Intravenous Antibiotic Susceptibility for Urinary Tract Infection Prior to Emergency Department Discharge

Abstract

Background

Urinary tract infection (UTI) is a common infectious disease managed in the emergency department (ED). Patients may be initially treated with an intravenous (IV) antibiotic and subsequently discharged with an oral antibiotic regimen.

Objective

The purpose of this study was to determine whether the current Infectious Diseases Society of America guideline recommendation for an initial dose of long-acting IV antibiotic for treatment of UTI when the prevalence of fluoroquinolone resistance exceeds 10% improves the likelihood of providing in vitro susceptibility to the isolated uropathogen.

Methods

This was a retrospective study of patients in ED presenting between May 2009 and August 2018 who received treatment for UTI. The primary outcome was susceptibility of uropathogen to the IV antibiotic administered. Secondary outcomes included susceptibility to the oral antibiotic regimen prescribed at discharge, repeat health care visit within 30 days related to UTI follow-up, adverse events (AEs) associated with antibiotic use, and identification of risk factors associated with pathogen resistance.

Results

A total of 255 patients were included for analysis. Of these patients, 230 (90.2%) had pathogens susceptible to the administered IV antibiotic. The oral regimen susceptibility was 81.6% with 29 patients returning for UTI follow-up and 4 patients reporting AEs related to antibiotic use. Men and long-term care facility residents were more likely to have resistant uropathogens.

Conclusion

Administration of a long-acting IV antibiotic for treatment of UTI prior to ED discharge is recommended when the fluoroquinolone resistance rate exceeds 10% to improve in vitro susceptibility coverage.

Background

Urinary tract infection (UTI) is a common infectious disease encountered by emergency medicine clinicians accounting for more than 3 million emergency department (ED) visits annually. ¹ Once the diagnosis of UTI is established, management involves selection of appropriate empiric antimicrobial treatment, thereby making it one of the leading causes for antibiotic administration in the United States. ² The Infectious Diseases Society of America (IDSA) clinical practice guidelines for treatment of acute uncomplicated cystitis and pyelonephritis in women provide recommendations for antimicrobial selection and treatment duration. ³ Fluoroquinolones (FQs) are recommended as first-line therapy for patients with acute uncomplicated pyelonephritis if the prevalence of resistance to Escherichia coli is less than 10%. Unfortunately, antimicrobial resistance has presented challenges for the treatment of UTI with most geographic areas across the United States reporting rates of E coli resistance to guideline-recommended FQs of greater than 10% and some areas with rates greater than 20%.^3,4

In the setting of presumed FQ resistance of greater than 10%, IDSA guidelines recommend an initial dose of a long-acting intravenous (IV) antibiotic. Guidelines recommending long-acting IV antibiotic options include a third-generation cephalosporin (eg, ceftriaxone) or a consolidated 24-hour dose of an aminoglycoside (eg, gentamicin). ³ Administering a single dose of an IV antibiotic before discharge theoretically increases the likelihood of empiric activity, which is essential for a clinical and bacteriological response in the setting of E coli resistance. Unfortunately, the use of IV antibiotics before ED discharge has been associated with increased rates of antibiotic-associated diarrhea, ED length of stay and health care costs.^5-7

While IDSA guidelines endorse a single IV dose of antibiotic therapy before ED discharge, there is little evidence to suggest that this recommendation increases the likelihood of providing in vitro coverage of commonly isolated urinary pathogens. A previous study reported that a 1-time IV dose did not appear to improve clinical cure rates; however, not all patients included received an initial IV antibiotic, and resistance rates were not reported. ⁸ The objective of this study was to determine whether the current IDSA guideline recommendation to use an initial long-acting IV antibiotic for the treatment of UTI prior to discharge from the ED improves the likelihood of providing in vitro susceptibility to the isolated urinary pathogen.

Methods

A retrospective chart review was conducted at a single academic medical center with more than 60 000 ED adult patient visits annually presenting between May 2009 and August 2018 who received an IV antibiotic for the initial treatment of UTI and were subsequently discharged from the ED. Patients were included if they had a urine culture with resulted susceptibilities (>10 000 colony-forming units/mL and growth of no more than 2 organisms), received a single dose of long-acting IV therapy (eg, ceftriaxone or gentamicin) or ciprofloxacin for the treatment of UTI, and were subsequently discharged on an empiric oral antibiotic regimen from the ED. Patients were excluded if they had a history of renal transplant, active pregnancy, intention of treatment was for suspected sexually transmitted infection, or if the infection was catheter associated. The study protocol was approved by the local institutional review board.

Data Collection

Data collection was completed by a trained research assistant who was blinded to primary and secondary outcomes. Ten percent of patient charts were reviewed by a study investigator for quality control purposes. The following data were collected: age, sex, uropathogens, urine culture susceptibilities, past medical history (denoted by the physician in the history and physical) including diabetes, recurrent UTI, residence in a long-term care facility (LTCF), exposure to antibiotics within the previous 90 days, prophylactic antibiotic use for UTI, IV antibiotic administered, outpatient oral antibiotic prescribed upon discharge, return visits within 30 days of ED visit for infectious reason including return location (eg, outpatient clinic, ED), diagnosis, and adverse events (AEs) from antibiotic regimen.

The primary outcome was susceptibility of urinary tract pathogens to the IV antibiotic administered in the ED. Secondary outcomes included susceptibility of uropathogens to the oral antibiotic regimen prescribed at discharge, repeat health care visit within 30 days of discharge for UTI follow-up, assessment of potential patient risk factors for pathogen resistance, and AEs associated with antibiotic use. At our institution, IV ciprofloxacin is sometimes administered for UTI prior to ED discharge. A subgroup analysis was conducted to evaluate the impact of receiving IV ciprofloxacin compared with ceftriaxone on in vitro susceptibility of the isolated urinary pathogen.

Statistical Analysis

Logistic regression was used to model susceptibility to IV and outpatient oral treatment and to assess differences in susceptibility models for IV ciprofloxacin versus ceftriaxone. Frequency tables and univariate analyses were used to examine categorical and continuous variables, respectively. We used multivariable logistic regression analysis to determine the associations of baseline characteristics (sex, LTCF residence, diabetes, recurrent UTI, prophylactic antibiotic use, and antibiotic use within 90 days) on predicting the likelihood of susceptibility to IV and outpatient oral treatment. The design and results reporting were completed in accordance with the Strengthening and Reporting of Observational Studies in Epidemiology statement. ⁹ Statistical analyses were performed using Statistical Analysis System version 9.4.

Results

During the study period, 319 patients met inclusion criteria with 255 patients included for final analysis (Figure 1). The median age was 39 years (interquartile range 25-57), with the majority of the population being women (n = 214, 84%). Ceftriaxone was the primary IV antibiotic administered (n = 186, 73%) and ciprofloxacin was the most commonly prescribed outpatient oral antibiotic (n = 173, 67.8%). The most common pathogen isolated was E coli (n = 273, 78.4%). Four patients had urine cultures with reported susceptibilities of 2 organisms, resulting in a total of 259 uropathogens included for analysis (Table 1).

Figure 1.

Study participants and design.

Note. IV = intravenous; UTI = urinary tract infection; ED = emergency department.

Table 1.

Baseline Characteristics.

Characteristics	Value Total patients, n = 255
Age, median y	39 (interquartile range 25–57)
Female	214 (84)
Resistant pathogen risk factors
Antibiotics within 90 d	83 (33)
Diabetes mellitus	45 (17.6)
History of recurrent UTI	43 (16.9)
Residence in long-term care facility	17 (6.7)
Prophylactic antibiotics for UTI	10 (4)
Intravenous antibiotic administered
Ceftriaxone	186 (73)
Ciprofloxacin	69 (27)
Gentamicin	0 (0)
Empiric outpatient antibiotic prescribed
Ciprofloxacin	173 (67.8)
Sulfamethoxazole-trimethoprim	31 (12.2)
Cephalexin	19 (7.4)
Nitrofurantoin	15 (5.9)
Levofloxacin	6 (2.4)
Othera	11 (4.3)
Uropathogens identifiedb
Escherichia coli	203 (78.4)
Klebsiella pneumoniae	15 (5.8)
Proteus mirabilis	9 (3.5)
Enterobacter spp.	8 (3)
Enterococcus spp.	7 (2.7)
Staphylococcus aureus	4 (1.5)
Pseudomonas spp.	3 (1.2)
Klebsiella oxytoca	3 (1.2)
Otherc	7 (2.7)

Note. Values are reported as no. (%). UTI = urinary tract infection.

^aOther antibiotics prescribed included 4 cefdinir, 3 amoxicillin-clavulanate, 1 amoxicillin, 1 cefadroxil, 1 cefixime, and 1 doxycycline.

^bTotal of 259 uropathogens, given 4 patients had 2 uropathogens.

^cOther pathogens were 2 Serratia marcescens, 2 Citrobacter freundii, 1 Acinetobacter baumannii, 1 Citrobacter koseri, and 1 Staphylococcus lugdunensis.

Two hundred and thirty patients (90.2%) had urinary isolates that were susceptible to the IV antibiotic administered prior to ED discharge. Of those patients who received ceftriaxone, 91.4% of isolates were susceptible compared with 87% for those who received ciprofloxacin. Overall, 208 patients (81.6%) were susceptible to the empiric outpatient regimen prescribed (Table 2). A total of 22 patients (8.6%) demonstrated resistance to both the IV and oral antibiotic regimen.

Table 2.

Antibiotic Susceptibility Rates for Uropathogens.

Administered antibiotic susceptibilitya
	Intravenous (total)	Ceftriaxone	Ciprofloxacin	Empiric oral regimen
All urine isolates	230/255 (90.2)	170/186 (91.4)	60/69 (87)	208/255 (81.6)
Escherichia coli isolates	193/203 (95.1)	138/143 (96.5)	55/60 (91.7)	176/203 (86.7)
In vitro susceptibility
		Ceftriaxone	Ciprofloxacin
All urine isolates (n = 259b)		245/259 (94.6)	235/259 (90.7)
Escherichia coli isolates (n = 203)		199/203 (98)	183/203 (90)

Values are reported as no. (%)

^aPercentages were calculated for a total of 255 patients, given 4 patients had 2 uropathogens.

^bPercentages were calculated for total of 259 uropathogens.

There were 29 patients (11.4%) who had a return visit for UTI follow-up due to a lack of clinical improvement within 30 days of ED discharge (Figure 2). One patient who lacked in vitro susceptibility to both the IV and oral antibiotic therapy administered was admitted to an intensive care unit with concern for urosepsis. An additional 13 patients required hospital admission, while the remainder were evaluated in the outpatient setting and received additional IV antibiotic therapy or a change to their oral antibiotic regimen apart from 1 patient who had no additional changes.

Figure 2.

Return to health care for UTI follow-up within 30 days.

Note. Values are reported as no. (%). UTI = urinary tract infection.

^†Escherichia coli isolates 23/29 (79.3).

^††Intensive care unit admission (n =1), general medicine admission (n = 13).

^§Of the 22 patients (8.6%) who demonstrated resistance to both the IV and empiric oral antibiotic regimen, 1 patient returned for outpatient management and 1 patient returned for hospital admission.

Male sex increased the odds of pathogen resistance for IV antibiotic administered (odds ratio [OR] 5.8, 95% confidence interval [CI], 2.2-14.9; P < .001) and empiric outpatient oral treatment regimen (OR 3.0, 95% CI, 1.3-6.9; P < .001), while LTCF residence (OR 7.9, 95% CI, 2.6-24.2; P < .001) increased the odds of resistance to the empiric outpatient oral treatment regimen only (Table 3). Adverse events included 3 reports of diarrhea and 1 occurrence of gastritis.

Table 3.

Risk Factors Associated With Resistance Impact on Susceptibility to Antibiotic Therapy.

	Antibiotic susceptibility		Odds ratio (95% confidence interval)
Risk Factor	Factor present	Factor absent
Male
IV	29/41 (71)	201/214 (94)	5.8 (2.2–14.9)
PO	27/41 (65.8)	181/214 (84.6)	3.0 (1.3–6.9)
Antibiotics within 90 d
IV	72/83 (86.7)	156/172 (90.7)	1.9 (0.7–4.9)
PO	61/83 (73.5)	144/172 (83.7)	1.9 (0.9–4.2)
Diabetes mellitus
IV	41/45 (91.1)	189/210 (90)	0.7 (0.2–2.5)
PO	38/45 (84.4)	170/210 (81)	0.5 (0.2–1.5)
History of recurrent UTI
IV	38/43 (88.4)	191/212 (90)	1.2 (0.3–4.3)
PO	31/43 (72)	176/212 (83)	1.8 (0.7–4.4)
Residence in long-term care facility
IV	13/17 (76.5)	216/238 (90.8)	2.9 (0.8–11.3)
PO	7/17 (41)	200/238 (84)	7.9 (2.6–24.2)
Prophylactic antibiotics for UTI
IV	9/10 (90)	217/245 (88.6)	0.6 (0.06–5.7)
PO	8/10 (80)	197/245 (80.4)	0.7 (0.1–3.6)

Note. Values are reported as no. (%). IV = intravenous; UTI: urinary tract infection; PO = oral.

Discussion

In our study, administration of a single dose of IV antibiotic therapy (eg, ceftriaxone or ciprofloxacin) for UTI prior to ED discharge increased the likelihood of urine pathogen susceptibility over the empiric oral antibiotic regimen. Contrary to the IDSA guideline recommendations for treatment of acute uncomplicated cystitis and pyelonephritis in women when FQ resistance exceeds 10%, a single IV dose of ciprofloxacin is occasionally administered at our institution prior to discharge. Results of the subgroup analysis revealed that the use of IDSA guideline–concordant ceftriaxone as initial therapy would have further increased patients' susceptibility by an additional 3.9% to nearly 95% for treatment of all uropathogens.

During the study period, the reported hospital-wide FQ resistance rate of E coli at our institution ranged between 16% and 20%. We found that 10% of E coli uropathogen isolates were resistant to FQ therapy, which is similar to resistance rates of 9% and 12.1% reported in previous studies conducted by Zatorski et al and Talan et al, respectively.^4,10 The modest discrepancy in FQ antibiotic susceptibility among patients in ED compared with the hospital-wide antibiogram at our institution is worth noting. A recent study comparing susceptibility of E coli urinary isolates in ED and hospital-wide patient populations with uncomplicated UTI showed significant between-group differences of more than 18%. ¹⁰ Our institutional antibiogram does not differentiate where the isolate is collected (eg, inpatient vs outpatient) or the source from which the pathogen was collected (eg, urine and blood), which likely results in large discrepancies in resistance rates among various patient populations. Our findings highlight the importance of development of stratified antibiograms for empiric antibiotic selection according to the source of infection, patient population (eg, ED vs inpatient), and other patient-specific risk factors.

Our return follow-up rate related to infection of 11.4% is similar to previously reported literature assessing clinical outcomes. ⁸ Repeat health care visits and hospitalization for UTI can contribute to additional complications, increased resource utilization, and significant cost. ¹¹ Interestingly, all patients who had follow-up visits and initially received ceftriaxone demonstrated in vitro susceptibility and only 2 of the 10 patients who received a FQ had in vitro resistance. Furthermore, of the patients who returned for a follow-up visit, 75% of the oral regimens prescribed at ED discharge were susceptible to the uropathogen. These return visits for treatment failures are likely multifactorial and highlight the discrepancy between in vitro susceptibility, clinical cure rates, and achievement of adequate antimicrobial concentrations at the site of infection. With the aid of clinical practice guidelines, clinicians should select empiric treatment based on severity of illness (eg, cystitis vs pyelonephritis), coexisting medical conditions, history of drug allergy, potential for drug-drug interactions, and prior microbiologic data when available.

Antibiotic-associated diarrhea is a well-known AE of antibiotic therapy and administration of an initial dose of IV antibiotics prior to ED discharge has been associated with higher rates of diarrheal illness.^6,12 Similar to a previous study, our patient population had a low incidence of reported AEs, albeit the most commonly cited AE was diarrhea. ⁵

Previous evidence suggests that certain patient characteristics increase the likelihood of having a multidrug-resistant (MDR) pathogen.^13–15 Although only 16% of the study population was men, we found a 5- and 3-fold increase in the odds of having a resistant pathogen to the selected IV and oral antibiotic regimens, respectively. In addition, patients residing in a LTCF had an 8-fold increase in the odds of having a urinary pathogen resistant to the empiric outpatient antimicrobial therapy. The treatment of UTI in men remains challenging, given anatomical differences from women and predisposing factors to infections extending beyond the urinary tract. ¹⁶ Long-term care residence remains a significant risk factor for MDR pathogens due to colonization along with the inappropriate and overuse of antibiotics. ¹⁷ Unfortunately, current IDSA guidelines regarding treatment of UTI do not include specific recommendations pertaining to men or patients with significant risk factors for antimicrobial resistance.^3,17,18 Based on our findings, empiric treatment for men and residents of an LTCF requires alternative considerations.

Our study has important limitations. First, this was a retrorespective data analysis which introduces risk of bias due to inadequate documentation. However, we chose variables that would likely be available and documented accurately, although some elements (eg, return health care vistis, AEs) may not have been described. Second, there was an absence of confirmed UTI defined by the International Classification of Diseases (ICD) codes. The absence of UTI diagnosis determined by ICD codes should not influence findings, given each patient had a urine culture with resulted susceptibilities and received IV antibiotics. Third, the relatively small number of patients over the study period may have affected uptake of practice guidelines and changes in antimicrobial resistance and prescribing patterns. Finally, we were unable to determine the impact of the administered antimicrobial therapy on clinical outcomes. The main objective was to determine the impact of IV antibiotic therapy on the likelihood of in vitro susceptibility, which in clinical practice has been associated with the improved likelihood of clinical and bacterial cure rates. ⁴

Conclusion and Relevance

In the setting of reported E coli FQ resistance rates exceeding 10%, the administration of a single dose of a long-acting IV antibiotic (eg, ceftriaxone) provided greater than 90% in vitro susceptibility coverage for patients with UTI prior to ED discharge. Further studies should evaluate different treatment strategies for male patients and residents of LTCF as their resistance patterns are unique.