Prospective Study Of An Emergency Department Febrile Neutropenia Pathway In Patients With Hematologic Malignancies

Febrile neutropenia (FN) is an oncologic emergency associated with high morbidity and mortality, particularly in patients (pts) with hematologic malignancies. Delays in antibiotic administration, which can occur in busy emergency departments (EDs), lead to worse outcomes. We instituted a FN pathway (FNP) in the Cleveland Clinic (CC) ED to reduce antibiotic delays.

This prospective study comparing patients from 06/12 - 06/13 to historical pts from 02/10 - 05/12, represented a collaboration among cancer center, ED, infectious disease, pharmacy, and electronic medical record representatives. Fever was defined as temperature >38°C either at home or in the ED, while neutropenia as absolute neutrophil count <0.5 x 10⁹/L. All CC cancer pts received a special “Neutropenic Risk Hospital Medical Alert Card,” which they presented upon CC ED registration with fever. The pathway formally recognized “fever with a history of cancer” to be a distinct chief complaint and FN to be categorized as Emergency Severity Index level 2 (equivalent to stroke or myocardial infarction) for immediate triage and care. ED-specific electronic FN order sets were created to facilitate antibiotic, laboratory, and blood culture ordering, with antibiotics administered prior to return of neutrophil count. The primary goal of the FNP is administration of empiric broad-spectrum antibiotics within 120 minutes of ED presentation, per Infectious Diseases Society of America guidelines; and the primary outcomes measured were time intervals related to it, e.g. time to blood draw, physician assessment, and antibiotic order/administration. Group comparisons were made using the chi-square, Kruskal-Wallis, Wilcoxon rank-sum, and log rank tests, as appropriate. All reported times were from ED registration.

In total, 137 consecutive FN episodes in 115 pts with hematologic malignancies occurred during the 12 month study period, 63 episodes in 44 pts in the historical cohort. All pts were triaged and treated using the ED FNP, but use of the specific FN order set was variable: episodes were thus classified as treated per the order set (n=53) or not (n=84 – pts still received antibiotics, but not necessarily per the order set). Overall 60% of pts (n=89) were male and the median age at the time of first ED encounter was 59 years (range 20-88). Cancers were: non-Hodgkin lymphoma (38%), acute myeloid leukemia (21%), other leukemias (15%), and myelodysplastic syndromes (8%). Compared to historical pts, FNP study pts had a higher median ANC (2.0 vs. 0.2, p<0.0001), were less likely to be on growth factors (26% vs. 41%, p=0.06) and more likely to have received prophylactic antibiotics (55% vs. 35%, p=0.006). For the outcome of interest, FNP study pts had significantly shorter time to having blood drawn (median 38.5 vs. 70 minutes, p<0.0001), seeing a doctor (median 44 vs. 71 minutes, p=0.0002) and to receiving antibiotics (median 79 vs. 228 minutes, p<0.0001). Time to admission was also shorter for FNP study pts (4.2 vs. 6.0 hours, p<0.0001), though study pts were less likely to be admitted than historical controls (83% vs. 97%, p=0.005). For FNP pts admitted to the hospital, there was a non-significant decrease in length of stay (median 3.8 vs 4.6 days, p=0.28), ICU admission (7% versus 11%, p=0.26), and length of ICU stay (median 1.9 vs 2.3 days, p=0.83) compared to historical controls.

Comparing the two FNP groups treated or not treated per the order set, those treated using the order set had shorter times to antibiotics being ordered (median 28.0 vs. 60.5 minutes, p<0.0001) and administered (median 66.0 vs 92.5 minutes, p=0.002). ED order set pts also had a higher rate of antibiotic use (100% vs 90%, p=0.02). Correct antibiotic use, antibiotic over-use, hospital and ICU admission rates, time to hospital admission, and length of hospital stay were all similar between the two groups (all p>0.28).

The FNP significantly decreased time from ED registration to all set time-points, including time to antibiotics by almost three-fold, compared to historical controls in pts with hematologic malignancies. Rate of hospitalization was significantly lower, and ICU and length of stay numerically lower. The FNP is an effective clinical tool to provide prompt antibiotic administration to FN pts and likely represents a significant mechanism for improved outcomes and cost-savings to patients with hematologic malignancies presenting with FN.

No relevant conflicts of interest to declare.