Several recent studies provide valuable insight surrounding the appropriateness of transitioning from intravenous to oral therapy when treating bacteremic urinary tract infections.
Gram-negative organisms are responsible for 25% to 50% of all bacteremias, with the urinary tract being the most common source.1-4 Many clinicians prefer to treat bacteremia with a complete course of intravenous (IV) antibiotic therapy and may be hesitant to switch patients to oral therapy.5-7 However, careful evaluation of the source, organism, and severity of infection can help identify patients who may be able to safely transition. Recently, literature evaluating the effectiveness of oral antibiotics for definitive therapy of bacteremic urinary tract infections (UTIs) has emerged.1,2,8 Several studies shed light on which oral agents may be appropriate choices, when clinicians should consider transitioning from IV to oral therapy, and which patients may be at a higher risk of failure upon transition. These findings are of interest to antimicrobial stewards, considering the potential for shorter hospital stays and reduced medication cost in patient groups who appear to have high success rates with definitive oral therapy.9
One of clinicians’ common concerns regarding oral antibiotic treatment of bacteremia is agent bioavailability.1,7 Several oral antibiotics achieve both excellent concentrations in the urinary tract and moderate to high systemic concentrations. Agents with low bioavailability, such as β-lactam antibiotics, may not provide adequate plasma exposure for the treatment of bacteremia because dosing is limited by tolerability. Optimal dosing and probability of target attainment of these low-bioavailability drugs are often understudied. As such, some oral antibiotics may produce better clinical outcomes than others. Two recent studies evaluated the effect of oral agent bioavailability on clinical outcomes of patients with gram-negative bacteremias.1,2
Kutob and colleagues evaluated the effect of oral bioavailability on clinical outcomes of gram-negative bacteremias, about 70% of which were from a urinary source (254 of 362 patients).1 Bioavailability was classified as high (≥95%, levofloxacin), moderate (75%-94%, ciprofloxacin and trimethoprim/sulfamethoxazole), and low (<75%, β-lactam antibiotics). Treatment failure rates were 2%, 12%, and 14% in patients receiving oral antibiotics with high, moderate, and low bioavailability, respectively (P = .02). Ciprofloxacin was the most common moderate bioavailability drug used, with 84% of patients receiving a 500-mg twice daily dose. After adjusting for end stage renal disease, cirrhosis, immunosuppression, central catheters, urologic complications, and resistant pathogens, low bioavailability agents were significantly associated with treatment failure, compared with high bioavailability agents (adjusted HR, 7.67; 95% CI, 1.9-51.5). Moderate bioavailability agents were also associated with increases in treatment failure, although differences between success within the fluoroquinolone class may in part be due to underdosing of ciprofloxacin or better compliance with once-daily dosing of levofloxacin.10,11 Of note, patients with a urinary source of bacteremia had a low rate of treatment failure overall (8%). No difference was found in failure rates between early (<5 days IV therapy) versus late (≥5 days) oral transition (10% vs 9%; P = .73). The mean duration of treatment in all 3 groups was 13.8 days (4.7 days IV followed by 9.1 days oral).
Mercuro and colleagues similarly compared fluoroquinolones versus β-lactams for oral step-down therapy of Enterobacteriaceae bacteremias.2 The urinary tract was again the most common source of infection: 69.4% (58 of 84 patients) in the β-lactam group and 72.3% (101 of 140 patients) in the fluoroquinolone group (P >.05). The authors observed similar rates of clinical success between early (≤3 days) and late (>3 days) step-down (86.7% vs 87.5%; P >.05). No difference was found in clinical success between the β-lactam and fluoroquinolone groups (86.9% vs 87.1%; mean difference, 0.2%; 97.5% CI, -10.3 to -10.7). Additionally, β-lactams were better tolerated than fluoroquinolones (91.7% vs 82.1%; P = .049).
Rieger and colleagues retrospectively evaluated the effectiveness of IV transitioned to oral antimicrobials for Enterobacteriaceae-associated bacteremic UTIs, although bioavailability was not a specific focus of the study.8 Failure rates were assessed among patients with positive urine and blood cultures receiving IV only (n = 106) versus IV transitioned to oral antibiotics (IV/oral; n = 135). The median number of days on IV therapy before switching to oral therapy was (interquartile range [IQR], 2-5), although total days of therapy was not reported. Discharge oral antibiotics were primarily ciprofloxacin (65.1%), a β-lactam (19%), or trimethoprim/sulfamethoxazole (9.1%) in the IV/oral group. In the IV-only group, discharge antibiotics were most commonly ceftriaxone (79.3%), cefepime (5.8%), and ertapenem (5.8%). Patients in the IV-only group were more likely to be admitted to the intensive care unit (39.6% vs 27.4%; P = .045), be infected with a multidrug-resistant pathogen (39% vs 23%; P = .01) and have a higher Charlson comorbidity index (CCI) (median [IQR], 8 [5-11] vs 6 [4-9]; P = .03). No differences in clinical failure were observed between treatment with IV-only versus IV/oral (3.8% vs 8.2%; P = .19). Neither treatment group nor CCI was significantly associated with treatment failure in multivariate analyses. Of note, 4 of 6 patients prescribed an oral fluoroquinolone at discharge who experienced treatment failure were also discharged on a polyvalent cation, potentially resulting in decreased systemic fluoroquinolone exposure. Hospital length of stay was significantly shorter in the IV/oral group (median [IQR], 4.6 [3.1-7.8] vs 7.1 [4.0-17.5]; P <.001); this difference remained significant after adjustment for CCI (P = .004).
Differences in results between the studies may be attributed to several factors. Institution-specific criteria in place for IV to oral conversion, as well as adequate source control, may have played a role in clinical success. With respect to medication adherence, Mercuro was able to confirm outpatient receipt of medications in 67.5% of patients, whereas adherence was not reported in others.2 Additionally, the follow-up period for infection recurrences differed between studies. Both Rieger and Mercuro followed patients for 30 days post discharge and completion of therapy, whereas Kutob had a 90-day follow-up period.1,2,8 A 90-day follow-up has been proposed as the optimal duration for bacteremias.12 The median time to treatment failure in the study by Kutob and colleagues was past the 30-day mark, at 35 days for mortality (range, 15-68 days) and 37 days for recurrence (range, 10-69 days). Because of this, the other 2 studies may have failed to capture possible treatment failures.
In conclusion, data from these retrospective studies suggest that fluoroquinolones (particularly those with high bioavailability) are the most consistently effective oral options for treating patients with bacteremic UTIs, generally for a total duration of 14 days. This strategy, however, must be weighed against the risk of adverse events associated with fluoroquinolone use.13 It appears that early transition to oral therapy (ie, within 3-5 days) is appropriate in clinically stable patients. This strategy was also associated with a shorter hospital stay. Oral β-lactam agents may be effective in some patients; however, their true place in the treatment of bacteremic UTIs is not well defined. When deciding on oral therapy, optimal dosing and likelihood of patient adherence must be considered. Lastly, patient-specific factors including urinary abnormality, complications of diabetes, liver cirrhosis, and immunocompromised status may predispose patients to treatment failure. Caution should be exercised when considering transition oral therapy in high-risk patients.
See Table of Study Characteristics
Dr. Mangino is a PGY-1 pharmacy practice resident at St. Peter’s Hospital in Albany, New York. She received a PharmD from Albany College of Pharmacy and Health Sciences also in Albany. Upon completion of her PGY- 1 residency, she will be pursuing a PGY-2 pharmacy residency in infectious diseases.
Dr. Bidell is an assistant professor of pharmacy practice at Albany College of Pharmacy and Health Sciences in Albany, New York. She maintains a clinical practice at St Peter’s Hospital, also in Albany.
Dr. O’Donnell is an assistant professor of pharmacy practice at Albany College of Pharmacy and Health Sciences in Albany, New York. He maintains a clinical practice at Albany Medical Center, also in Albany.