
|
31/32 (97%)
|
10
|
<0.05
|
|
Iron metabolism
|
Salmochelin
|
4/32 (13%)
|
15/32 (47%)
|
6
|
<0.05
|
|
Yersiniabactin
|
23/32 (72%)
|
30/32 (94%)
|
6
|
<0.05
|
|
Fec
|
13/32 (41%)
|
22/32 (69%)
|
3
|
<0.05
|
|
Fhu
|
12/32 (38%)
|
32/32 (100%)
|
48
|
<0.005
|
|
ChuA
|
6/32 (19%)
|
27/32 (84%)
|
22
|
<0.005
|
|
Exotoxins
|
Colibactin
|
2/32 (6%)
|
10/32 (31%)
|
7
|
<0.05
|
|
Immunomodulation
|
IbeB
|
14/32 (44%)
|
23/32 (72%)
|
3
|
<0.05
|
|
Iss
|
17/32 (53%)
|
26/32 (81%)
|
4
|
<0.05
|
|
OmpT
|
13/32 (41%)
|
25/32 (78%)
|
5
|
<0.005
|
|
Tol-Pal
|
22/32 (69%)
|
32/32 (100%)
|
14
|
<0.005
|
|
Capsule
|
K1
|
4/32 (13%)
|
12/32 (38%)
|
4
|
<0.05
|
|
Bacteriocins
|
Microcin H4
|
0/32 (0%)
|
6/32 (19%)
|
7
|
<0.05
|
Abbreviations: ESBL, extended-spectrum β-lactamase; OR, odds ratio.
Fig. 1 Schematic illustration of the different virulence factors for Escherichia coli.
Discussion
The fecal screening for ESBL-producing E. coli conducted here revealed a relatively high perinatal colonization frequency, with
18% detected in mothers and 5% in newborns within days after delivery. The overall
occurrence of virulence factors associated with extraintestinal infections was significantly
lower in fecal isolates compared with those from extraintestinal sources. It is reassuring
that the ESBL-producing E. coli from our study exhibit fewer virulence traits compared with invasive isolates. This
is consistent with findings from Rottier et al, where the positive predictive value
of fecal carriage of Enterobacteriaceae with ESBL to develop symptomatic infection
was relatively low.[5]
The establishment of serious invasive infections in neonates with the opportunistic
pathogen E. coli is mainly caused by bacterial translocation from the intestine to the bloodstream.
The prerequisites for this process are successful bacterial colonization of the gut
and effective interaction of the bacteria with the host. Shortly after birth, a newborn
rapidly acquires colonization, primarily from the mother's microbiome.[6] Isolates from phylogroup non-B2 are less frequently associated with invasive infections
and are often considered true commensals of the gut. In contrast, E. coli from phylogroup B2 not only possesses acquired virulence factors, providing them
with the prerequisites for invasiveness but has also been linked to persistent colonization
of the intestine and invasive infections.[7]
Screening for ESBL-producing Enterobacteriaceae in neonatal intensive care units is
commonly justified to increase awareness for outbreak-preventing strategies, given
the catastrophic consequences of uncontrolled spread.[8] Moreover, adequate treatment of ESBL-producing isolates in an individual can be
life-saving when a symptomatic infection arises.[9] However, the understanding of the risk of symptomatic infection attributable to
intestinal carriage of ESBL-producing Enterobacteriaceae remains limited.[10] In recent years, it has been acknowledged that patients and their caregivers may
undergo feelings of worry, anxiety, or stigma upon receiving positive test results
and guidance from health care providers regarding the implications of ESBL colonization
in the gut can facilitate coping.[11]
Conclusion
All the above underscore the necessity for a deeper understanding of the implications
of ESBL-producing Enterobacteriaceae colonization, particularly concerning antimicrobial
treatment decisions. Uncertainty in risk assessment of carriage of resistant bacteria
may result in unnecessary use of broad-spectrum antibiotics and impact the intestinal
microbiome.
