Optimizing the Management of Uncomplicated Gram-Negative Bloodstream Infections in Children: Translating Evidence From Adults Into Pediatric Practice Free

Gram-negative bloodstream infections (BSI) continue to cause significant morbidity and death in children. Uncomplicated Gram-negative BSI has been treated traditionally with approximately 14 days of intravenous (IV) antibiotics, but data to support this practice are lacking [1]. Prolonged durations of IV therapy lead to patient inconvenience and missed days of school or work [2]. They also increase the risk for adverse drug events, the development of antibiotic resistance, and Clostridioides difficile infections [3]. In children, prolonged courses of therapy for BSI have been associated with an increased risk of candidemia [4]. Moreover, longer durations of IV therapy and the subsequent need for a peripherally inserted central catheter can lead to secondary infections and thrombotic complications [5]. These risks are justifiable in the setting of a strong foundation of evidence that supports the need for prolonged IV therapy for Gram-negative BSI. It is important to explore whether such an evidence base exists.

Over the past several years at The Johns Hopkins Hospital, we have recommended early transition to oral therapy and generally 7 days of therapy for most children with uncomplicated Gram-negative BSI. We define “uncomplicated” BSI as bacteremia without evidence of endocarditis, endovascular disease, central nervous system infection, or osteoarticular infection. Based on the available published evidence and our clinical experience, herein we discuss our recommended approach to the management of uncomplicated Gram-negative BSI in children.

IDENTIFYING THE SOURCE OF INFECTION AND SOURCE-CONTROL MEASURES

Identifying the source of bacteremia is a critical first step in management. Common BSI sources are urinary, intra-abdominal, respiratory, catheter-related, and skin and soft-tissue infections (Table 1). Adequate source control is defined as the removal of any infected hardware or device, drainage of infected fluid collections, and resolution of biliary or urinary obstruction. Patients who lack adequate source control have largely been excluded from studies, which makes extrapolations from available data to these populations challenging. Although no existing data indicate that a lack of source control precludes either the conversion to oral therapy or the prescription of a shorter course of therapy, clinicians should feel reassured that available data show that patients who have achieved source control can generally be treated with oral antibiotics and 7 days of therapy [4, 6–8].

A frequently encountered scenario in the pediatric population is the retention of central lines. We endeavor to distinguish catheter-related BSI (CRBSI) from other potential sources resulting in bacteremia in a patient who also has a central line. For example, if a patient with a central line has an Escherichia coli BSI and associated urinary symptoms, line removal is not pursued except in the setting of prolonged bacteremia (“prolonged bacteremia” is generally characterized as >48 hours of positive blood cultures in the presence of in vitro active antibiotic therapy). In the absence of prolonged bacteremia, line retention in a patient with uncomplicated BSI and an alternative infectious source is not considered a failure of source control and does not affect our decision regarding the duration or route of therapy. In contrast, all efforts should be made to remove any vascular catheter thought to be the source of bacteremia; however, line removal is not always feasible. In patients with CRBSI in whom the central line is retained, we often continue IV therapy for the full duration of treatment (in general, 7 days total) for the theoretical benefit of administering antibiotics directly into an infected line. When a central line is removed, we transition to oral therapy once other criteria (enumerated below) are met and treat the patient for a total of 7 days.

TRANSITIONING FROM IV TO ORAL THERAPY

Transitioning hospitalized adults with Enterobacteriaceae bacteremia to oral therapy has resulted in shorter lengths of hospital stay without any increases in the 30-day mortality rate [10, 11]. Early appropriate IV antibiotics are critically important in the treatment of Gram-negative BSI, and IV therapy should continue for the duration of hemodynamic instability [12, 13]. Transition to oral therapy should be entertained as soon as certain conditions are met. Our criteria for converting a patient with uncomplicated Gram-negative BSI to oral therapy include the following: (1) return to baseline clinical status (eg, afebrile, normotensive, baseline respiratory and mental status), (2) ability to consume and absorb oral antibiotics, (3) no evidence of persistent bacteremia or metastatic disease (ie, foci of infection distant to the primary site of infection), and (4) availability of an appropriate oral agent with good bioavailability and antibiotic susceptibility confirmation (Table 2). If any of these criteria are not met (eg, impaired intestinal absorption), we recommend continuing IV antibiotics for the duration of therapy.

The question of compliance with treatment is more nuanced. Oral therapy can improve adherence because it is less disruptive for parents and children at home and can lead to earlier hospital discharge [10, 11]. Alternatively, home nursing visits can ensure more reliable administration of IV antibiotics, but home visits are not often an option for patients who transition to oral therapy. No studies have compared compliance with antibiotic therapy in the outpatient setting depending on route of administration. We incorporate patient and family preferences and our perceived likelihood of compliance with oral medications into our decision-making process when considering discharging a child from the hospital with continued antibiotic therapy.

Neonates are less likely to meet the criteria for conversion to oral therapy and are therefore, by default, more likely to continue IV therapy for the completion of their antibiotic course. However, we consider transition to oral therapy in clinically stable neonates with a short duration of bacteremia (no longer than 1–2 days), normal cerebrospinal fluid cell counts and chemistry results along with negative cerebrospinal fluid culture results, no evidence of metastatic disease, and rapid clinical improvement on a case-by-case basis.

SELECTING THE APPROPRIATE ORAL ANTIBIOTIC AGENT

Data regarding the effect of the bioavailability of antibiotic agents on clinical success in the treatment of Gram-negative BSI are inconclusive. Studies have failed to find a consistent difference in favorable clinical responses between antibiotics with relatively low (eg, oral third-generation cephalosporins) and those with high (eg, oral fluoroquinolones) bioavailability [10, 15, 16] (Table 2). However, most studies have not been adequately powered to address this issue. Furthermore, the timing of transition and source of infection have been variable across studies; bioavailability might be less important for transitions at later time points once the burden of infection has been reduced adequately.

We generally prescribe antibiotics with high bioavailability and sustained bloodstream concentrations for the treatment of bacteremia and preferentially select fluoroquinolones or trimethoprim-sulfamethoxazole. The possible adverse events associated with fluoroquinolones, particularly tendinopathies, have received a lot of attention (Table 2). Although potential antibiotic-associated adverse events should always be factored into the decision-making process and reviewed with patients and their families, they should be weighed against the considerable risks associated with the prolonged use of a vascular catheter [17].

Although most studies have not assessed the use of specific agents or classes of antibiotics as oral options for BSI, there are more data regarding transitioning patients with urinary tract infection to oral therapy than patients with other infectious sources. We generally limit the use of oral β-lactams to those with urinary tract infections and associated bacteremia and are often more reticent to do so for those with other sources of infection. When prescribing oral β-lactams in the setting of uncomplicated BSI, we typically use dosages at the higher end of dosage ranges and at the shortest dosing interval to maximize time of the free drug above the minimum inhibitory concentration. It should be noted that although tetracyclines such as doxycycline are highly bioavailable, their blood concentrations are not sustained. Therefore, we find them to be suboptimal for the treatment of BSI.

There are other considerations when selecting oral agents, including drug allergies and potential interactions with supplements, nutrition, or other agents (Table 2). These considerations are particularly important for the fluoroquinolones, which are poorly absorbed if administered in close proximity to the intake of a calcium-, magnesium-, aluminum-, zinc-, or iron-containing product (eg, antacids, dairy products, tube feeds).

SELECTING THE DURATION OF THERAPY

Data supporting the position that approximately 7 days of therapy are effective for most cases of uncomplicated Gram-negative BSI are mounting [4, 6–8]. Retrospective studies in both children and adults have yielded similar results for Enterobacteriaceae and Pseudomonas aeruginosa BSI [4, 7, 8]. A recent randomized controlled trial of uncomplicated Gram-negative bacteremia found no difference in mortality rates or suppurative complications in adults who received 7 days of antibiotics and those who received 14 days of antibiotics [6]. Furthermore, patients who received a shorter course had an earlier time to return to baseline activity. Although available data are drawn primarily from adult populations, a number of patients in these studies had complex medical issues or were immunosuppressed. There is little reason to suspect that these conclusions cannot be applied to children, who have similar risk factors for BSI and are often healthier at baseline. Therefore, we generally limit antibiotic therapy for uncomplicated Gram-negative BSI to 7 days in children who (1) have had no more than 2 days of bacteremia, (2) have achieved source control, and (3) have shown clinical improvement within the first few days of IV therapy. Our practice has been to define a 7-day treatment course for uncomplicated BSI as 7 total days of active antibiotic therapy rather than beginning from the resolution of bacteremia or source control.

We consider a duration of 7 days even for children with a retained central line. Relapse is more common in children with Gram-negative BSI and a retained central line; 1 study found a 26% likelihood of relapse among children with a retained line compared to 1% in children whose line has been removed [4]. No data indicate that this risk is mitigated by prolonging therapy. Our practice has not been to extend the duration of therapy in the absence of line removal. We generally consider the recurrence of bacteremia in children with a retained central line to be a failure of source control and advocate for line removal in these instances rather than prolonging the duration of therapy.

Most of the data evaluating duration of therapy for Gram-negative BSI have been limited to immunocompetent patients. However, some studies have also included immunocompromised patients, such as neutropenic patients and recipients of a hematopoietic stem cell or solid organ transplant [4, 7]. Although the degree and anticipated duration of immunosuppression should be considered in determining treatment duration, we generally advocate for 7 to 10 days of therapy for immunocompromised patients with uncomplicated Gram-negative BSI, including those who remain neutropenic at the end of therapy.

CONCLUSIONS

The body of evidence supporting the use of a shorter duration of therapy and earlier transition to oral therapy in the management of uncomplicated Gram-negative BSI is growing. We believe that increasing the application of these strategies to children who meet appropriate clinical criteria will reduce risks associated with the prolonged use of IV antibiotics.

Notes

Financial support. This work was supported by funding from the National Institutes of Health (grants K23-AI127935 [to P. D. T.] and T32-AI052071 [to R. G. S.]).

Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References