A risk model that accurately predicts pediatric oncology patients at risk for blood stream infections (BSIs) has been streamlined and has the potential to greatly impact antibiotic stewardship.
Research published in the American Cancer Society’s Cancer journal tests the risk model developed at Vanderbilt on a type of patient for whom there is little evidence-based data to guide optimal management. The research concludes that substantially fewer antibiotics may be needed than are typically prescribed for those patients.
The Esbenshade/Vanderbilt (EsVan) model for predicting risk of BSIs in pediatric oncology patients with fever had previously been tested internally at Monroe Carell Jr. Children’s Hospital at Vanderbilt and the latest published research represents the first external validation of the model.
“Our previous standard practice of giving empiric ceftriaxone resulted in antibiotic exposure to all with bacteremia in <10%,” said Adam Esbenshade, M.D., associate professor of Pediatrics. “Furthermore, the ceftriaxone was only covering 45 percent of the actual bacterial infections we isolated, and yet the vast majority of patients were well-appearing after the blood culture grew. The model supported our hypothesis that the majority of patients could be safely managed without empiric antibiotics, waiting for blood culture results.”
Model Provides First Guidelines for Patients without Severe Neutropenia
Evidence-based guidelines support antibiotic use for febrile pediatric oncology patients with severe neutropenia – defined as an absolute neutrophil count (ANC) <500/μL. The EsVan model was created to predict the likelihood of BSI in patients without severe neutropenia, defined as an ANC ≥500/μL, who did not have treatment guidelines until its development. For the external test of the EsVan model, two less reliable variables were removed, with the new model dubbed EsVan2. The EsVan2 model was tested at six external institutions and found that 85 percent of patients had a risk under 10 percent for developing BSIs. The results suggest that febrile pediatric oncology patients with ANC ≥500/μLmay not need empiricant antibiotics.
“The external validation allowed refinement of the model and now is reliable in identifying those at low risk of bacterial blood stream infections and ready for implementation at other institutions to guide management,” according to Esbenshade.
Great Promise for Limiting Antibiotic Use
“We have used these data to implement the model at the Monroe Carell Jr. Children’s Hospital at Vanderbilt and have eliminated use of empiric antibiotics in over 80 percent of low-risk fever episodes.”
In the EsVan model, tunneled external catheters, highest temperature reported or measured within 24 hours prior to presentation or within 1 hour after presentation, reported or observed chills, hypotension, and increased ANC were all determined to be variables that increased the risk of BSI.
Exposure to chemotherapy drugs that cause fever within 24 hours of the fever; older age; and acute lymphoblastic leukemia (ALL) were associated with decreased risk. Location at time of presentation, presence of upper respiratory symptoms and history of stem cell transplant were also included in the model. The EsVan2 model removed ALL diagnosis and location at presentation (inpatient vs. outpatient) as variables. The EsVan2 model was found to be more accurate in predicting BSI risk than the original model. Researchers pointed out that limiting antibiotics because of the model is not advised for subjects with any concern for severe sepsis. But the EsVan2 model does show great promise for guiding the management for predicting BSI for patients without severe neutropenia, limiting antibiotics, and identifying patients at high risk for BSIs.
“We have used these data to implement the model at the Monroe Carell Jr. Children’s Hospital at Vanderbilt and have eliminated use of empiric antibiotics in over 80 percent of low-risk fever episodes,” said Esbenshade. “With an international prospective validation and implementation study launched over the last year, the EsVan2model may be able to support clinical decision-making at pediatric oncology centers across the globe.”
