Pediatrics and Neonatology (2020) 61, 195e200

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: http://www.pediatr-neonatol.com

Original Article

Diagnosis and imaging of neonatal UTIs

Laura Walawender a, David S. Hains b, Andrew L. Schwaderer b,*

a
Nationwide Children’s Hospital, Department of Pediatrics, United States
b
Indiana University, Department of Pediatrics, Division of Nephrology, United States

Received Apr 16, 2019; received in revised form Jul 30, 2019; accepted Oct 21, 2019
Available online 5 November 2019

Key Words Background: The 2011 American Academy of Pediatrics clinical practice guideline recommends
cystitis; when to obtain renal and bladder ultrasound (RBUS) and voiding cystourethrography (VCUG)
cystourethrogram; following febrile urinary tract infection (UTI) for children age 2e24 months. However, there
ultrasound; is not consensus about when to obtain imaging in neonates. The objective of this study is to
urinalysis evaluate UTI diagnostic criteria along with RBUS and VCUG in neonates admitted to the NICU
in the first 3 months of life.
Methods: A retrospective electronic medical record review was performed of neonates
admitted to Nationwide Children’s Hospital system NICUs between January 2010 and December
2014 with UTI as a primary or secondary diagnosis. Urine culture results were evaluated versus
established UTI criteria and renal US and VCUG results were compared.
Results: Of 81 patients with a straight catheterized urine culture obtained, 28 patients met
laboratory criteria for diagnosis of UTI and all but 4 had a RBUS. Urine cultures had an equal
distribution of Enterobacter cloacae, Escherichia coli, Klebsiella pneumoniae, and Coagulase
negative staphylococcus. RBUS showed dilation of the collecting system in 37.5% of patients
with UTI compared to 41.3% without UTI. VCUG showed vesicourteral reflux (VUR) on 41.7%
of those with UTI compared to 34.8% without UTI. For patients with UTI, the sensitivity of RBUS
for VUR on VCUG was 60% with CI [0.17, 0.93] and specificity was 43% with CI [0.12, 0.80]. In
patients without UTI, sensitivity of RBUS for VUR on VCUG was 63% with CI [0.26, 0.90] and
specificity was 71% with CI [0.42, 0.90].
Conclusions: Fewer than half of neonates that were diagnosed clinically with UTI met labora-
tory criteria for a UTI. Positive urine cultures grew a wide variety of organisms. The sensitivity
of renal ultrasound for VUR is only about 60%.
Copyright ª 2019, Taiwan Pediatric Association. Published by Elsevier Taiwan LLC. This is an
open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/
by-nc-nd/4.0/).

* Corresponding author. 699 Riley Hospital Dr, Indianapolis, IN 46201, United States. Fax: þ1 317 278.3599.
E-mail address: Schwadea@iu.edu (A.L. Schwaderer).

https://doi.org/10.1016/j.pedneo.2019.10.003
1875-9572/Copyright ª 2019, Taiwan Pediatric Association. Published by Elsevier Taiwan LLC. This is an open access article under the CC BY-
NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
196
1. Background
The 2011 American Academy of Pediatrics clinical practice
guideline applies to infants age 2e24 months with febrile
urinary tract infection (UTI) and recommends when pro-
viders should obtain a renal and bladder ultrasound (RBUS)
and a voiding cystourethrography (VCUG).1 Given unique
characteristics of infants less than 2 months of age, these
aforementioned guidelines may not be applicable to neo-
nates.1 Currently, practice varies among practitioners as to
when imaging is obtained in neonatal UTIs.
Studies show that diagnosis of UTI in neonates is diffi-
cult. The prevalence of UTIs in term newborns ranges from
0.1% to 1% and is higher in premature neonates, ranging
from 4% to 25%.2,3 In children age 2e24 months, a UTI is
defined as pyuria on urinalysis and a urine culture with
>50,000 colony forming units/mL (CFU/mL) of a single or-
ganism.1 Neonates void frequently, which can lead to lower
colony counts on urine cultures; therefore, 10,000 CFU/mL
is sometimes used as a cutoff for the diagnosis of neonatal
UTI.2 However, there is not consensus about what estab-
lishes a diagnosis of UTI in neonates or when to obtain a
RBUS and/or a VCUG.2,4 Studies show various utility of RBUS
in detecting dilation of the collecting system which can be
seen secondary to vesicoureteral reflux (VUR). A 2015 study
showed RBUS had 86.7% sensitivity for detecting grades
IVV VUR but only 32.7% for detecting all grades of VUR.5
Another study from 2015 found no relationship between
renal pelvic dilation on RBUS and VUR.2
Key gaps in the care of neonates include (a) how to di-
agnose UTI and (b) what imaging modalities should be used
to screen for structural kidney and urinary tract anomalies
following the diagnosis of UTI. The objectives of this study
were to determine how UTIs were diagnosed in neonates
and evaluate the utility of RBUS and VCUG in neonates
admitted to the NICU with a clinically diagnosed UTI in the
first 3 months of life.
2. Methods
A retrospective electronic medical record review was per-
formed of neonates admitted to Nationwide Children’s
Hospital (NCH) system NICUs between January 2010 and
December 2014. Infants with an ICD9 code 599.0 (urinary
tract infection) as a primary or secondary diagnosis were
included. Patients were excluded if there was no evidence
of a catheterized urine sample sent for culture in the first 3
months of life or the urine sample was obtained at another
institution.
All patients had been clinically diagnosed with a UTI.
However, the exact reasons for this diagnosis were not al-
ways clear from chart review. Therefore, this study defined
laboratory criteria for UTI diagnosis as either: (a) a urine
culture with 50,000 CFU/mL or (b) a urine culture with
10,000 to 49,999 CFU/mL and positive leukocyte esterase or
nitrites on urinalysis. This definition was used because of
the current clinical practice guidelines and the fact that
some NICUs will use a minimum of 10,000 CFU/mL on urine
culture when UA findings also indicate UTI.1,2 Patients were
then categorized as either having a UTI or not based upon
these laboratory criteria.
L. Walawender et al
Gestational age at birth and infant sex were obtained.
Infant age on the day of urine sample collection was
recorded. From urine cultures, the organism(s) and CFU/mL
were recorded. For patients with UTI that met laboratory
criteria, only organisms that grew 10,000 CFU/mL were
included. If multiple organisms grew only organisms that
grew 10,000 CFU/mL were included. For patients that did
not meet UTI laboratory criteria, all organisms that grew
were included. For this group, the organisms were divided
into those that grew <10,000 CFU/mL and those that grew
10,000 CFU/mL. If urinalysis was obtained, evidence of
small, moderate, or large leukocyte esterase and positive
nitrites were noted. Charts were reviewed for RBUS within
4 months of the urine sample and VCUG within 6 months.
Findings of hydronephrosis, collecting system dilation, or
urothelial thickening were recorded from RBUS reports.
Vesicoureteral reflux and grade were recorded from VCUG
reports.
Continuous data was compared using the t-test if para-
metric or ManneWhitney test if non parametric using
GraphPad Prism (SanDiego CA). The chi squared or, if the
expected cell was <5, the Fisher exact test, compared
percentages or proportions. The sensitivity and specificity
of RBUS to predict VUR on VCUG was analyzed with a 95%
confidence interval (CI). Vassarstats was used for the Chi
square and Fisher exact test analysis and to calculate
sensitivities, specificities and positive predictive values.6
3. Results
3.1. Patients
One hundred and nine neonates were identified by the ICD9
code, of whom 28 were excluded. Eighty-one patients, 51
males and 30 females were included. The mean gestation
age of included neonates was 212  36 days. The mean day
of life at the time of urine collection was 49  22 days.
3.2. Urine culture results
For all urine samples obtained, 20 (25%) cultures exhibited
no growth, 36 (44%) grew < 50,000 CFU/mL, 4 (5%) grew
50,000e100,000 CFU/mL, and 21 (26%) grew >100,000 CFU/
mL. Only 46 (57%) patients had urinalysis sent. Of these,
15 (33%) were leukocyte esterase positive and 4 (9%) were
nitrite positive. Twenty-eight (35%) patients met laboratory
criteria for diagnosis of UTI and 53 (65%) did not meet UTI
criteria (Fig. 1). There were not significant differences in
the demographics in neonates that met UTI criteria versus
those who did not (Table 1).
3.3. Organisms
A wide variety of organisms were identified. Gram-negative
organisms were most prominent, accounting for 23 (74%)
and 36 (80%) of the organisms that grew in urine cultures in
those that did and did not meet UTI criteria, respectively.
Urine cultures from the 28 (35%) patients that met lab
criteria for UTI grew 16 organisms. Multiple organisms grew
in 7 (25%) patients in the UTI group and 10 (19%) patients in
Neonatal UTI Diagnosis 197

Figure 1 Flow chart of UTI diagnosis. White boxes indicate no UTI present, light grey boxed indicate indeterminate UTI status,
dark grey boxes indicate that UTI criteria was met and the black box indicates excluded patients.

the non-UTI group. Enterobacter cloacae, E. coli, Klebsiella
pneumoniae, and Coagulase negative staphylococcus each
grew in 13% of cultures. Regarding the 53 (65%) patients
that did not meet laboratory criteria for UTI, 13 organisms
were identified on cultures that grew at less than
10,000 CFU/mL and 10 different organisms were
identified on cultures that grew greater than or equal to
10,000 CFU/mL with Enterococcus faecalis being the most
common organism in each group. There were no significant
differences in the types of organisms that grew in cultures
that met UTI criteria versus those that did not (Table 2).
3.4. Imaging workup and results
RBUS was obtained within 4 months of the infection in 24
(86%) patients that met laboratory criteria for UTI and 45
(85%) of those that did not. A VCUG was obtained within 6
months of the infection in 12 (43%) patients that met lab-
oratory criteria for UTI and 23 (43%) of those that did not.
RBUS showed hydronephrosis or dilation of the collecting
system in 9 (37.5%) of patients with UTI compared to 18
(40.0%) without UTI. VCUG showed reflux in 5 (41.7%) of
those with UTI compared to 8 (34.8%) without UTI.

Table 1 Demographics.
Met UTI criteria n Z 28 Did not meet UTI criteria n Z 53 P value
a
Gestational age (median days 25% and 75%) 207 (183, 257) 203 (179, 244) 0.48
Mean day of life at urine culture 45  21 51  22 0.26
collection  standard deviationb
M/F 17:11 (1.5:1) 34:19 (1.8:1) 0.76
a
Data non parametric, analyzed with ManneWhitney test.
b
Data parametric, analyzed with t test.
198 L. Walawender et al

Table 2 Culture results.

Organism UTI criteria meta UTI criteria not metb P value
Number of cultures % Number of cultures %
Bacillus species 0 0 1 2.2 1.00
Beta streptococcus group B 1 3.2 1 2.2 1.00
Candida 0 0 1 2.2 1.00
Citrobacter freundii 2 6.5 0 0 0.16
Citrobactera koseri 1 3.2 0 0 0.40
Coag negative staph, not saprophyticus 4 12.9 2 4.4 0.22
Coag negative staph 0 0 1 2.2 1.00
E. coli 4 12.9 5 11.1 1.00
Enterobacter aerogenes 0 0 2 4.4 0.51
Enterobacter cloacae 4 12.9 1 2.2 0.15
Enterococcus faecalis 3 9.7 10 22.2 0.21
Gardnerella vaginalis 1 3.2 0 0 0.40
Klebsiella oxytoca 1 3.2 6 13.3 0.22
Klebsiella pneumoniae 4 12.9 4 8.9 0.71
Lactose fermenting gram negative rod 0 0 1 2.2 0.51
Pantoea species 1 3.2 0 0 0.40
Proteus mirabilis 1 3.2 3 6.7 0.64
Pseudomonas aeruginosa 1 3.2 3 6.7 0.64
Serratia marcescens 1 3.2 1 2.2 1.00
Staph aureus 1 3.2 2 4.4 1.00
Viridans group Streptococcus 1 3.2 1 2.2 1.00
a
n Z 31 cultures with growth from 28 patients.
b
n Z 45 cultures with growth from 33 patients (33 of 53 patients that did not meet UTI criteria had some growth on urine culture).

Of the patients that met laboratory criteria for UTI,
RBUS showed hydronephrosis or dilation of the collecting
system in nine (37.5%) patients. Seven of these patients had
VCUG obtained and three showed VUR. Five of the patients
without hydronephrosis or dilation on the RBUS had VCUG
obtained and two showed VUR. For the patients that did not
meet laboratory criteria for UTI, RBUS showed hydro-
nephrosis or dilation of the collecting system in 18 (40.0%)
patients. Nine of these patients had VCUG obtained, and
five showed VUR. Thirteen of the patients without hydro-
nephrosis or dilation on the RBUS had VCUG obtained and
three showed VUR. All VUR identified was grade II or III
except for 1 patient that had grade IV. A summary of the
imaging results is presented in Table 3.
For patients with UTI, the sensitivity of RBUS for VUR on
VCUG was 60% with CI [0.17, 0.93] and specificity was 43%
with CI [0.12, 0.80]. In patients without UTI, sensitivity of
RBUS for VUR on VCUG was 63% with CI [0.26, 0.90] and
specificity was 71% with CI [0.42, 0.90].
3.5. Source data
Source data is presented as supplemental file S1.
4. Discussion
In this study we evaluate urine culture results and imaging
results in a neonatal population. We identified that a wide
range of organisms grow in neonatal urine cultures, that
children who are evaluated for UTI often do not meet
published threshold of CFU/mL to be considered positive,
and that sensitivity of renal ultrasound results to identify
VUR is only w60%.

Table 3 Imaging results.

Imaging modality and results Met UTI criteria Did not meet UTI criteria p-value
RBUS performed 24/28 (86%) 45/53 (85%) 1.00
Echogenic medullary pyramidsa or echogenic focib 4/28 (17%) 15/53 (28%) 0.16
HN on RBUS 9/24 (38%) 18/45 (40%) 0.84
VCUG performed 12/28 (42%) 23c/53 (43%) 1.00
VUR when RBUS (þ) for HN 3/7 (43%) 5/9 (56%) 0.99
VUR when RBUS () for HN 2/5 (40%) 3/13 (23%) 0.58
a
Echogenic medullary pyramids, Tamm Horsfall proteinuria vs. nephrocalcinosis.
b
Nephrocalcinosis, stones, calculi or possible fungal ball.
c
One patient had VCUG performed but not RBUS.
Neonatal UTI Diagnosis
We used a laboratory definition of UTI that included not
only patients with 50,000 CFU/mL or greater on urine cul-
ture but also those patients with 10,000e49,999 CFU/mL if
there was positive (“small”or greater) leukocyte esterase
or nitrites on urinalysis. Even with this less stringent defi-
nition, fewer than half of the patients clinically diagnosed
with UTI actually met lab criteria. There is a wide variety of
definitions currently being used to define a UTI by lab
criteria in the neonatal population. In a study by Weems
et al., UTI was defined as a positive urine culture with
colony counts ranging from 100 to >100,000 CFU/mL.2 A
study by Foglia and Lorch defined a positive culture as at
least 1000 CFU/mL of a single organism or at least
10,000 CFU/mL of 2 organisms.4 This variety of definitions
in conjunction with the large percentage of patients in our
study that did not meet lab criteria but were clinically
diagnosed with UTI exemplifies the need for better UTI
criteria in this unique population.
Only w13% of cultures grew E. coli in the neonates in
this study. In contrast E. coli grows in >80% of urine cul-
tures in older children.7 Sequencing technology and
expanded culture techniques have demonstrated that the
urine microbiota at baseline and during disease states is
more diverse than what is identified by clinical
cultures.8e10 Criteria for largely E. coli pediatric UTIs may
not be applicable to primarily non-E. coli neonatal UTIs.
Also the organisms were similar between patients that met
criteria for UTI and those that did not (Table 2). Thus it
does not appear that organisms considered “uropatho-
genic” are highly prevalent in the results that met UTI
criteria, nor are organisms considered “contaminants”
highly prevalent in results that did not meet UTI criteria.
We submit that UTI diagnostic criteria should include a
lower CFU/mL threshold and enhanced cultures may war-
rant evaluation in the neonatal population.11
Renal ultrasound was obtained fairly consistently in
patients clinically determined to have UTI. Of patients
that met laboratory criteria for UTI, 86% were evaluated
with a renal ultrasound, which is similar to those that did
not meet lab criteria at 85%. Interestingly, the patients
that did not meet lab criteria but had renal imaging ob-
tained had higher percentage of hydronephrosis or dilation
on renal ultrasound (40.0%) compared to those that met
lab criteria (37.5%). However, VCUG to further evaluate
abnormal RBUS findings was inconsistently obtained (59%),
but patients that met lab criteria for UTI were more
frequently evaluated with VCUG (78%). VUR was found
more frequently in those that did not meet lab criteria,
55% compared to 43% of those that met lab criteria, but
this difference was not statistically significant. In patients
with a normal RBUS that also had a VCUG, VUR was found
in 40% of patients that met lab criteria compared to 23% of
those that did not meet criteria. Wallace et al., note that
few studies demonstrate the accuracy of detecting VUR
with RBUS in neonates, and those that exist suggest poor
sensitivity for low grade VUR with moderate to good
sensitivity for higher grade VUR but cases can still be
missed. In the study by Wallace et al., the sensitivity of
RBUS for VUR in infants younger than 2 months of age that
were diagnosed with UTI was evaluated and found to be
33% for grade IeV VUR but 86% when limited to grade
IVV VUR.5
199
The sensitivity of RBUS for VUR is only w60%. RBUS was
not very sensitive or specific for VUR in either group of pa-
tients. Given that the majority of the VUR identified in our
patients was low grade and similar rates of VUR were iden-
tified in all groups except for patients that did not meet
laboratory criteria for UTI and had normal RBUS, it is
reasonable to suggest that all neonates with clinically diag-
nosed UTI should have RBUS obtained. Moreover, given the
low sensitivity and specificity of RBUS for VUR, it is reason-
able to also obtain VCUG on all neonates with UTI. Effec-
tively, RBUS is being utilized to grossly evaluate the structure
of the kidneys and bladder and VCUG to assess for VUR.
Lastly, there was a wide variety of organisms that were
identified on cultures. A retrospective chart review
completed by Clarke et al. looked at urine cultures of infants
admitted to the NICU and major organisms identified were E.
coli, C. negative staphylococci, and Klebsiella species.12 We
identified similar organisms with E. faecalis being the most
frequently identified. E. coli is prevalent in older patients
and accounts for 85e90% of infections.13 This was not the
most frequent organism identified in our study. It is likely
that the incompletely developed neonatal immune system
has not only increased their risk of infection but makes them
susceptible to additional organisms compared to older chil-
dren.13 A better understanding of which organisms
commonly cause neonatal UTIs would help to further define
guidelines for when a neonate should be diagnosed with a
UTI and require renal imaging. This study has limitations as a
retrospective chart review and due to the number of pa-
tients that received a VCUG. Future directions consist of
optimizing neonatal criteria for UTI diagnosis. Additionally,
there needs to be long-term follow up of renal function and
infectious outcomes in neonates that had VCUG compared to
those that did not. Up to a quarter of cultures grew multiple
organisms. Because our cultures were obtained by cathe-
terization, as long as growth of at least one organism met UTI
criteria it was considered a UTI. Whether presence of mul-
tiple organisms in catheter-obtained NICU samples repre-
sents contaminants should be addressed in future studies
and NICU UTI diagnostic guidelines. We chose the 4-and 6-
month time periods to include US and VCUG respectively
because studies past this time period might be due to a new
problem. It is possible that we missed some imaging studies
related to the urine culture that were performed at very late
time periods.
5. Conclusions
Overall our findings regarding the sensitivity of RBUS to
identify VUR are similar to the limited previously published
studies of UTIs in the NICU and neonatal population.2,5 We
demonstrated that a wide range of bacteria are responsible
for neonatal UTIs. The organisms that grew in cultures were
not different between the patients that met UTI criteria and
those that did not. This suggests that the cultures that did
not meet UTI criteria may not be contaminants but lower
levels of growth of infecting organisms. This study further
highlights the wide variation of clinical practice when diag-
nosing UTIs and obtaining renal imaging in these patients.
Ultimately, this reinforces the need for clinical guidelines
specific to this unique patient population to be developed
200
because extrapolating the clinical practice guidelines for
infants age 2e24 months would likely underdiagnose UTIs in
neonates and lead to insufficient renal imaging.
Declaration of Competing Interest
ALS has consulted for Allena Pharmaceuticals on a topic
unrelated to this manuscript. Otherwise, the authors have
no conflicts of interests relevant to this article.
Acknowledgements
ALS and DSH received support from Lilly Endowment, Inc.
Physician Scientist Initiative. ALS also received support
from Nationwide Children Hospital internal funds.
References
1. Subcommittee on Urinary Tract Infection, Steering Committee
on Quality Improvement and Management, Roberts KB. Urinary
tract infection: clinical practice guideline for the diagnosis and
management of the initial UTI in febrile infants and children 2
to 24 months. Pediatrics 2011;128:595e610.
2. Weems MF, Wei D, Ramanathan R, Barton L, Vachon L,
Sardesai S. Urinary tract infections in a neonatal intensive care
unit. Am J Perinatol 2015;32:695e702.
3. Bauer S, Eliakim A, Pomeranz A, Regev R, Litmanovits I,
Arnon S, et al. Urinary tract infection in very low birth weight
preterm infants. Pediatr Infect Dis J 2003;22:426e30.
4. Foglia EE, Lorch SA. Clinical predictors of urinary tract infec-
tion in the neonatal intensive care unit. J Neonatal Perinatal
Med 2012;5:327e33.
5. Wallace SS, Zhang W, Mahmood NF, Williams JL, Cruz AT,
Macias CG, et al. Renal ultrasound for infants younger than 2
L. Walawender et al
Months with a febrile urinary tract infection. AJR Am J
Roentgenol 2015;205:894e8.
6. Lowry R. VassarStats: website for statistical computation.
Vassar College; 2004.
7. Craig JC, Simpson JM, Williams GJ, Lowe A, Reynolds GJ,
McTaggart SJ, et al. Antibiotic prophylaxis and recurrent uri-
nary tract infection in children. N Engl J Med 2009;361:
1748e59.
8. Thomas-White K, Brady M, Wolfe AJ, Mueller ER. The bladder is
not sterile: history and current discoveries on the urinary
microbiome. Curr Bladder Dysfunct Rep 2016;11:18e24.
9. Price TK, Dune T, Hilt EE, Thomas-White KJ, Kliethermes S,
Brincat C, et al. The clinical urine culture: enhanced tech-
niques improve detection of clinically relevant microorgan-
isms. J Clin Microbiol 2016;54:1216e22.
10. Thomas-White KJ, Kliethermes S, Rickey L, Lukacz ES,
Richter HE, Moalli P, et al. Evaluation of the urinary microbiota
of women with uncomplicated stress urinary incontinence. Am
J Obstet Gynecol 2017;216:55.e1e55.e16.
11. Hilt EE, McKinley K, Pearce MM, Rosenfeld AB, Zilliox MJ,
Mueller ER, et al. Urine is not sterile: use of enhanced urine
culture techniques to detect resident bacterial flora in the
adult female bladder. J Clin Microbiol 2014;52:871e6.
12. Clarke D, Gowrishankar M, Etches P, Lee BE, Robinson JL.
Management and outcome of positive urine cultures in a
neonatal intensive care unit. J Infect Public Health 2010;3:
152e8.
13. Becknell B, Schober M, Korbel L, Spencer JD. The diagnosis,
evaluation and treatment of acute and recurrent pediatric
urinary tract infections. Expert Rev Anti Infect Ther 2015;13:
81e90.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.pedneo.2019.10.003.