Comparative Outcomes of Pediatric Hematologic Malignancy Patients with Fluoroquinolone-Resistant vs. Susceptible Gram-Negative Bacteremia

Background

Bacterial bloodstream infections (BSI) are a leading cause of morbidity and mortality in children with neutropenia secondary to chemotherapy for hematological malignancies (1, 2). Gram-negative organisms are the etiology of approximately 40% of neutropenic BSI events and are potentially more pathogenic than some of the gram-positive organisms. Levofloxacin prophylaxis significantly reduces bacteremia incidence in pediatric patients with prolonged neutropenia after receipt of chemotherapy for hematological malignancy, but breakthrough infections still occur in approximately 20% of patients (3). Many of these breakthrough infections are fluoroquinolone resistance which has led to concern about the negative consequences of fluoroquinolone prophylaxis. However, the impact of fluoroquinolone resistance on outcomes among patients with gram-negative BSI has not been assessed. This study aims to compare outcomes in pediatric hematologic malignancy patients with fluoroquinolone-resistant vs. fluoroquinolone-susceptible gram-negative bacteremia.

Methods

This observational cohort study of patients with neutropenia associated gram-negative BSI during was assembled from pediatric hematologic malignancy patients cared for at the Children’s Hospital of Philadelphia from 2014 to 2023. The exposure was gram-negative bacteremia, dichotomized as fluoroquinolone-resistant vs. fluoroquinolone-susceptible. Primary outcomes were ICU admission and mortality, while secondary outcomes included the need for at least two fluid boluses, use of vasopressors, and non-invasive mechanical ventilation (NIMV) or intubation/mechanical ventilation. All outcomes were evaluated from the day before to 30 days after the positive blood culture was obtained. In an effort to account for confounding by indication, propensity score adjustment, using inverse probability of treatment weighting (IPTW), was performed to balance the following baseline covariates (age, year of diagnosis, time from malignancy onset, cumulative neutropenia, and G-CSF use). IPTW weighted logistic regression models were constructed to compare outcomes between the two groups.

Results

Among 119 pediatric patients with gram-negative bacteremia, 82 (68.9%) had a fluoroquinolone-susceptible BSI, and 37 (31.1%) had a fluoroquinolone-resistant BSI. After IPTW, all covariates were well balanced (standard mean difference <0.1) (Table 1). Outcomes were worse in the resistant group across all parameters, but only mortality reached statistical significance (Table 2). ICU admissions were more frequent (28.5% vs. 44.5%, OR: 2.01, 95% CI: 0.75–5.44, p = 0.166), as was the need for at least two fluid boluses (33.9% vs. 43.9%, OR: 1.53, 95% CI: 0.58–4.04, p = 0.391), need for vasopressors (19.8% vs. 47.9%, OR: 1.95, 95% CI: 0.66–5.77, p = 0.391), and NIMV or intubation/mechanical ventilation (16.8% vs. 34.6%, OR: 2.62, 95% CI: 0.83–8.32, p = 0.100). Mortality was significantly higher in the resistant group (2.1% vs. 17.9%, OR: 10.24, 95% CI: 1.62–64.56, p = 0.014).

Conclusions

Fluoroquinolone-resistant gram-negative bacteremia in pediatric hematologic malignancies was associated with worse clinical outcomes. Despite attempts to adjust for confounding by indication using IPTW informed models, residual confounding remains a concern. Post hoc analyses of this cohort and analyses of other larger cohorts are planned to further assess for residual confounding and to better understand the impact of resistance on outcomes in this vulnerable population.

References