A 42-year-old woman with uncontrolled diabetes presents to the ED with fever, chills, dysuria, and flank pain for 3 days. On exam, she is febrile and tachycardic. Lab results show leukocytosis and urinalysis is consistent with infection. CT scan shows acute pyelonephritis without complication. She is admitted to the hospital and started on ceftriaxone 2 g/24 hrs. On hospital day 2, her blood cultures show gram-negative bacteria.
Management of gram-negative (GN) bacteremia remains a challenging clinical situation for inpatient providers. With the push for high-value care and reductions in length of stay, recent literature has focused on reviewing current practices and attempting to standardize care. Despite this, no overarching guidelines exist to direct practice and clinicians are left to make decisions based on prior experience and expert opinion. Three key clinical questions exist when caring for a hospitalized patient with GN bacteremia: Should blood cultures be repeated? When is transition to oral antibiotics appropriate? And for what duration should antibiotics be given?
Overview of the data
When considering repeating blood cultures, it is important to understand that current literature does not support the practice for all GN bacteremias.
Canzoneri et al. retrospectively studied GN bacteremia and found that it took 17 repeat blood cultures being drawn to yield 1 positive result, which suggests that they are not necessary in all cases.1 Furthermore, repeat blood cultures increase cost of hospitalization, length of stay, and inconvenience to patients.2
However, Mushtaq et al. noted that repeating blood cultures can provide valuable information to confirm the response to treatment in patients with endovascular infection. Furthermore, they found that repeated blood cultures are also reasonable when the following scenarios are suspected: endocarditis or central line–associated infection, concern for multidrug resistant GN bacilli, and ongoing evidence of sepsis or patient decompensation.3
Consideration of a transition from intravenous to oral antibiotics is a key decision point in the care of GN bacteremia. Without guidelines, clinicians are left to evaluate patients on a case-by-case basis.4 Studies have suggested that the transition should be guided by the condition of the patient, the type of infection, and the culture-derived sensitivities.5 Additionally, bioavailability of antibiotics (see Table 1) is an important consideration and a recent examination of oral antibiotic failure rates demonstrated that lower bioavailability antibiotics have an increased risk of failure (2% vs. 16%).6
In their study, Kutob et al. highlighted the importance of choosing not only an antibiotic of high bioavailability, but also an antibiotic dose which will support a high concentration of the antibiotic in the bloodstream.6 For example, they identify ciprofloxacin as a moderate bioavailability medication, but note that most cases they examined utilized 500 mg b.i.d., where the concentration-dependent killing and dose-dependent bioavailability would advocate for the use of 750 mg b.i.d. or 500 mg every 8 hours.
The heterogeneity of GN bloodstream infections also creates difficulty in standardization of care. The literature suggests that infection source plays a significant role in the type of GN bacteria isolated.6,7 The best data for the transition to oral antibiotics exists with urologic sources and it remains unclear whether bacteria from other sources have higher risks of oral antibiotic failure.8
One recent study of 66 patients examined bacteremia in the setting of cholangitis and found that, once patients had stabilized, a switch from intravenous to oral antibiotics was noninferior, but randomized, prospective trials have not been performed. Notably, patients were transitioned to orals only after they were found to have a fluoroquinolone-sensitive infection, allowing the study authors to use higher-bioavailability agents for the transition to orals.9 Multiple studies have highlighted the unique care required for certain infections, such as pseudomonal infections, which most experts agree requires a more conservative approach.5,6
Fluoroquinolones are the bedrock of therapy for GN bacteremia because of historic in vivo experience and in vitro findings about bioavailability and dose-dependent killing, but they are also the antibiotic class associated with the highest hospitalization rates for antibiotic-associated adverse events.8 A recent noninferiority trial comparing the use of beta-lactams with fluoroquinolones found that beta-lactams were noninferior, though the study was flawed by the limited number of beta-lactam–using patients identified.8 It is clear that more investigation is needed before recommendations can be made regarding ideal oral antibiotics for GN bacteremia.
The transition to oral is reasonable given the following criteria: the patient has improved on intravenous antibiotics and source control has been achieved; the culture data have demonstrated sensitivity to the oral antibiotic of choice, with special care given to higher-risk bacteria such as Pseudomonas; the patient is able to take the oral antibiotic; and the oral antibiotic of choice has the highest bioavailability possible and is given at an appropriate dose to reach its highest killing and bioavailability concentrations.7
After evaluating the appropriateness of transition to oral antibiotics, the final decision is about duration of antibiotic therapy. Current Infectious Disease Society of America guidelines are based on expert opinion and recommend 7-14 days of therapy. As with many common infections, recent studies have focused on evaluating reduction in antibiotic durations.
Chotiprasitsakul et al. demonstrated no difference in mortality or morbidity in 385 propensity-matched pairs with treatment of Enterobacteriaceae bacteremia for 8 versus 15 days.10 A mixed meta-analysis performed in 2011 evaluated 24 randomized, controlled trials and found shorter durations (5-7 days) had similar outcomes to prolonged durations (7-21 days).11 Recently, Yahav et al. performed a randomized control trial comparing 7- and 14-day regimens for uncomplicated GN bacteremia and found a 7-day course to be noninferior if patients were clinically stable by day 5 and had source control.12
It should be noted that not all studies have found that reduced durations are without harm. Nelson et al. performed a retrospective cohort analysis and found that reduced durations of antibiotics (7-10 days) increased mortality and recurrent infection when compared with a longer course (greater than 10 days).13 These contrary findings highlight the need for provider discretion in selecting a course of antibiotics as well as the need for further studies about optimal duration of antibiotics.