Infectious Diseases

ISSN: 2374-4235 (Print) 2374-4243 (Online) Journal homepage: https://www.tandfonline.com/loi/infd20

Risk factors for community acquired urinary

tract infections caused by extended spectrum
β-lactamase (ESBL) producing Escherichia coli in
children: a case control study

Frank H. Zhu, Maria P. Rodado, Basim I. Asmar, Hossein Salimnia, Ronald

Thomas & Nahed Abdel-Haq

To cite this article: Frank H. Zhu, Maria P. Rodado, Basim I. Asmar, Hossein Salimnia, Ronald
Thomas & Nahed Abdel-Haq (2019): Risk factors for community acquired urinary tract infections
caused by extended spectrum β-lactamase (ESBL) producing Escherichia�coli in children: a case
control study, Infectious Diseases, DOI: 10.1080/23744235.2019.1654127

To link to this article: https://doi.org/10.1080/23744235.2019.1654127

Published online: 20 Aug 2019.

Submit your article to this journal

Article views: 14

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=infd20
INFECTIOUS DISEASES, https://doi.org/10.1080/23744235.2019.1654127
2019; VOL. 0,
NO. 0, 1–8

ORIGINAL ARTICLE

Risk factors for community acquired urinary tract

infections caused by extended spectrum b-lactamase
(ESBL) producing Escherichia coli in children: a case
control study

Frank H. Zhua, Maria P. Rodadoa, Basim I. Asmara,b, Hossein Salimniac,d, Ronald Thomasb and
Nahed Abdel-Haqa,b
a
Division of Infectious Diseases, Children’s Hospital of Michigan, Detroit, MI, USA; bCarman and Ann Adams Department of
Pediatrics, Wayne State University, Detroit, MI, USA; cDepartment of Pathology, Wayne State University, Detroit, MI, USA;
d
Detroit Medical Center University Laboratories, Detroit, MI, USA

ABSTRACT
Background: We noted a recent increase in cases of urinary tract infection due to community-acquired ESBL-producing
Escherichia coli in children treated at our institution. Risk factors of urinary tract infection due to ESBL-producing E. coli in
children in the USA remain unclear.

Methods: A single center retrospective case control study of UTI due to CA-ESBL-producing E. coli during a 5-year period
(2012–2016). Control cases with non-ESBL-producing E. coli urinary tract infection were matched by age, gender and year
of infection.

Results: A total of 111 patients with ESBL-producing E coli urinary tract infection and 103 controls were included. The propor-
tion of ESBL-producing E coli urinary tract infection ranged from 7% to 15% of all UTI cases. The distribution of ESBL cases per
year: 27 in 2012; 18 in 2013; 22 in 2014; 15 in 2015 and 29 in 2016. Median age was 4 years with female predominance (84%).
The ESBL group was predominantly African American (32%) followed by individuals of Middle Eastern ethnic background
(31%). Risk factors by univariate analysis were vesicoureteral reflux: (20.9 ESBL group vs 6% controls; p ¼ .002), prior antibiotic
usage in the last 3 months (including b-lactams), prior UTI (last 3 months), recent hospitalization (last 3 months) and Middle
Eastern ethnic background. However, multivariate analysis showed that only prior antibiotic usage (p ¼ .001) and Middle
Eastern ethnic background (p < .001) were independent risk factors. ESBL-producing strains were more frequently resistant to
trimethoprim-sulfamethoxazole (72% vs 25%) and ciprofloxacin (73% vs 5%) than strains not producing ESBL.

Conclusion: Risk factors for community-acquired ESBL-producing E coli urinary tract in our pediatric patient population
were antibiotic usage within the previous 3 months and Middle Eastern ethnic background. This may be related to
increased risk of intestinal colonization with resistant bacterial strains.

KEYWORDS ARTICLE HISTORY CONTACT

Children Received 25 April 2019 Nahed Abdel-Haq
UTI Revised 1 August 2019 nabdel@dmc.org
ESBL Accepted 5 August 2019 Division of Infectious Diseases, Children’s
E. coli Hospital of Michigan, 3901 Beaubien Blvd,
risk factors Detroit, MI 48201, USA

ß 2019 Society for Scandinavian Journal of Infectious Diseases

2 F. H. ZHU ET AL.
Background
Extended-spectrum b lactamases (ESBLs) are b-lacta-
mases that hydrolyze extended-spectrum cephalosporins
with an oxyimino side chain (cefotaxime, ceftriaxone,
and ceftazidime), as well as the oxyimino-monobactam
aztreonam. Thus ESBLs confer resistance to these antibi-
otics and related oxyimino-b lactams [1,2].
Frequent use of antibiotics in humans and animals
has caused increased resistance in Gram-negative
(G ve) bacteria in recent years. Increasing resistance
has been reported in G ve bacteria to antibiotics com-
monly used in the treatment of urinary tract infections
(UTI) such as trimethoprim-sulfamethoxazole, fluoroqui-
nolones and other standard antibiotics [3]. However, in
recent years, G ve bacteria that produce ESBLs have
emerged [4]. Initially these organisms were isolated from
patients who were hospitalized or received antibiotic
treatment. However, an increasing number of patients
with community-acquired infections has been noted.
These patients were mainly adults and their infections
with ESBL producing organisms were associated with
increased morbidity and mortality [5–7].
The most common organisms associated with com-
munity-acquired ESBL-positive infections are Escherichia
coli and Klebsiella species, bacteria that cause acute
uncomplicated UTI, thus making the treatment of UTI
more challenging. The identification of risk factors for
antimicrobial resistance may improve the empirical treat-
ment of UTI. In adult patients, reported risk factors for
community-acquired UTI due to ESBL-producing G ve
bacteria included older age, female sex, diabetes melli-
tus, recurrent UTI, invasive urological procedures, previ-
ous urine catheters and prior use of antibiotics [8–10].
Previous use of cephalosporins, penicillins or fluoroqui-
nolones was associated with infections caused by ESBL-
producing G ve bacteria [8,9].
Studies of risk factors associated with infections by
ESBL-producing G ve bacteria in children focused
mainly on hospitalized children [11,12]. Few studies on
children with community acquired infections caused by
ESBL-positive bacteria have been published [13,14].
Topaloglu et al. found that an underlying disease and
hospitalization within the last 3 months were risk factors
for infection with ESBL-producing E coli and Klebsiella in
children [13]. Previous exposure to antibiotics and young
age (<1 year) have also been reported to be risk factors
[15]. However, these studies were done in countries with
a known high prevalence of infections with ESBL-posi-
tive bacteria or included a high proportion of patients
who were receiving antibiotic prophylaxis [16]. The rates
of infection with ESBL-positive bacteria in children in the
USA are increasing [17], but the risk factors of ESBL-pro-
ducing bacteria in community acquired urinary tract
infection (CA-UTI) in children in the USA remain unclear.
In recent years, increasing numbers of children with
CA-UTI due to ESBL-producing organisms, especially
E. coli, have been observed at our institution. The pri-
mary aims of this study were to determine the fre-
quency of CA-UTIs caused by ESBL-producing E. coli in
children treated at Children’s Hospital of Michigan dur-
ing 2012–2016 and to investigate the characteristics of
these children to determine risk factors associated with
these infections.
Methods
A retrospective, case-control study of children present-
ing with CA-UTI due to ESBL-producing E. coli from
January 2012 to December 2016 at the Children’s
Hospital of Michigan, a 220-bed tertiary care center in
southeastern Michigan, was performed. The study was
approved by the Institutional Review Board at Wayne
State University, Detroit, MI, USA.
Urine cultures positive for ESBL-producing E. coli were
identified from the microbiology laboratory records.
Bacterial isolates and their antibiotic susceptibilities were
identified by BD Phoenix identification and susceptibility
combo panels (NMIC/ID303). The ESBL screen by BD
Phoenix Gram Negative panel involves six test wells of
different cephalosporins with and without clavulanic
acid. This methodology meets the recommendations of
the Clinical and Laboratory Standards Institute (CLSI)
and the European Committee on Antimicrobial
Susceptibility Testing (EUCAST). The drug combinations
in the wells, as well as the algorithm used for interpret-
ation of results, is proprietary to Becton Dickinson,
Sparks, MD, USA [18].
Positive urine cultures were defined according to the
method of collection of the urine sample. Bag speci-
mens were not included in the analysis. For midstream
urine specimens, culture was defined as positive if 105
colony forming units (CFU)/mL of bacterial growth was
present. Cultures with 104–105 CFU/mL of growth were
considered positive if there was pyuria of 10 leuko-
cytes per high power field. For specimens obtained by
bladder catheterization, growth of 104–105 CFU/mL was
defined as a positive urine culture.
CA-UTI was defined as the presence of a positive
urine culture with a clinical diagnosis of UTI

documented in the electronic medical record within the
first 48 hours of hospitalization. Clinical diagnosis of UTI
was defined as a final diagnosis of UTI by the physician
in charge with at least 1 supportive finding on urinalysis
(leukocyte esterase, nitrate, pyuria of >5 WBCs/hpf, pres-
ence of bacteria in).
Cases with asymptomatic bacteriuria (in age-appropri-
ate patients) were excluded. Patients with positive urine
cultures obtained >48 hours after hospitalization, post-
operative UTI within 10 days of surgery, and history of
long-term care facility stay within the preceding
3 months were excluded.
Patients who met the criteria for CA-UTI with urine
culture positive for ESBL-producing E. coli were included.
A control group consisting of children with CA-UTI with
positive urine cultures for non-ESBL-producing E. coli
was identified. Patients in the control group were
matched by age, gender, and year of infection to
improve the statistical precision of the study by ensuring
roughly equal numbers of cases and controls in these
categories. Each patient was included once in the study.
If more than one UTI with ESBL-producing E. coli was
documented, the first clinical episode was included.
Medical records of patients with UTI caused by ESBL-
producing and non-ESBL producing E coli were manually
reviewed to obtain demographic characteristics, history
of hospital visits, clinical findings, culture results and
antimicrobial susceptibilities, laboratory and imaging
studies, comorbidities, hospital course, treatment modal-
ities, complications and outcome.
The US Census Bureau definition of ethnicity
(Hispanic/Latino or Non-Hispanic/Non-Latino) and race
(Caucasian, African American, Asian, American Indian,
Native Hawaiian or Other Pacific Islander) are individu-
ally poor categories for our patient population.
Therefore, in our study, patients were defined by the
category of ethnic background which further subdivided
non-Hispanic ethnicity by race into African American,
Caucasian and Middle Eastern. Middle Eastern ethnic
background was included separately in our study des-
pite its current definition as part of the Caucasian race
by US Census Bureau. Middle Eastern ethnic background
was defined as family origin from Middle Eastern coun-
tries such as Yemen, Iraq, Lebanon and Syria. The infor-
mation was extracted from medical records. This
distinction was made due to a desire to separately ana-
lyze a large number of patients of Middle Eastern des-
cent served by our institution.
Information collected was analyzed for potential risk
factors for ESBL-positive infection. A 3 month period was
INFECTIOUS DISEASES 3
used to categorize risks of prior healthcare exposure
(antibiotic usage, surgery, and hospitalization) as well as
recurrent UTI. Subcategories of prior b-lactam antibiotic
use (prior antibiotic usage) and intraurinary tract inter-
vention (prior surgery) were included and analyzed sep-
arately from their parent categories. Additional risk
factors analyzed were presence of genitourinary (GU)
tract abnormalities (defined as presence of hydronephro-
sis, ureteropelvic junction obstruction, duplex collection
system/kidney, or other congenital abnormalities of the
renal/GU tract), functional abnormalities (defined as
neurogenic bladder, voiding dysfunction, constipation),
presence/use of intraurinary tract device (defined as use
of clean intermittent catherization or intraurinary tract
stents), and presence of immunosuppression (defined as
use of any chronic immunosuppressive agent or chemo-
therapy regimen in the previous 3 months).
Statistical analysis
Data on different clinical variables and frequencies were
analyzed using SPSS version 20. A non-parametric
Fisher’s Exact test was employed to examine potential
differences in categorical variables between study
groups. Variables found significant in univariate analysis
were entered into a binary logistic regression equation
to find independent predictors of acquiring infection
with ESBL positive E. coli. A p value of <.05 was consid-
ered statistically significant.
Results
A total of 111 cases of community-acquired UTI caused
by ESBL-producing E. coli were identified for the period
2012-2016. A total of 103 control cases of community-
acquired UTI with non-ESBL producing E. coli were
matched by age, gender, and year in the same time
period. No matching cases were found for 8 of the ESBL
positive cases, resulting in a slight imbalance between
the groups.
When compared to total number of UTIs caused by
E. coli that were diagnosed during the study period, the
prevalence of infections with ESBL-producing E. coli var-
ied from 7% to 15% (27 (11%) in 2012, 18 (7%) in 2013,
22 (9%) in 2014, 15 (7%) in 2015, and 29 (15%) in 2016).
Patients with community-acquired UTI with ESBL pro-
ducing E. coli were predominantly female (84%) and had
a median age of 4 years with a range of 1 month to
18 years. The ethnic background of the ESBL-positive
group was predominantly African American (32%),
4 F. H. ZHU ET AL.

Table 1. Demographic information and laboratory values of children with ESBL-producing E. coli UTI compared
to controls.
ESBL-producing E. coli (n ¼ 111) Non ESBL-producing E. coli (n ¼ 103) p-value
Demographicsa
Median Age in years (range) 4 (1 mo-18 yrs, n ¼ 111) 4 (1 mo-18 yrs, n ¼ 103)
Female Gender n ¼ 92 (83%) n ¼ 91 (88%)
Ethnic background
Caucasian n ¼ 19 (17%) n ¼ 18 (17%)
African American n ¼ 35 (32%) n ¼ 50 (49%)
Hispanic n ¼ 17(15%) n ¼ 23 (22%)
Middle Eastern n ¼ 34 (31%) n ¼ 10 (10%)
Other n ¼ 4 (5%) n ¼ 2 (2%)
Laboratory value (mean)
WBC (103/mm3) 14.5 (n ¼ 88) 14.3 (n ¼ 58) .340
Hemoglobin (g/dL) 11.9 (n ¼ 88) 11.4 (n ¼ 58) .340
Hematocrit (%) 34.9 (n ¼ 88) 33.7 (n ¼ 58) .212
Platelets (103/ml) 318.5 (n ¼ 88) 318.2 (n ¼ 58) .775
BUN (mg/dL) 13.4 (n ¼ 63) 11.6 (n ¼ 46) .437
Creatinine (mg/dL) 0.5 (n ¼ 62) 1.6 (n ¼ 46) .028
CRP (mg/dL) 81.5 (n ¼ 23) 60.8 (n ¼ 20) .661
ESBL: Extended spectrum b-lactamase; CRP: C-reactive protein.
Denominators for percentages noted in header row of table.
a
Patients were case-controlled by age, sex, and year of UTI. No matches were found for 8 patients leading to a slightly smaller Non-
ESBL cohort.

followed by Middle Eastern (31%), Caucasian (17%), Table 2. Antimicrobial resistance of ESBL-producing E. coli and non
Hispanic (16%) and unidentified (4%). The ethnic back- ESBL-producing E. coli.
Resistance to ESBL-producing Non ESBL-producing
ground of the control group was predominantly African antimicrobial agents E. coli: n (%) E. coli: n (%)
American (49%), followed by Hispanic (22%), Caucasian Amikacin n¼0 n¼0
(17%), Middle Eastern (10%) and unidentified (2%). Ampicillin/Sulbactam n ¼ 13 (92%) n ¼ 16 (22%)
Ampicillin n ¼ 111 (100%) n ¼ 65 (63%)
Middle Eastern ethnic background was significantly Aztreonam n ¼ 30 (81%) n¼0
Cefazolin n ¼ 29 (100%) n¼0
more prevalent in the ESBL-positive group than in the Cefepime n ¼ 98 (88%) n¼0
control group (p < .001) (Table 1). Cefoxitin n ¼ 6 (86%) n¼0
Ceftriaxone n ¼ 111 (100%) n¼0
Laboratory values of ESBL producing and non-ESBL Ciprofloxacin n ¼ 68 (73%) n ¼ 5 (5%)
producing E. coli UTI groups are presented in Table 1. Ertapenem n ¼ 2 (2%) n¼0
Gentamicin n ¼ 39 (36%) n ¼ 5 (5%)
No significant difference was found in WBC, hemoglo- Imipenem n¼0 n¼0
Meropenem n¼0 n¼0
bin/hematocrit, platelet count, blood urea nitrogen Nitrofurantoin n ¼ 1 (1%) n¼0
(BUN), or C-reactive protein (CRP) at presentation. The Piperacillin/Tazobactam n ¼ 33 (75%) n ¼ 17 (17%)
Tobramycin n ¼ 43 (39%) n ¼ 2 (2%)
ESBL producing group had significantly lower mean Trimethoprim/Sulfamethoxazole n ¼ 80 (72%) n ¼ 26 (25%)
serum creatinine at presentation (0.48 mg/dl ESBL vs
1.57 mg/dl non-ESBL; p ¼ .028). Clinical presentation was
often resistant to gentamicin (36% ESBL vs 5% non-
excluded from analysis since a large proportion of the
ESBL), tobramycin (39% ESBL vs 2% non-ESBL), and
patients were too young to verbalize symptoms (ie dys- trimethoprim-sulfamethoxazole (72% ESBL vs 25% non-
uria, frequency, flank pain). However, there was no sig- ESBL) and ciprofloxacin (73% ESBL vs 5% non-ESBL).
nificant difference in the frequency of fever between the Univariate analysis of risk factors for community
two groups (59.6% ESBL vs 64.1% non-ESBL; p ¼ .572). acquired UTI with ESBL-producing E. coli is shown in
Antibiotic susceptibility of E. coli strains recovered Table 3. Patients in the ESBL group had increased preva-
from the two groups is shown in Table 2. ESBL-produc- lence of pre-existing medical conditions (46% ESBL vs
ing E. coli strains were universally susceptible to amika- 10% non-ESBL) with higher rates of myelomeningocele
cin, meropenem, and imipenem. Only two strains (2%) (10% ESBL vs 6% non-ESBL) and cerebral palsy (3% ESBL
in the ESBL group were resistant to ertapenem and one vs 0% non-ESBL). However, univariate analysis of risk fac-
(1%) to nitrofurantoin. Resistance to b-lactam antibiotics tors associated with these underlying medical conditions
was nearly universal: cefepime 88%, cefoxitin 86%, cef- did not show statistical significance (Table 3). These risk
triaxone 100%, cefazolin 100%. However, given the con- factors included functional abnormalities (neurogenic
firmed presence of ESBL in these isolates, these strains bladder, voiding dysfunction, concurrent neurogenic
were considered resistant in clinical practice despite bladder and voiding dysfunction, constipation), use of
in vitro sensitivity. ESBL-positive E. coli strains were more intraurinary device (clean intermittent catheterization,
 INFECTIOUS DISEASES 5

Table 3. Univariate analysis of risk factors for UTI with ESBL-producing E. coli compared to controls.
Risk factors ESBL-producing E. coli Non ESBL-producing E. coli p-value
VUR n ¼ 23 (20.9%) n ¼ 6 (5.9%) .002
Prior Antibiotic use (last 3 months) n ¼ 60 (54.5%) n ¼ 15 (14.6%) <.001
Prior b-lactam use (last 3 months) n ¼ 34 (30.9%) n ¼ 9 (8.8%) <.001
Prior UTI (last 3 months) n ¼ 26 (23.6%) n ¼ 6 (5.9%) <.001
Prior Hospitalization (last 3 months) n ¼ 27 (24.5%) n ¼ 10 (10.1%) .007
Prior Surgery (last 3 months) n ¼ 8 (7.3%) n ¼ 4 (3.9%) .378
Intraurinary tract intervention n ¼ 5 (4.6%) n ¼ 4 (3.9%) 1.000
Middle Eastern ethnic background n ¼ 34 (30.6%) n ¼ 10 (9.8%) <.001
GU Abnormalitiesa n ¼ 26 (22.2%) n ¼ 26 (30.8%) .182
Intraurinary tract deviceb n ¼ 13 (11.8%) n ¼ 7 (6.9%) .247
Functional abnormalities
Neurogenic Bladder n ¼ 7 (7.9%) n ¼ 1 (1.4%) .077
Voiding Dysfunction n ¼ 3 (3.5%) n ¼ 4 (5.4%) .706
Neurogenic Bladder and voiding dysfunction n ¼ 9 (9.9%) n ¼ 15 (12.5%) .632
Constipation n ¼ 10 (10.9%) n ¼ 17 (19.5%) .143
Immunosuppression n ¼ 9 (8.2%) n ¼ 5 (4.9%) .412
a
GU abnormalities included hydronephrosis, ureteropelvic junction obstruction, duplex collection system, duplex kidney, other renal/genito-
urinary tract abnormalities.
b
Intraurinary tract device included use of clean intermittent catheterization and presence of any intraurinary tract stents.

Table 4. Multivariate analysis of risk factors for ESBL-producing
E. coli UTI compared to controls.
Risk factor p-value Odds ratio
GU abnormalities .065 0.44
History of antibiotic .001 4.17
usage (last 3 months)
VUR .152 2.27
Prior UTI (last 3 months) .097 2.78
Middle Eastern ethnic background <.001 4.00
ureteral stent placement), presence of GU abnormalities,
and intraurinary tract intervention or prior surgery in the
last 3 months (excluding surgery in the last 10 days).
Additionally, immunosuppressed status was not signifi-
cantly associated. The only risk factor related to an
underlying medical condition significantly associated
was presence of vesicoureteral reflux (VUR) (20.9% ESBL
vs 6% non-ESBL; p ¼ .002). Further analysis of patients
with VUR revealed that 11 of 23 patients in the ESBL
group were on long term trimethoprim-sulfamethoxa-
zole prophylaxis vs 0 of 6 patients in the non-
ESBL group.
Significant risk factors found by univariate analysis
included: prior antibiotic usage in the last 3 months
(54.5% ESBL vs 14.6% non-ESBL, p < .001), prior b-lactam
use in the last 3 months (30.9% ESBL vs 8.8% non-ESBL;
p < .001), presence of VUR (20.9% VUR vs 5.9% non-
ESBL; p ¼ .002), prior UTI in the last 3 months (24% ESBL
vs 6% non-ESBL; p ¼ .002), prior hospitalization in last
3 months (24.5% ESBL vs 10.1% non-ESBL; p ¼ .007),
Middle Eastern ethnic background (30.6% ESBL vs 9.8%
non-ESBL; p < .001).
History of antibiotic usage in the last 3 months
(p ¼ .001, OR 4.17, 95% CI 1.85–9.09) and Middle Eastern
ethnic background (p < .001, OR 4.00, 95% CI 1.69–9.09)
remained significantly associated in multivariate analysis
(Table 4). Prior UTI (p ¼ .097), VUR (p ¼ .152), and pres-
95% CI for ence of GU abnormalities (p ¼ .065) were not signifi-
odds ratio
0.19–1.05
cantly associated.
1.85–9.09
0.74–7.14 Discussion
0.83–9.09
1.69–9.09 While infections with ESBL-producing E. coli have been
described as predominantly healthcare-associated since
the late 1980s [19], an increasing incidence of commu-
nity-acquired infections has been reported since the
turn of the century. Nearly half of ESBL-producing E. coli
strains isolated from a hospital in Seville, Spain between
2001 and 2002 were defined as community-acquired.
Risk factors for ESBL-production E. coli were diabetes
mellitus, recent use of fluoroquinolones, and hospital
admission in the prior year. [7]. In a recent survey of 5
centers across the United States during 2009–2010 (New
York City, Detroit, Pittsburg, Iowa City, San Antonio)
more than one-third of community-onset infections with
ESBL-producing E. coli were community-acquired [20].
There is no consensus on possible risk factors for
community acquired infections with ESBL-producing
E. coli in pediatric patients from the limited data avail-
able. A recent retrospective observational study from
2015–2016 of pediatric patients in Toledo, Spain pro-
posed male sex, hospitalization within last 30 days, VUR,
and urological pathology as possible risk factors [21]. In
a prospective study of 24 French medical centers Madhi
et al. found that the main risk factor that was associated
with febrile UTIs due to ESBL-positive Enterobacteriaceae
in children was recent antibiotic use, followed by previ-
ous hospitalization and travel history. The frequency of
isolation of ESBL-Enterobacteriaceae ranged from 0.8%
6 F. H. ZHU ET AL.
to 10% per year over the 3-year study period [14].
Underlying neurological diseases, developmental delay,
recurrent UTI, recent hospitalization, and prior antibiotic
use were proposed as risk factors in a study in Taiwan
in 2002–2005 [22]. In a study from 2004 to 2006 in
Turkey, Topaloglu et al. proposed underlying disease
and hospitalization, infection and antibiotic use, all in
the prior 3 months as possible risk factors for UTI with
ESBL-producing bacteria in children [13].
A possible explanation for these divergent risk factors
is the regional variation of the study populations. ESBL
prevalence varies by geographic location with the high-
est rates of >50% of E. coli isolates in South-East Asia
(China, India), followed by 25–50% in the Middle East
(Turkey, Egypt, Iraq, Iran), and 20–25% in the United
States and Western Europe (UK, France, Spain, Portugal)
[23]. The higher prevalence is reflected in the fecal col-
onization of healthy individuals with ESBL-producing
Enterobacteriaceae (46% in West Pacific, 22% in South-
East Asia, and 22% in Africa vs 4% in Europe and 2% in
the Americas) [24]. Additionally, international travel, par-
ticularly to India and South-East Asia, is a risk factor for
the development of community-acquired infection with
ESBL-producing E. coli [25].
The significantly higher proportion of patients of
Middle Eastern origin with ESBL-producing E. coli UTI
compared to other ethnic groups (Caucasian, African
American, and Hispanic) in our study suggests that eth-
nic background may be an important risk factor for the
development of community-acquired UTI with ESBL-pro-
ducing E. coli in children. The majority of these patients
resided in the city of Dearborn, MI, which has the largest
proportion of Arab Americans in the United States. The
increased risk in this group may reflect variations in diet-
ary habits, international travel, and fecal colonization.
Indeed, travel to the Middle East and North Africa
increased the risk of fecal colonization with ESBL-pro-
ducing Enterobacteriaceae [26]. However, fecal carriage
of ESBL-producing organisms acquired through travel
rarely persists for longer than 3 months and further
investigation into possible risk factors which may con-
tribute to a higher risk of community-acquired UTI with
ESBL-producing E. coli in this ethnic group is warranted
[27]. Additionally, Ravensbergen et al have shown that
asylum seekers originating from Syria, Iraq, Afghanistan,
and Eritrea in the Netherlands have a 20.3% carriage
rate of multidrug-resistant Enterobacteriaceae (predomin-
antly ESBL) [28]. Dearborn, MI, has recently had an influx
of refugees from similar war-torn regions of the Middle
East, which may have resulted in an increased carriage
rate of ESBL Enterobacteriaceae in the city and in the
surrounding region resulting in increased prevalence of
UTI caused by ESBL-producing E. coli in the predominant
ethnic group in the area.
The second significant risk factor found in multivari-
ate analysis was prior use of antibiotics within the last
3 months. Additionally, the subset of prior b-lactam anti-
biotic use increased the risk of UTI with ESBL-producing
E. coli in univariate analysis. Prior antibiotic use has been
a proposed risk factor in prior studies in both pediatric
and adult patients [8,9,13,22,29]. Prolonged antibiotic
use increased the prevalence of Enterobacteriaceae
resistant to ampicillin and trimethoprim-sulfamethoxa-
zole in the gastrointestinal tract of patients with cystic
fibrosis [30]. Therefore, prior exposure to antibiotics, par-
ticularly b-lactam antibiotics, may select ESBL-producing
E. coli in the gastrointestinal tract. As intestinal coloniza-
tion generally precedes the onset of infection by gram-
negative pathogens, the increased colonization by ESBL-
producing E. coli in the gastrointestinal tract may lead
to increased risk of UTI with ESBL-producing E. coli [31].
Prior UTI in the last 3 months, hospitalization in the
last 3 months, and the presence of VUR were significant
risk factors in univariate analysis but not in multivariate
analysis. These risk factors are likely associated with prior
antibiotic use thereby lowering their significance in
multivariate analysis. Patients with prior UTI and a sig-
nificant proportion of previously hospitalized patients
would have been treated with antibiotics. The presence
of VUR increases exposure to antibiotics through
increased risk for UTI and the use of antibiotic prophy-
laxis. Indeed, 11 of 60 patients exposed to prior anti-
biotic use in the ESBL group were receiving
trimethoprim-sulfamethoxazole prophylaxis for VUR.
As the genes for ESBL and resistance genes for other
classes of antibiotics are often encoded together on
transferable plasmids, ESBL producing organisms are
often the multidrug-resistant (MDR) [32]. The higher
rates of co-resistance to non-b-lactam antibiotics (tri-
methoprim-sulfamethoxazole, aminoglycosides, and qui-
nolones) in the ESBL E. coli isolates when compared to
non-ESBL isolates is consistent with previously observed
resistance patterns in ESBL-producing organisms [19,33].
High resistance rates in ESBL-producing E. coli to non-
b-lactam antibiotics have been reported throughout the
world including Iran, Turkey, Spain, and Israel [8,9,34,35].
Interestingly, the pattern of preserved sensitivity to ami-
kacin in our ESBL-producing E. coli strains was also seen
in ESBL-producing Enterobacteriaceae in France [14].
Exposure to non b-lactam classes of antibiotics,

including trimethoprim-sulfamethoxazole used in the
prophylaxis of VUR, can select ESBL-producing E. coli in
the gastrointestinal tract. Judicious use of antibiotics in
the community is necessary to stem the increasing
prevalence of community acquired ESBL-producing
E. coli strains.
While underlying medical conditions, including neuro-
logical conditions, were suggested as possible risk fac-
tors in prior studies [21], no such association was found
in our study in univariate analysis. There was a prepon-
derance of myelomeningocele in the ESBL group (10%
ESBL vs 5.9% non-ESBL). However, neurological abnor-
malities associated with myelomeningocele (neurogenic
bladder, voiding dysfunction, concurrent neurogenic
bladder and voiding dysfunction, use of intraurinary
device such as clean intermittent catheterization or
ureteral stent placement) were not significantly associ-
ated in univariate analysis.
The optimal antibiotic therapy of ESBL-producing
E. coli UTI is not clear. Our data suggest that the drug of
choice in patients who require parenteral therapy is a
carbapenem or amikacin. Potential oral options include
trimethroprim-sulfamethoxazole or ciprofloxacin if the
isolate is susceptible.
The primary limitation of this study is the lack of
detailed patient history due to the retrospective nature
of the study. Upper and lower UTI could not be distin-
guished due to inconsistent documentation of symp-
toms, history, and physical exam. Indications for
antibiotic prophylaxis in patients with VUR (11/23 in
ESBL group vs 0/6 in non-ESBL group) were unclear as
both groups had a wide range of VUR grades (1–5).
Finally, only limited background details were available
from patients with Middle Eastern ethnic background
allowing only conjectural analysis of the root cause of
this risk factor.
Conclusion
Currently, there is limited data available on possible risk
factors for the development of community acquired
ESBL-producing E. coli UTI in the United States. In our
study, prior antibiotic use in the last 3 months and
Middle Eastern ethnic background were significant risk
factors in multivariate analysis. Prior antibiotic use as a
risk factor reinforces the need for judicious use of antibi-
otics. As 18% of patients with prior antibiotic use in the
ESBL group were on long-term trimethoprim-sulfameth-
oxazole prophylaxis for VUR, further investigation of the
impact of antibiotic prophylaxis on the prevalence of
INFECTIOUS DISEASES 7
UTIs with ESBL-producing E. coli are warranted. Indeed,
this practice may paradoxically increase the prevalence
of UTIs with ESBL-producing E. coli in a patient popula-
tion already prone to recurrent UTIs. Additionally, the
increased risk of community-acquired ESBL-producing
E. coli among children of Middle Eastern ethnic back-
ground suggests that a history of international travel or
of contact with international travelers should be
explored when evaluating a child for possible commu-
nity-acquired ESBL infection.
Disclosure statement
No potential conflict of interest was reported by the authors.
References
[1] Bradford PA. Extended-spectrum beta-lactamases in the
21st century: characterization, epidemiology, and detection
of this important resistance threat. Clin Microbiol Rev.
2001;14:933–951.
[2] Pitout JD, Laupland KB. Extended-spectrum beta-lactamase-
producing Enterobacteriaceae: an emerging public-health
concern. Lancet Infect Dis. 2008;8:159–166.
[3] Arslan H, Azap OK, Ergonul O, et al. Risk factors for cipro-
floxacin resistance among Escherichia coli strains isolated
from community-acquired urinary tract infections in Turkey.
J Antimicrob Chemother. 2005;56:914–918.
[4] Doi Y, Adams J, O’Keefe A, et al. Community-acquired
extended-spectrum beta-lactamase producers, United
States. Emerg Infect Dis. 2007;13:1121–1123.
[5] Meier S, Weber R, Zbinden R, et al. Extended-spectrum
beta-lactamase-producing Gram-negative pathogens in
community-acquired urinary tract infections: an increasing
challenge for antimicrobial therapy. Infection. 2011;39:
333–340.
[6] Pitout JD, Nordmann P, Laupland KB, et al. Emergence
of Enterobacteriaceae producing extended-spectrum beta-
lactamases (ESBLs) in the community. J Antimicrob
Chemother. 2005;56:52–59.
[7] Rodriguez-Bano J, Navarro MD, Romero L, et al.
Epidemiology and clinical features of infections caused by
extended-spectrum beta-lactamase-producing Escherichia
coli in non-hospitalized patients. J Clin Microbiol. 2004;42:
1089–1094.
[8] Colodner R, Rock W, Chazan B, et al. Risk factors for the
development of extended-spectrum beta-lactamase-pro-
ducing bacteria in nonhospitalized patients. Eur J Clin
Microbiol Infect Dis. 2004;23:163–167.
[9] Calbo E, Romani V, Xercavins M, et al. Risk factors for com-
munity-onset urinary tract infections due to Escherichia coli
harbouring extended-spectrum beta-lactamases. J
Antimicrob Chemother. 2006;57:780–783.
[10] Goyal D, Dean N, Neill S, et al. Risk factors for community-
acquired extended-spectrum beta-lactamase-producing
Enterobacteriaceae infections-a retrospective study of
8 F. H. ZHU ET AL.
symptomatic urinary tract infections. Open Forum Infect
Dis. 2019;6:ofy357.
[11] Zaoutis TE, Goyal M, Chu JH, et al. Risk factors for and out-
comes of bloodstream infection caused by extended-spec-
trum beta-lactamase-producing Escherichia coli and
Klebsiella species in children. Pediatrics. 2005;115:942–949.
[12] Kuo KC, Shen YH, Hwang KP. Clinical implications and risk
factors of extended-spectrum beta-lactamase-producing
Klebsiella pneumoniae infection in children: a case-control
retrospective study in a medical center in southern Taiwan.
J Microbiol Immunol Infect. 2007;40:248–254.
[13] Topaloglu R, Er I, Dogan BG, et al. Risk factors in commu-
nity-acquired urinary tract infections caused by ESBL-pro-
ducing bacteria in children. Pediatr Nephrol. 2010;25:
919–925.
[14] Madhi F, Jung C, Timsit S, et al. Febrile urinary-tract infec-
tion due to extended-spectrum beta-lactamase-producing
Enterobacteriaceae in children: a French prospective multi-
center study. PloS one. 2018;13:e0190910.
[15] Kizilca O, Siraneci R, Yilmaz A, et al. Risk factors for commu-
nity-acquired urinary tract infection caused by ESBL-pro-
ducing bacteria in children. Pediatr Int. 2012;54:858–862.
[16] Uyar Aksu N, Ekinci Z, Dundar D, et al. Childhood urinary
tract infection caused by extended-spectrum beta-lacta-
mase-producing bacteria: Risk factors and empiric therapy.
Pediatrics Int. 2017;59:176–180.
[17] Logan LK, Braykov NP, Weinstein RA, et al. Extended-
Spectrum beta-lactamase-producing and third-generation
cephalosporin-resistant Enterobacteriaceae in children:
trends in the United States, 1999-2011. J Pediatric Infect
Dis Soc. 2014;3:320–328.
[18] BD Phoenix TM NMIC/ID Panel [Package Insert]. Becton,
Sparks, MD, USA: Dickinson and Company; 2009.
[19] Paterson DL, Bonomo RA. Extended-spectrum beta-lacta-
mases: a clinical update. Clin Microbiol Rev. 2005;18:
657–686.
[20] Doi Y, Park YS, Rivera JI, et al. Community-associated
extended-spectrum beta-lactamase-producing Escherichia
coli infection in the United States. Clin Infect Dis. 2013;56:
641.
[21] Perez Heras I, Sanchez-Gomez JC, Beneyto-Martin P, et al.
Community-onset extended-spectrum b-lactamase produc-
ing Escherichia coli in urinary tract infections in children
from 2015 to 2016: prevalence, risk factors, and resistances
. Medicine (Baltimore).. 2017;96:e8571.
[22] Fan NC, Chen HH, Chen CL, et al. Rise of community-onset
urinary tract infection caused by extended-spectrum
beta-lactamase-producing Escherichia coli in children.
J Microbiol Immunol Infect. 2014;47:399–405.
[23] Doi Y, Iovleva A, Bonomo RA. The ecology of extended-
spectrum b-lactamases (ESBLs) in the developed world.
J Travel Med. 2017;24:S44–S51.
[24] Karanika S, Karantanos T, Arvanitis M, et al. Fecal coloniza-
tion with extended-spectrum beta-lactamase-producing
Enterobacteriaceae and risk factors among healthy individ-
uals: a systematic review and meta-analysis. Clin Infect Dis.
2016;63:310–318.
[25] Kantele A, Laaveri T, Mero S, et al. Antimicrobials increase
travelers’ risk of colonization by extended-spectrum beta-
lactamase-producing Enterobacteriaceae. Clin Infect Dis.
2015;60:837–846.
[26] Tangden T, Cars O, Melhus A, et al. Foreign travel is a
major risk factor for colonization with Escherichia coli pro-
ducing CTX-M-type extended-spectrum beta-lactamases: a
prospective study with Swedish volunteers. Antimicrob
Agents Chemother. 2010;54:3564–3568.
[27] Ruppe E, Armand-Lefevre L, Estellat C, et al. High rate of
acquisition but short duration of carriage of multidrug-
resistant Enterobacteriaceae after travel to the tropics. Clin
Infect Dis. 2015;61:593–600.
[28] Ravensbergen SJ, Berends M, Stienstra Y, et al. High preva-
lence of MRSA and ESBL among asylum seekers in the
Netherlands. PLoS One. 2017;12:e0176481.
[29] Ena J, Arjona F, Martinez-Peinado C, et al. Epidemiology of
urinary tract infections caused by extended-spectrum beta-
lactamase-producing Escherichia coli. Urology. 2006;68:
1169–1174.
[30] Knudsen PK, Brandtzaeg P, Hoiby EA, et al. Impact of
extensive antibiotic treatment on faecal carriage of anti-
biotic-resistant Enterobacteria in children in a low resist-
ance prevalence setting. PloS one. 2017;12:e0187618.
[31] Donskey CJ. Antibiotic regimens and intestinal colonization
with antibiotic-resistant gram-negative bacilli. Clin Infect
Dis. 2006;43:S62–S69.
[32] Carattoli A. Resistance plasmid families in
Enterobacteriaceae. Antimicrob Agents Chemother. 2009;
53:2227–2238.
[33] Lob SH, Nicolle LE, Hoban DJ, et al. Susceptibility patterns
and ESBL rates of Escherichia coli from urinary tract infec-
tions in Canada and the United States, SMART 2010-2014.
Diagn Microbiol Infect Dis. 2016;85:459–465.
[34] Soltani R, Ehsanpoor M, Khorvash F, et al. Antimicrobial
susceptibility pattern of extended-spectrum beta-lacta-
mase-producing bacteria causing nosocomial urinary tract
infections in an Iranian referral teaching hospital. J Res
Pharm Pract. 2014;3:6–11.
[35] € Arslan H, Şerefhanog
Azap OK, lu K, et al. Risk factors for
extended-spectrum beta-lactamase positivity in uropatho-
genic Escherichia coli isolated from community-acquired
urinary tract infections. Clin Microbiol Infect. 2010;16:
147–151.