YAJEM-56769; No of Pages 6

American Journal of Emergency Medicine xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

American Journal of Emergency Medicine

journal homepage: www.elsevier.com/locate/ajem

Risk factors for early return visits to the emergency department in patients with urinary
tract infection☆
Sarah Jorgensen, Pharm. D. a,b, Mira Zurayk, Pharm. D. a, Samantha Yeung, MSc, Pharm. D. b, Jill Terry, Pharm. D. a,
Maureen Dunn, MD c, Paul Nieberg, MD d, Annie Wong-Beringer, Pharm. D. a,b,⁎
a
Department of Pharmacy, Huntington Hospital, 100 W California Blvd, Pasadena, CA 91105, United States
b
University of Southern California, School of Pharmacy, 1985 Zonal Ave, Los Angeles, CA 90089, United States
c
Division of Emergency Medicine, Department of Medicine, Huntington Hospital, 100 W California Blvd, Pasadena, CA 91105, United States
d
Division of Infectious Diseases, Department of Medicine, Huntington Hospital, 100 W California Blvd, Pasadena, CA 91105, United States

a r t i c l e i n f o a b s t r a c t

Article history: Background: Optimal management of urinary tract infections (UTIs) in the emergency department (ED) is chal-
Received 5 April 2017 lenging due to high patient turnover, decreased continuity of care, and treatment decisions made in the absence
Received in revised form 15 June 2017 of microbiologic data. We sought to identify risk factors for return visits in ED patients treated for UTI.
Accepted 21 June 2017 Methods: A random sample of 350 adult ED patients with UTI by ICD 9/10 codes was selected for review. Relevant
Available online xxxx
data was extracted from medical charts and compared between patients with and without ED return visits within
30 days (ERVs).
Keywords:
Urinary tract infection
Results: We identified 51 patients (15%) with 59 ERVs, of whom 6% returned within 72 h. Nearly half of ERVs
Return visits (47%) were UTI-related and 33% of ERV patients required hospitalization. ERVs were significantly more likely
Emergency department (P b 0.05) in patients with the following: age ≥ 65 years; pregnancy; skilled nursing facility residence; dementia;
psychiatric disorder; obstructive uropathy; healthcare exposure; temperature ≥ 38 °C heart rate N 100; and bac-
teremia. Escherichia coli was the most common uropathogen (70%) and susceptibility rates to most oral antibi-
otics were below 80% in both groups except nitrofurantoin (99% susceptible).
Cephalexin was the most frequently prescribed antibiotic (51% vs. 44%; P = 0.32). Cephalexin bug-drug mis-
matches were more common in ERV patients (41% vs. 15%; P = 0.02). Culture follow-up occurred less frequently
in ERV patients (75% vs. 100%; P b 0.05).
Conclusions: ERV in UTI patients may be minimized by using ED-source specific antibiogram data to guide empiric
treatment decisions and by targeting at-risk patients for post-discharge follow-up.
© 2017 Published by Elsevier Inc.

1. Introduction
Urinary tract infection (UTI) is a leading cause of infection among
patients presenting to the emergency department (ED), accounting for
nearly 2 million visits in females of all ages and 160,000 visits in males
age 65 and older in the US in 2013 [1]. Relapse or recurrent infection oc-
curs in up to 44% of women with community-acquired cystitis [2]. The
likelihood of treatment failure increases with age and in those with
complicated infection [3]. Several investigators have also identified
UTI to be among the most common diagnosis in patients with return
visits to the ED [4-6]. Revisits to the ED are expensive and add strain
to already overburdened EDs [6]. The evolving structure of healthcare
reimbursement places increasing emphasis on value driven care. Thus,
☆ Investigator initiated grant support from Merck Inc. to AWB
⁎ Corresponding author at: University of Southern California, School of Pharmacy,
United States.
E-mail address: anniew@usc.edu (A. Wong-Beringer).
accurate identification of patients at risk for return visits is critical to in-
form the development of interventions aimed at improving outcomes
and healthcare resource utilization.
Optimal management of UTIs in the ED setting is particularly chal-
lenging due to high patient turnover, decreased continuity of care, and
therapeutic decisions made in the absence of microbiologic data. Fur-
thermore, the increasing emergence of antimicrobial resistance among
uropathogens in the community and inpatient settings [7,8] presents a
significant challenge for ED clinicians to balance prompt initiation of ef-
fective empiric antibiotic therapy without overprescribing broad spec-
trum agents. ED return visits within 30 days (ERVs) in patients with
UTI may be related to a number of factors including resistance to antimi-
crobial therapy, uncorrected functional or structural urinary tract ab-
normalities, recurrent infection or unrelated issues. Previous studies
have sought to identify risk factors for ERVs in the general ED popula-
tion [5,9-11]. Several revisit prediction models have also been devel-
oped for the elderly, for patients within specific payor groups and/or
within specific disease groups [12-16] but none, to our knowledge,

http://dx.doi.org/10.1016/j.ajem.2017.06.041
0735-6757/© 2017 Published by Elsevier Inc.

Please cite this article as: Jorgensen S, et al, Risk factors for early return visits to the emergency department in patients with urinary tract infection,
American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.06.041
2 S. Jorgensen et al. / American Journal of Emergency Medicine xxx (2017) xxx–xxx
have investigated the risk factors for ERVs in ED patients treated for UTI.
Thus, we sought to examine host factors, antimicrobial resistance and
treatment decisions in ED patients with UTI to identify at risk groups
for ERVs.
2. Methods
This was a retrospective cohort study conducted at a 636-bed, non-
profit, community teaching hospital in Pasadena, CA. The 50-bed Level II
Trauma Center and ED have an annual census of N70,000 patient visits.
The study protocol was approved by the Huntington Hospital and the
University of Southern California Institutional Review Boards with
waivers for informed consent.
Patients presented to the ED with a primary or secondary UTI diag-
nosis by ICD codes (ICD-9595.9, 599.0; ICD-10 N30.00, N30.01, N39.0)
and discharged directly from the ED between July 2015 and June 2016
were identified. A total of 350 patients were randomly selected to rep-
resent approximately 10% of the population of interest. Exclusion
criteria were: 1) age b 18 years, 2) refusal of evaluation and/or treat-
ment, and 3) incomplete medical record. Only the first visit was counted
as the index visit for patients with multiple ED visits during the study
period. Time to ERV was recorded as within 72 h or beyond 72 h of
the index visit. Relevant patient demographic, laboratory and clinical
data were extracted from the electronic medical record using a struc-
tured data collection form and entered into REDCap (Research Electron-
ic Data Capture, Vanderbilt University), an electronic data capture tool
hosted at the University of Southern California.
Urinalysis and urine cultures were collected from ED patients in the
course of routine medical care. For the purposes of this study, a positive
urinalysis was defined as one or more of the following: leukocytes
≥ trace and/or nitrite positive on urine dipstick; leukocyte esterase
≥ trace and/or nitrite positive on macroscopic urinalysis; and/or WBC
≥ 10/HPF and/or bacteria ≥ small on microscopic urinalysis.
Uropathogen colony counts ≥103 cfu/mL in a voided or catheter speci-
men were reported as positive. Urine cultures showing three or more
organisms were reported as probable contamination. Automated bacte-
rial identification and susceptibility testing was conducted on positive
urine cultures using the BD Phoenix Automated Microbiology System
(BD Diagnostic Systems, Sparks, MD). Antibiotic susceptibility was
interpreted using the CLSI 2015 breakpoint criteria [17]. An E. coli
cefazolin breakpoint of ≤16 mcg/mL was used as a surrogate for cepha-
lexin susceptibility in patients with uncomplicated cystitis while a
breakpoint of ≤2 mcg/mL was used for complicated UTI [17]. Intermedi-
ate susceptibility was considered non-susceptible. ESBL phenotype was
detected via the Phoenix BDXpert system (BD Diagnostic Systems,
Sparks, MD). Phenotypic ESBL confirmation testing, using disk-diffusion
with both cefotaxime and ceftazidime alone and in combination with
clavulanic acid, was performed on isolates demonstrating indetermi-
nate susceptibility patterns by the BDXpert system.
Symptomatic presentation was classified as UTI-specific (dysuria,
urinary urgency, urinary frequency, flank/back pain) or non-specific
(nausea/vomiting, abdominal pain, subjective fevers/chills, acutely al-
tered mental status, fall, anorexia, malaise, lethargy, dizziness, and
new or worsening incontinence). Patients without UTI-specific or non-
specific symptoms were classified as asymptomatic. We observed that,
irrespective of the reason prompting the ERV, UTI was often listed as
an active diagnosis in patients returning within a week of their index
visit and coded as such. Thus we considered ERVs to be UTI-related
based on both ICD codes and documentation of UTI signs or symptoms
in the medical record. Symptomatic patients were considered to have
a complicated UTI if any of the following characteristics were present:
age ≥ 65 years; male gender; pregnancy; history of structural or func-
tional urinary tract abnormality (history of kidney stones, prostate pa-
thology, urinary catheter within 30 days, genitourinary procedure
within 60 days, neurogenic urinary retention, ureteral stricture, renal
or bladder cancer, renal transplant, single kidney, and nephrostomy
Please cite this article as: Jorgensen S, et al, Risk factors for early return visits to the emergency department in patients with urinary tract infection,
American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.06.041
tubes); immunocomprimising condition (diabetes with A1c N 8.5%
and/or random glucose N200 mg/dL, active cancer, chemotherapy or bi-
ological agent within 30 days, prednisone or equivalent ≥ 20 mg/day
≥2 weeks, dialysis and HIV infection); failure of outpatient management
for index UTI; history of recurrent UTI (≥ 2 UTI/6 months or ≥ 3 UTI/
12 months) or presence of signs or symptoms of upper tract/systemic
infection (subjective fever/chills, flank/back pain, acutely altered mental
status, temperature ≥ 38 °C, WBC N 12 × 103/μL, bands N5%, systolic
blood pressure b 100 mm Hg). All other symptomatic patients were
considered to have an uncomplicated UTI.
A pharmacist-managed culture follow-up program has been opera-
tional in our ED since 2014. Clinical pharmacists with specialized post-
graduate emergency medicine training are present in the ED between
0700 h and 2300 h daily during which time computerized decision soft-
ware alerts them to positive cultures for patients discharged from the
ED. For patients with discordant therapy, the ED pharmacist and physi-
cian determine a follow-up plan which may include a change in therapy
or patient re-evaluation. The pharmacist is responsible for contacting
the patient by telephone to communicate the plan.
3. Data analysis
Comparisons between categorical variables in patients with and
without ERVs were performed using Chi-square or Fisher's exact tests
as appropriate. Differences among continuous variables were evaluated
using Mann Whitney U test. All statistical tests were performed using
Prism version 7.0 (GraphPad Software, San Diego, CA). A 2-tailed P
value of b 0.05 was considered to be statistically significant.
4. Results
We identified 51/350 patients (15%) with 59 ERVs during the study
period, of whom 6% (20/350) returned within 72 h of their index visit
and 2% (8/350) returned twice within 30 days. Nearly half (47%, 28/
59) of ERVs were UTI-related. The proportions of patients with a UTI-re-
lated reason for return did not differ between those returning within
72 h and those returning beyond 72 h (55% vs. 44%; P = 0.41). Overall
one-third of ERV patients (17/51) were subsequently hospitalized, bias-
ing towards those returning within 72 h compared to those returning
later (40% vs. 29%; P = 0.14) and those returning for a non-UTI-related
reason (39% vs. 18%; P = 0.09).
Patient characteristics are shown in Table 1. Patients with ERVs were
significantly older than those who did not return to the ED (mean age
57 years vs. 49 years; P = 0.03). Specifically, nearly half of ERV patients
were 65 years or older (45% vs. 27%; P = 0.008) and had Medicare insur-
ance (31% vs. 17%; P = 0.01). In contrast, the proportions of patients
with Medicaid, private insurance or no insurance were similar between
groups. ERVs were more likely in patients with the following character-
istics in descending order of odds ratio: pregnancy (odds ratio [OR], 6.3;
95% confidence interval [CI], 1.8–22.0; P = 0.02); residence in a SNF
(OR, 5.6; 95% CI, 1.4–15.7; P = 0.001); obstructive uropathy (OR, 5.3;
95% CI 1.7–19.0; P = 0.03); comorbid dementia (OR, 4.5; 95% CI, 1.7–
11.1; P = 0.0009); healthcare exposure within 6 months of the index
ED visit (OR, 2.3; 95% CI 1.2–4.1; P = 0.01) and psychiatric disorder
(OR, 2.2; 95% CI, 1.1–4.5; P = 0.03).
A comparison of the clinical presentation in those with and without
ERVs is shown in Table 2. The majority of patients in both groups pre-
sented with a complicated UTI (73% vs. 67%; P = 0.67) and less than
half reported UTI-specific symptoms (45% vs. 46%; P = 0.96). Approxi-
mately 15% of patients in both groups were asymptomatic (16% vs.
14%; P = 0.81). ERVs were significantly more likely in patients with
fever (T ≥ 38C; OR 5.3; 95% CI, 1.7–19.0; P = 0.01) and tachycardia
(HR N 100; OR 3.6; 95% CI, 1.6–8.4; P = 0.002).
Microbiological characteristics are shown in Table 3. Urinalysis was
performed in all study patients and the results were positive in N99%
(347/350). Among the patients with a negative urinalysis result (n =
 S. Jorgensen et al. / American Journal of Emergency Medicine xxx (2017) xxx–xxx 3

Table 1
Patient characteristics.

Characteristic Overall ED cohort Patients with return visits Patients without return visits Odds ratio (95% confidence interval); P value
N = 350 N = 51 N = 299

Age (yr) mean (± SD) 50.3 ± 23.1 56.6 ± 24.1 49.2 ± 22.8 0.03
≥65 years 103 (29.4) 23 (45.1) 80 (26.8) 2.25 (1.23–4.20); 0.008
Gender, female 304 (86.9) 44 (86.3) 260 (87.0) 0.94 (0.39–2.37); 0.89
Pregnant 8 (2.3) 4 (7.8) 4 (1.3) 6.28 (1.76–21.98); 0.02
Payer
Medicaid 100 (28.6) 12 (23.5) 88 (29.4) 0.74 (0.38–1.44); 0.39
Medicare 66 (18.9) 16 (31.4) 50 (16.7) 2.28 (1.19–4.37); 0.013
Private 152 (43.4) 20 (39.2) 132 (44.1) 0.82 (0.44–1.49); 0.51
Uninsured 32 (9.1) 3 (5.9) 29 (9.7) 0.58 (0.18–1.90); 0.60
SNF residence prior to presentation 13 (3.7) 6 (11.8) 7 (2.3) 5.56 (1.37–15.7); 0.001
Three or more comorbidities 16 (4.6) 5 (9.8) 11 (3.8) 2.85 (1.05–7.95); 0.07
Cardiovascular disease 25 (7.1) 6 (11.8) 19 (6.4) 1.97 (0.76–5.04); 0.17
Diabetes 52 (14.9) 6 (11.8) 46 (15.4) 0.73 (0.31–1.72); 0.50
Dementia 20 (5.7) 8 (15.7) 12 (4.0) 4.45 (1.73–11.11)0.0009
Psychiatric disorder 54 (15.4) 13 (25.5) 41 (13.7) 2.15 (1.07–4.45); 0.031
Obstructive uropathy 11 (3.1) 5 (9.8) 6 (2.0) 5.31 (1.74–18.97); 0.013
Long-term/intermittent catheter 11 (3.1) 2 (3.9) 9 (3.0) 1.32 (0.28–5.26); 0.67
History of recurrent UTI 32 (9.1) 8 (15.7) 24 (8.0) 2.13 (0.95–5.00); 0.08
Hospitalization (≥48 h) within 6 months 41 (11.7) 9 (17.6) 32 (10.7) 1.79 (0.77–4.05); 0.15
Healthcare exposure within 6 months 87 (24.9) 20 (40.0) 67 (22.4) 2.23 (1.16–4.10); 0.01
IV antibiotics within 90 days 17 (4.9) 5 (9.8) 12 (4.0) 2.6 (0.97–7.74); 0.084
PO antibiotics within 90 days 34 (9.7) 7 (13.7) 27 (9.0) 1.60 (0.61–3.84); 0.30
History of MDRO 5 (1.4) 1 (2.0) 4 (1.3) 1.48 (0.12–9.16); 0.55
Estimated CrCl
Cr measured 281 (80.3) 47 (92.2) 234 (78.3) 3.26 (1.18–8.72); 0.02
N60 mL/min 167 () 26 (55.3) 141 (60.3) 0.82 (0.44–1.57); 0.53
30–60 mL/min 92 (26.3) 15 (31.9) 77 (32.9) 0.96 (0.50–1.85); 0.89
b30 mL/min 22 (6.3) 6 (12.8) 16 (6.8) 1.99 (0.75–5.51); 0.17

ED emergency department; UTI urinary tract infection; BMI body mass index; SNF skilled nursing facility; IV intravenous; PO oral; MDRO multidrug resistant pathogen; CrCl creatinine
clearance.

3), 2 had received antibiotics within the week prior their ED visits. Urine
culture was collected from the majority of patients (82% vs. 77%; P =
0.47) and similar proportions of patients in both groups had positive
cultures (41% vs. 42%; P = 0.86). Contamination rates were relatively
high in the 2 groups (48% vs. 41%; P = 0.43). More patients in the ERV
group had concurrent bacteremia (6% vs. 0.3%; P = 0.01). E. coli was
the most commonly isolated uropathogen (65% vs. 71%; P = 0.61).
ESBL was detected in 2 E. coli isolates from each group (P = 0.1). The
study groups did not differ with respect to E. coli susceptibility rates
which were below 80% for most oral UTI antibiotics but close to 100%

Table 2
Clinical presentation.

Overall ED cohort Patients with return visits Patients without return visits Odds ratio (95% confidence interval); P value
N = 350 N = 51 N = 299

UTI classification
Uncomplicated 54 (15.4) 6 (11.8) 48 (16.1) 0.70 (0.30–1.63); 0.43
Complicated 245 (70.0) 37 (72.5) 208 (69.5) 1.16 (0.59–2.22); 0.67
Asymptomatic 51 (14.6) 8 (15.7) 43 (14.4) 1.11 (0.51–2.49); 0.81
Symptoms
UTI-specific symptoms onlya 62 (17.7) 8 (15.7) 54 (18.1) 0.84 (0.39–1.85); 0.68
Non-specific symptom onlyb 141 (40.3) 20 (39.2) 121 (40.5) 0.95 (0.51–1.74); 0.87
Both specific & non-specific symptoms 97 (27.7) 15 (29.4) 82 (27.4) 1.10 (0.59–2.15); 0.77
Presence of systemic symptoms
Subjective fever/chills 44 (12.6) 9 (17.6) 35 (11.7) 1.62 (0.70–3.60); 0.24
Nausea/vomiting 109 (31.1) 18 (35.3) 91 (30.4) 1.25 (0.68–2.34); 0.49
Acutely altered mental status 9 (2.6) 2 (3.9) 7 (2.3) 1.7 (0.35–8.01); 0.62
Fall 18/350 (5.1) 2/51 (3.9) 16 (5.4) 0.72 (0.16–2.74); N0.99
Weakness/dizziness 62 (17.7) 12 (23.5) 50 (16.7) 1.53 (0.77–3.14); 0.24
Vital signs
Tmax ≥38 °C 11 (3.1) 5 (9.8) 6 (2.0) 5.31 (1.74–18.97); 0.01
RR ≥ 22 32 (9.1) 6 (11.8) 26 (8.7) 1.4 (0.57–3.6); 0.48
HR N 100 29 (8.3) 10 (19.6) 19 (6.4) 3.59 (1.58–8.41); 0.0015
SBP b 100 mm Hg 31 (8.9) 7 (13.7) 24 (8.0) 1.82 (0.69–4.49); 0.19
Labs
WBC N12 or b4 × 103/mcL 56 (16.0) 8 (15.7) 48 (16.1) 0.97 (0.45–2.16); 0.95
Bands N5% 1 (0.3) 1 (2.0) 0 ∞ (0.65–∞); 0.15
Lactate N19.8 mg/dL 4 (1.1) 1 (2.0) 3 (1.0) 1.97 (0.15–13.42); 0.47
qSOFA score ≥ 2 6 (1.7) 1 (2.0) 5 (1.7) 1.18 (0.10–8.86); N0.999
≥2 SIRS criteria positive 32 (9.1) 5 (9.8) 27 (9.0) 1.10 (0.44–2.79); 0.80

UTI urinary tract infection; T temperature; RR respiratory rate; HR heart rate; SBP systolic blood pressure; WBC white blood cell; qSOFA quick sepsis related organ failure assessment; SIRS
systemic inflammatory response syndrome
a
UTI-specific symptoms: dysuria, frequency, urgency, flank/back pain.
b
Non-specific symptoms: Subjective fever/chills, nausea/vomiting, anorexia, acutely altered mental status, lethargy/weakness/dizziness, malaise, new/worsening incontinence.

Please cite this article as: Jorgensen S, et al, Risk factors for early return visits to the emergency department in patients with urinary tract infection,
American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.06.041
4 S. Jorgensen et al. / American Journal of Emergency Medicine xxx (2017) xxx–xxx

Table 3
Microbiology and antibiotic prescribed.

Overall ED Patients with early return Patients without early return Odds ratio (95% confidence interval); P
cohort visits visits value
N = 350 N = 51 N = 299

Urine culture collected 271 (77.4) 42 (82.4) 229 (76.6) 1.43 (0.66–3.10); 0.36
Urine culture result N = 271 N = 42 N = 229
Positive 113 (41.7) 17 (40.5) 96 (41.9) 0.94 (0.47–1.87); 0.86
Contaminated 114 (42.1) 20 (47.6) 94 (41.0) 1.31 (0.69–2.45); 0.43
No growth 44(16.2) 5 (11.9) 39 (17.0) 0.66 (0.27–1.77); 0.41
Positive blood culture 4 (1.1) 3 (6) 1 (0.3) 17.54 (2.52–228); 0.01
Escherichia coli positive urine culture N = 113 N = 17 N = 96 0.75 (0.27–2.31); 0.61
79 (69.9) 11 (64.7) 68 (70.8)
E. coli isolatesa 81 12 69
ESBL-producing E. coli N = 81 N = 12 N = 69 6.7 (0.93–44.72); 0.10
4 (4.9) 2 (16.7) 2 (2.9)
E. coli antibiotic resistance (% resistant) N = 81 N = 12 N = 69
Cefazolin 22 (27.2) 6 (50) 16 (23.2) 3.31 (0.93–10.9); 0.06
Ciprofloxacin 18 (22.2) 3 (25.0) 15 (21.7) 1.2 (0.32–4.64); 0.73
TMP/SMX 27 (33.3) 5 (41.7) 22 (31.9) 1.53 (0.47–5.22); 0.52
Nitrofurantoin 1(1.2) 0 1 (1.4) 0 (0–51.8); N0.999
IV/IM antibiotic in ED 173 (49.4) 30 (58.8) 143 (47.8) 1.56 (0.85–2.84); 0.15
Ceftriaxone 163 (94.8) 29 (96.7) 135 (94.4) 1.72 (0.24–19.69); N0.999
Discharge antibiotic 345 (98.6) 51 (100) 294 (98.3) ∞ (0.25–∞); N0.999
N = 345 N = 51 N = 294
Cephalexin 154 (44.6) 26 (51.0) 128 (43.5) 1.35 (0.73–2.50); 0.32
Ciprofloxacin 59 (17.1) 7 (13.7) 52 (17.7) 0.74 (0.3–1.71); 0.49
TMP/SMX 52 (15.1) 11 (21.6) 41 (13.9) 1.70 (0.78–3.46); 0.16
Nitrofurantoin 65 (18.9) 5 (9.8) 60 (20.4) 0.42 (0.18–1.05); 0.08
Discharge bug-drug mismatches N = 113 N = 17 N = 96 2.27 (0.76–6.40); 0.12
35 (31.0) 8 (47.1) 27 (28.1)
Cephalexin 22 (19.5) 7 (41.4) 15 (15.6) 3.78 (1.32–11.6); 0.02
Ciprofloxacin 5 (4.4) 0 5 (5.2) 0 (0–4.04); N0.99
SXT 6 (5.3) 1 (5.9) 5 (5.2) 1.14 (0.09–9.7); N0.99
Nitrofurantoin 1 (0.9) 0 1 (1.0) 0 (0–50.8); N0.99
N = 35 N=8 N = 27
Culture follow-up intervention performed 32 (91.4) 6 (75.0) 27 (100) 0 (0–0.59); 0.05
New antibiotic with in vitro activity 25 (78.1) 3 (50) 22 (81.5)
prescribed
Scheduled follow-up with PCP 5 (15.6) 0 5 (18.5)
Scheduled follow-up in ED 3 (9.3) 3 (50) 0

ED emergency department; ESBL extended spectrum beta-lactamase; IV intra venous; IM intramuscular; SXT trimethoprim-sulfamethoxazole; PCP primary care provider
a
Two patients with 2 different E. coli strains in a single urine culture.

for nitrofurantoin. Approximately half of patients received parenteral an-
tibiotics in the ED (59% vs. 48%; P = 0.15) with ceftriaxone prescribed in
nearly all instances (97% vs. 94%; P N 0.99). Only 5 patients (non-ERV
group, 1.7%) did not receive a discharge UTI antibiotic prescription. Ceph-
alexin was the most commonly prescribed oral agent in both groups (51%
vs. 44%; P = 0.3) while nitrofurantoin was prescribed less frequently (10%
vs. 20%; P = 0.08). Among patients in the ERV group, resistance to the dis-
charge antibiotic was observed in 47% (8/17) of urine-culture positive pa-
tients compared to 28% (27/96) of patients without ERVs (P = 0.12).
Patients with ERVs were significantly more likely to have cephalexin-
bug mismatches (OR, 3.8; 95% CI, 1.3–11.6; P = 0.02). Nitrofurantoin
was mismatched in just 1 patient (non-ERV group).
As shown in Table 3, 75% (6/8) of EVR patients with bug-drug mis-
matches received follow-up through the culture review program com-
pared to 100% (27/27) of those without ERVs (P b 0.5). ERV patients
who did not receive follow-up (2/8), returned to the ED b 72 h following
the index visit and prior to availability of culture susceptibilities. Follow-
up interventions included modification of the empiric antibiotic regi-
men based on susceptibility results (3/6, 50% vs. 22/27, 81%; P =
0.14); scheduled re-assessment with the patient's primary care physi-
cian (non-ERV group 5/27, 19%); and request that the patient return
to the ED for re-evaluation due to lack of oral treatment options second-
ary to resistance (ERV group 3/6, 50%).
5. Discussion
This study focused on risk factors for return visits within 30 days in
patients treated for UTI in the ED setting. In contrast to the majority of
previous studies based on national databases or single institutions
that examined ED revisits for all patients [5,6,9-11], we limited our
study to a cohort with a particular clinical condition that has been
found to be among the top diagnosis in patients with ERVs[4-6],
UTI. To our knowledge, we are the first to examine risk factors for re-
turn visits in ED patients with UTI. We documented an ERV rate with-
in 30 days of 15% which is similar to recent studies investigating
30 day return visits among the general ED population [4,11] and in
the elderly [18]. Our 72 h revisit rate of 6% however is higher than re-
ported in previous studies that have captured return visits in the
overall ED population and have been in the range of 2.7% to 3.4%
[10,19,20]. Our 72 h return rate is more consistent with the rate of
5.4% reported for the UTI subgroup in a recent multistate investiga-
tion of ED revisits and associated costs by Duseja et al. [6]. Consider-
ing the high volume of UTI cases being managed in the ED setting
(approximately 3500 cases annually at our institution and N 2 million
cases nationwide [1]), the ERV rate within 72 h of 6% translates into a
substantial number of return visits.
In addition to advanced age, Medicare insurance and comorbid psy-
chiatric illness recognized here and in previous studies in the general ED
population [5,12,16], we have identified additional risk factors for re-
turn visits specific to ED patients with UTI. Obstructive uropathy is a
well-documented risk factor for recurrent infection while residence in
a SNF and recent healthcare exposure in patients with UTI are associated
with antibiotic-resistant pathogens [3], thus it is plausible that these
factors should place patients at increased risk for ERVs as observed in
our study. Fever and tachycardia at presentation were more common
in ERV patients raising the possibility of unrecognized early

Please cite this article as: Jorgensen S, et al, Risk factors for early return visits to the emergency department in patients with urinary tract infection,
American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.06.041
 S. Jorgensen et al. / American Journal of Emergency Medicine xxx (2017) xxx–xxx
pyelonephritis and underscores the diagnostic uncertainty present in a
significant number of UTI ED encounters.
Because our overall goal was to improve the management and out-
comes of patients diagnosed with and treated for UTI in the ED setting,
we based our inclusion criteria on physician diagnosis rather than con-
sensus UTI definitions. Not unexpectedly, we observed that 15% of pa-
tients were asymptomatic and 40% presented with non-specific
symptoms only. Furthermore, although the urinalysis was positive in
nearly all patients, less than half (42%) of patients had a positive urine
culture. These findings point to the need for education pertaining to
the appropriate diagnosis of UTI as well as antimicrobial stewardship ef-
forts to reduce treatment of asymptomatic patients.
When deciding upon disposition, ED physicians must balance the
expected benefits of hospitalization against the clinical uncertainty of
outpatient management as well as further infectious risks and financial
costs associated with hospitalization. The admission rate among our UTI
patients with ERVs (33%) was within the range of our ED's overall ad-
mission rate of approximately 25–33%, suggesting that in majority of
cases, a trial of outpatient management was warranted even though
this approach did not always work.
Over 40% of patients in our cohort had a contaminated urine culture
result affirming the need to improve urine culture collection technique
in ED patients. Accurate culture results are of critical importance partic-
ularly in the setting of our relatively high E. coli resistance rates to in-
form the development of empiric treatment guidelines.
Considering the high rates of E. coli resistance rates to trimethoprim-
sulfamethoxazole and the fluoroquinolones, leaving only beta-lactams
and nitrofurantoin as oral options, our finding that cephalexin was the
most commonly prescribed discharge antibiotic was not unexpected.
We hypothesize that the popularity of cephalexin was driven by limited
alternative options in the setting of a predominantly elderly population
exhibiting age-related declines in renal function in whom
nitrofurantoin use may be perceived as problematic. Of concern, cepha-
lexin bug-drug mismatches occurred in 41% of culture-positive ERV pa-
tients and were significantly more common in patients with ERV
compared to those not returning to the ED. Importantly, the significant
difference in E. coli susceptibility to cefazolin between that reported in
the institutional antibiogram and our cohort (83% vs. 73%; P = 0.02)
highlights the need for up-to-date ED source-specific antibiogram data
to guide therapy selection in the ED.
In the subgroup of patients with bug-drug mismatches to empiric
therapy, culture follow-up interventions were significantly less com-
mon in those with ERVs underscoring the value of such programs in im-
proving antibiotic utilization and averting preventable return visits. The
success of these programs however, hinges on a reliable method to con-
tact patients post-discharge. Although not observed among our study
patients, we have had several cases in which interventions could not
be attempted for patients without a fixed address or telephone. Further-
more among our cohort 2 patients returned to the ED before interven-
tions could be attempted thus also highlighting the importance of
appropriate initial empiric therapy. It is noteworthy that the interven-
tion in 50% (3/6) of ERV patients consisted of arranging for the patient
to return to the ED due to persistent symptoms in the setting of multi-
drug resistance with no appropriate oral treatment options. More seri-
ous complications were likely prevented through timely reassessment
of these patients.
Among our study patients with an E. coli positive urine culture,
nitrofurantoin appears to be the most active oral antibiotic supporting
its placement as a preferred agent in the current Infectious Diseases So-
ciety of America/European Society for Microbiology and Infectious Dis-
eases guideline for the management of uncomplicated UTI [21].
Furthermore, recent studies have documented an association between
nitrofurantoin prescription for UTI and reduced risk for inadequate ther-
apy in the ED setting or re-consultation in the ambulatory setting [22,
23]. While a minority of patients in our study were considered to have
uncomplicated UTI (15%), evidence also suggests nitrofurantoin is
Please cite this article as: Jorgensen S, et al, Risk factors for early return visits to the emergency department in patients with urinary tract infection,
American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.06.041
5
both safe and effective in healthy elders with mild to moderate renal im-
pairment [24,25]. With the lowering of the creatinine clearance (CrCl)
breakpoint to 30 mL/min (previously 60 mL/min) in the updated
Beers Criteria [26], a greater number of elderly patients may be appro-
priate candidates for treatment with nitrofurantoin. This encompasses
over one-quarter (26%) of our cohort who had an estimated CrCl in
the range 30–60 mL/min. Thus, in the context of low level resistance
and limited ecological collateral damage, nitrofurantoin use should be
encouraged in patients who do not have pyelonephritis or suspected
systemic involvement.
Our study has several important limitations. First, the study is sub-
ject to inherent bias with its retrospective design, although our in-
depth chart review allowed us to obtain detailed patient level data
that has not been available in the majority of previous studies that uti-
lized large administrative databases to investigate the rates and risk fac-
tors for ERVs to the ED. As a result of using pragmatic inclusion criteria
based on ICD coding, we observed that a substantial portion of our co-
hort did not meet the consensus definition of UTI based on symptom
and laboratory/microbiology criteria. Second, as our patients did not be-
long to a closed healthcare system, we were unable to capture data
pertaining to ERVs to other institutions. Thus our ERV rate is likely an
underestimate of the true burden of this problem. All patients with
UTI seen in our ED are routinely instructed to follow-up with their
primary care provider (PCP) within a few days of their index visit
and return to the ED for severe or worsening symptoms. As such,
our goal was to capture those in the latter group to identify risk fac-
tors requiring more intense use of resources in their care and alter-
native initial management. We did not review data pertaining to
follow-up ambulatory care visits, however future research aimed at
identifying factors influencing patients' decisions to follow-up with
their PCP vs. return to the ED would be valuable. Finally, our data
were obtained from a single institution and may not be generalizable
to other EDs in distinct geographic areas or serving a different pa-
tient population by age, comorbid conditions or risk factors for anti-
microbial resistance.
6. Conclusions
Because of the high burden of UTIs and complexity of treatment de-
cisions among the ED population, we sought to examine risk factors for
return ED visits within 30 days. Several risk factors can be readily iden-
tified during the initial patient assessment including advanced age, res-
idence in a SNF, recent healthcare exposure and comorbid dementia or
psychiatric disorders. Furthermore, simple clinical parameters such as
fever and tachycardia can enable detection of at risk individuals who
may be targeted for close follow-up after discharge from the ED and
for prompt intervention if needed. Patients for whom reliable outpa-
tient follow-up is not readily available present a particular management
challenge to ED clinicians. While this may need to be factored into dis-
position decisions, it is equally important that novel programs are de-
veloped to help these patients navigate the healthcare system and
facilitate safe ED discharge. In our ED, the assistance of case managers
can be requested to organize follow-up appointments at our internal
medicine resident-run ambulatory-care clinic or at several local clinics
that accept uninsured/low income patients. We also have an
underutilized protocol whereby antibiotics are filled at contracted retail
pharmacies and billed to the hospital for uninsured/indigent patients.
Promotion of these services for in need patients identified to be at risk
for ERV is critical to ensuring efficient use of resources and safe care
transitions.
Importantly, appropriate choice of empiric agent for treatment (e.g.
nitrofurantoin instead of cephalexin) at discharge based on an ED-
source specific antibiogram [27] may improve the chance of treatment
success and reduce the risk of avoidable ERVs by decreasing the likeli-
hood of bug-drug mismatch.
6 S. Jorgensen et al. / American Journal of Emergency Medicine xxx (2017) xxx–xxx
Acknowledgment
This study was presented in part at the ASM Microbe Annual Meet-
ing, June 3, 2017; New Orleans, LA, USA (Abstract #299).
References
[1] Rui P, Kang K, Albert M. National Hospital Ambulatory Medical Care Survey: 2013
Emergency Department Summary Tables. Available from http://www.cdc.gov/
nchs/data/ahcd/nhamcs_emergency/2013_ed_web_tables.pdf. Accessed 1/25/2017.
[2] Ikäheimo R, Siitonen A, Heiskanen T, Kärkkäinen U, Kuosmanen P, Lipponen P, et al.
Recurrence of urinary tract infection in a primary care setting: analysis of a 1-year
follow-up of 179 women. Clin Infect Dis 1996 Jan;22(1):91–9.
[3] Pallett A, Hand K. Complicated urinary tract infections: practical solutions for the
treatment of multiresistant Gram-negative bacteria. J Antimicrob Chemother 2010
Nov;65(Suppl. 3):iii25–33.
[4] Sabbatini AK, Kocher KE, Basu A, Hsia RY. In-hospital outcomes and costs among pa-
tients hospitalized during a return visit to the emergency department. JAMA 2016;
315(7):663–71.
[5] Martin-Gill C, Reiser RC. Risk factors for 72-hour admission to the ED. Am J Emerg
Med 2004 Oct;22(6):448–53.
[6] Duseja R, Bardach NS, Lin GA, Yazdany J, Dean ML, Clay TH, et al. Revisit rates and
associated costs after an emergency department encounter: a multistate analysis.
Ann Intern Med 2015 Jun 2;162(11):750–6.
[7] Lob SH, Nicolle LE, Hoban DJ, Kazmierczak KM, Badal RE, Sahm DF. Susceptibility pat-
terns and ESBL rates of Escherichia coli from urinary tract infections in Canada and
the United States, SMART 2010–2014. Diagn Microbiol Infect Dis 2016 Aug;85(4):
459–65.
[8] Sanchez GV, Baird AM, Karlowsky JA, Master RN, Bordon JM. Nitrofurantoin retains
antimicrobial activity against multidrug-resistant urinary Escherichia coli from US
outpatients. J Antimicrob Chemother 2014;69(12):3259–62.
[9] Lerman B, Kobernick MS. Return visits to the emergency department. J Emerg Med
1987 Sep–Oct;5(5):359–62.
[10] Verelst S, Pierloot S, Desruelles D, Gillet JB, Bergs J. Short-term unscheduled return
visits of adult patients to the emergency department. J Emerg Med 2014 Aug;
47(2):131–9.
[11] Gabayan GZ, Asch SM, Hsia RY, Zingmond D, Liang LJ, Han W, et al. Factors associated
with short-term bounce-back admissions after emergency department discharge.
Ann Emerg Med 2013 Aug;62(2):136–44.
[12] Shelton P, Sager MA, Schraeder C. The community assessment risk screen (CARS):
identifying elderly persons at risk for hospitalization or emergency department
visit. Am J Manag Care 2000 Aug;6(8):925–33.
[13] Buurman BM, van den Berg W, Korevaar JC, Milisen K, de Haan RJ, de Rooij SE. Risk
for poor outcomes in older patients discharged from an emergency department: fea-
sibility of four screening instruments. Eur J Emerg Med 2011 Aug;18(4):215–20.
Please cite this article as: Jorgensen S, et al, Risk factors for early return visits to the emergency department in patients with urinary tract infection,
American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.06.041
[14] Moons P, De Ridder K, Geyskens K, Sabbe M, Braes T, Flamaing J, et al. Screening for
risk of readmission of patients aged 65 years and above after discharge from the
emergency department: predictive value of four instruments. Eur J Emerg Med
2007 Dec;14:315–23.
[15] Sin DD, Man SF. Low-dose inhaled corticosteroid therapy and risk of emergency de-
partment visits for asthma. Arch Intern Med 2002 Jul;162(14):1591–5.
[16] Newton AS, Rosychuk RJ, Niu X, Radomski AD, McGrath PJ. Predicting time to emer-
gency department return for anxiety disorders and acute stress reactions in children
and adolescents: a cohort study. Soc Psychiatry Psychiatr Epidemiol 2015 Aug;
50(8):1199–206.
[17] CLSI. Performance Standards for Antimicrobial Susceptibility Testing 25th Informa-
tional Supplement. CLSI document M100-S25 Wayne. PA Clinical and Laboratory
Standards Institute; 2015.
[18] Lowthian J, Straney LD, Brand CA, Barker AL, Smit Pde V, Newnham H, et al. Un-
planned early return to the emergency department by older patients: the safe elder-
ly emergency department discharge (SEED) project. Age Ageing 2016 Mar;45(2):
255–61.
[19] Keith KD, Bocka JJ, Kobernick MS, Krome RL, Ross MA. Emergency department re-
visits. Ann Emerg Med 1989 Sep;18(9):964–8.
[20] Gordon JA, An LC, Hayward RA, William BC. Initial emergency department diagnosis
and return visits: risk versus perception. Ann Emerg Med 1998 Nov;32(5):569–73.
[21] Gupta Kalpana, Hooton Thomas M, Naber Kurt G, Wullt Björn, Colgan Richard, Miller
Loren G, et al. International clinical practice guidelines for the treatment of acute un-
complicated cystitis and pyelonephritis in women: a 2010 update by the infectious
diseases society of america and the european society for microbiology and infectious
diseases. Clin Infect Dis 2011;52(5):e103–20.
[22] Rosa R, Abbo LM, Raney K, Tookes TE, Supino M. Antimicrobial resistance in urinary
tract infections at a large urban emergency department: factors contributing to em-
piric treatment failure. Am J Emerg Med 2017 Mar;35(3):397–401.
[23] Tandan M, Duane S, Cormican M, Murphy AW, Vellinga A. Reconsultation and anti-
microbial treatment of urinary tract infection in male and female patients in general
practice. Antibiotics (Basel) 2016 Sep 15;5(3).
[24] Bains A, Buna D, Hoag NA. A retrospective review assessing the efficacy and safety of
nitrofurantoin in renal impairment. Can Pharm J 2009 Sept;142(5):248–52.
[25] Cunha BA, Cunha CB, Lam B, Giuga J, Chin J, Zafonte VF, et al. Nitrofurantoin safety
and effectiveness in treating acute uncomplicated cystitis (AUC) inhospitalized pa-
tients with renal insufficiency: antibiotic stewardship implications. Eur J Clin
Microbiol Infect Dis 2017 Feb;2.
[26] The American Geriatrics Society 2015 Beers Criteria Update Expert Panel. 2015.
American geriatric society 2015 updated beers criteria for potentially inappropriate
medication use in older adults. J Am Geriatr Soc, 63(11):2227–46.
[27] Jorgensen S, Zurayk M, Yeung S, Terry J, Dunn M, Nieberg P, Wong-Beringer A. Emer-
gency department urinary antibiograms differ by specific patient groups. J Clin
Microbiol Accepted manuscript posted online 2017 June 14.