Discussion
Empiric antibiotic selection for the treatment of UTIs continues to be the cornerstone of antibiotic management for the treatment of such a disease state.7 The noted drug-bug match rate of 88% in this study demonstrates effective initial empiric coverage and ensures a vast majority of veterans receive adequate coverage for identified pathogens. Additionally, this rate shows that the current system seems to be functioning appropriately and refutes the author’s preconceived ideas that the mismatch rate was higher at RRVAMC. However, these findings also demonstrate a predominant use of fluoroquinolones for empiric treatment in a majority of patients who could be better served with narrower spectrum agents. Only 2 of the outpatient regimens were for the treatment of pyelonephritis, the only indication in which a fluoroquinolone would be the standard of care per guideline recommendations.7
These findings were consistent with a similar study in which 83% of ED collected urine cultures ultimately grew bacteria susceptibleto empiric treatment.8 This number was similar to the current study despite the latter study consisting of predominantly female patients (93%) and excluding patients with a history of benign prostatic hypertrophy, catheter use, or history of genitourinary cancer, which are frequently found within the VA population. Thus, despite having a differing patient population at the current study’s facility with characteristics that would classify most to be treated as a complicated UTI, empiric coverage rates remained similar. The lower than anticipated intervention rate by the pharmacist on rotation in the ED can be directly attributed to this high empiric match rate, which could in turn be attributed to the extensive use of broad-spectrum antibiotics for treatment.
Empiric antimicrobial selection is based largely on local resistance patterns.7 Of particular importance is the resistance patterns of E coli, as it is the primary isolate responsible for UTIs worldwide. Thus, it is not unexpected that the most frequently isolated pathogen in the current study also was E coli. While clinical practice guidelines state that hospital-wide antibiograms often are skewed by cultures collected from inpatients or those with complicated infection, the current study found hospital-wide E coli resistance patterns, specifically those related to fluoroquinolone use, to be similar to those collected in the ED alone (78.5% hospital-wide susceptibility vs. 75% ED susceptibility). This was expected, as similar studies comparing E coli resistance patterns from ED-collected urine cultures to those institution-wide also have found similar rates of resistance.8,9 These findings are of particular importance as E coli resistance is noted to be increasing, varies with geographic area, and local resistance patterns are rarely known.7 Thus, these findings may aid ED providers in their empiric antimicrobial selections.
Ciprofloxacin was the most frequently used medication for the treatment of UTIs. While overall empiric selections were found to have favorable resistance patterns, it is difficult to interpret the appropriateness of ciprofloxacin’s use in the present study. First, there is a distinct lack of US-based clinical practice guidelines for the treatment of complicated UTIs. As the majority of this study population was male, it is difficult to directly extrapolate from the current Infectious Diseases Society of America treatment guidelines for uncomplicated cystitis and apply to the study population. Although recommended for the treatment of pyelonephritis, it is unclear whether ciprofloxacin should be utilized as a first-line empiric option for the treatment of UTIs in males.
Despite the lack of disease-specific recommendations for ciprofloxacin, recommendations exist regarding its use when local resistance patterns are known.7 It is currently recommended that these agents not be used when resistance rates of E coli exceed 20% for trimethoprim-sulfamethoxazole or 10% for fluoroquinolones. As this study demonstrated a nearly 25% resistance rate for E coli to fluoroquinolones in both the ED and institution-wide sample populations, it could potentially be ascertained that ciprofloxacin is an inappropriate choice for the empiric treatment of UTIs in this patient population. However, as noted, it is unknown whether this recommendation would still be applicable when applied to the treatment of complicated cystitis and greater male population, as overall rates of susceptible cultures to all organisms was similar to other published studies.8,9
While there is scant specific guidance related to the treatment of complicated UTIs, there is emerging guidance on the use of fluoroquinolones, both in general and specifically related to the treatment of UTIs. In July 2016, the FDA issued a drug safety communication regarding the use of and warnings for fluoroquinolones, which explicitly stated that “health care professionals should not prescribe systemic fluoroquinolones to patients who have other treatment options for acute bacterial sinusitis, acute bacterial exacerbation of chronic bronchitis, and uncomplicated UTIs because the risks outweigh the benefits in these patients.”10
This guidance has the potential to impact fluoroquinolone prescribing significantly at RRVAMC. Given the large number of fluoroquinolones prescribed for UTIs, the downstream effects that this shift in prescribing would have is unknown. As most nonfluoroquinolones used for UTI typically are narrower in antimicrobial spectrum (eg, trimethoprim/sulfamethoxazole, nitrofurantoin, cephalexin, etc) the possibility exists that the match rate for empiric therapy may decrease. Thus, a larger need for closer follow-up to assure adequate coverage may arise, posing a more expanded role for an ED-based pharmacist than was demonstrated in the current study.
This new guidance also may place providers in an area of larger uncertainty with regards to treating both complicated and uncomplicated cystitis. Given the enhanced warnings on fluoroquinolone use, it is unknown whether prescribers would gravitate to utilizing similar options as their peers as alternatives to fluoroquinolones. Similarly, duration of therapy with nonfluoroquinolone agents is unclear as well; as the present study demonstrated a large range in treatment duration of outpatients (3-14 days). While the average observed duration of 8.3 days is intuitively fitting, as the majority of cases were in males, no published guideline exists that affirms the appropriateness of this finding. Such uncertainty and potential inconsistency between providers affords a large opportunity for developing a standardized treatment pathway for the treatment of UTIs to ensure both effective and guideline concordant treatment for patients, specifically with regards to antimicrobial selection and duration of treatment.
It is noteworthy to mention that all follow-ups on positive cultures inadequately covered by empiric therapy took place on the day organism identification and susceptibility data were released. This finding was somewhat surprising, as it was originally theorized that most ED-collected urine cultures were not monitored to completion by a pharmacist and that would be necessary in order to ensure proper follow-up of culture results. What is not clear is whether there is a robust process for the follow-up of urine cultures in the ED. Most of the bug-drug mismatches coincidentally were admitted to the inpatient teams where there were appropriate personnel to follow up and adjust the antibiotic selection. If there was a bug-drug mismatch, and the patient was not admitted, it is unclear whether there is a consistent process for follow-up.
Given the limited number of mismatched cultures that required change in therapy, it is unknown if this role would expand if more narrow-spectrum agents were utilized, theoretically leading to a higher mismatch rate and necessitating closer follow-up. Furthermore, given the common practice of mailed prescriptions at the VA, it is all the more imperative that the cultures be acted upon on the day they were identified, as the mailing and processing time of prescriptions may limit the clinical utility in switching from a more broad-spectrum agent, to one more targeted for an identified organism. While a patient traveling back to the medical center for expedited prescription pickup at the pharmacy would alleviate this problem, many patients at the facility travel great distances or may not have readily available travel means to return to the medical center.