Citation: British Journal of Haematology. 2019, 185, 161-62
Author: D'Arienzo P.; Shaw R.J.; Johnstone M.; Salim R.; Turtle L.
Abstract: Febrile neutropaenia is the commonest life-threatening complication of chemotherapy. Haemato-oncology patients represent a unique cohort regarding the severity and duration of immunosuppression. In this study we aim to compare our management of immunocompromised sepsis to local policy and NICE guidelines, and by reviewing microbiological isolates as well as patient outcomes, we aim to assess if our antimicrobial policy is optimal. Data was collected prospectively from 1st September 2018-15th December 2018; due to insufficient numbers, we extended this to a retrospective review from 15th July 2018 - 31st August 2018. All inpatients (N = 149) admitted under a Haemato-Oncology consultant were reviewed for febrile episodes. As per NICE guidance, a pyrexial episode was defined as temperature >=38degreeC. The electronic patient record was used to identify patient characteristics including age, sex, diagnosis, treatment & transplant status. The electronic prescribing system was used to identify antimicrobial data. The Integrated Clinical Environment software was used to identify neutrophil count, resistant organism screens and blood cultures. Statistical analyses were carried out using SPSS 25.0. Forty-one patients (median age 52 yrs, 61% male) experienced 55 pyrexial episodes. The underlying condition was high-grade lymphoma in 12 (29%), AML in 11 (27%), myeloma in 6 (15%), ALL in 4 (10%). Fourteen patients were stem cell transplant recipients (50% autologous, 50% allogeneic). Neutropenia (neutrophil count < 1.0) was present in 51% of episodes; median neutropenia duration was 7 days (range 1 - 365). Most episodes occurred out of hours (76%); 20% occurred in A&E or on outlying wards. National early warning score was recorded in all episodes (median 3, range 1 - 10). Resistant organism screens had been performed within 3 months in 89% of cases. On presentation, peripheral blood cultures were sent in 71%, line blood cultures in 81%. Amongst the 36 patients not already receiving antimicrobials, antibiotics were administrated within 1 hr in 22 (39%) cases (median time to antibiotic administration 75 min, range 9 - 368). Median time to administration of antibiotics was 77 min in the Haematology department vs. 68 in other departments (Mann-Whitney U test, p = 0.44). Similarly, there was no difference in time to administration between episodes occurring during working hours vs. out of hours (Mann-Whitney U test, p = 0.76). Prescribed antibiotics were in line with local policy (piperacillin/tazobactam plus gentamicin or meropenem) in 39 cases (71%); in 17 cases additional antimicrobials were given (teicoplanin in 12, daptomycin in 3, amikacin in 2). Microbiological isolates were grown in 15 episodes, the commonest organisms being E. coli (33%), K. pneumoniae (13%), Enterococcus sp (13%). Outcomes included resolution in 52% of cases, complications in 40% (ongoing bacteraemia, need for line removal or admission to critical care) and death in 8%. Our study demonstrated particular areas of management which need addressing: <40% of patients received antimicrobials within 1 hr and a significant proportion did not have paired peripheral and line blood cultures sent. Antimicrobial choice was variable; many patients had resistant organism screens, but the results were not always clearly reflected in the antimicrobials given. Gram negative bacilli were the most common organisms grown in our cohort; data collection of resistance profiles of these isolates is ongoing.
