Results
Ninety febrile neutropenic events occurred in 66 patients. The proportion of men in control group was higher than intervention group (57.8% versus 35.6%, P =0.035). The mean age of all patients was 51.6±15.6 years. Most patients were diagnosed with cytotoxic chemotherapy-induced febrile neutropenia (74.4%) while twenty patients were identified as a febrile neutropenia during period of initial hematologic abnormalities diagnosis (22.2%) and only three patients were diagnosed as a febrile neutropenia from other cause such as vitamin B12 deficiency, severe infection, and zidovudine-induce pancytopenia. The majority of our patients had hematologic malignancy (80%) and 8.9% had solid cancer. The MASCC risk index median score was 20 (interquartile range [IQR] 17-21). The median absolute neutrophil count was 153.9 cells/mm3 (IQR 19-520). Fifty-one percent of patients had history of febrile neutropenia and 55.6 % of patients had been exposed to antibiotics within the past 3 months. Median duration of neutropenia was 7 days. The frequency of infectious diseases consultation was similar in both groups. Baseline characteristics are displayed in Table 1.
The major causative organisms were Gram negative bacteria (43.3%), followed by Gram positive bacteria (13.3%) and fungi (3.3%). The most common causative Gram negative bacteria were Escherichia coli(33.3%), Klebsiella pneumoniae (25.6%), and Pseudomonas aeruginosa (10.3%). Most Gram negative bacteria exhibited multiple-drug resistant (69%). More carbapenem resistance Gram negative bacteria were often found in the pharmacist-driven ASP group compared to control group (8.9% versus 2.2%, P =0.167) while extended spectrum beta-lactamase (ESBL) producing Gram negative bacteria were lower than the control group (6.7% versus 20%, P =0.063). The most common causative Gram positive bacteria were Enterococcispp. (41.7%), Staphylococcus aureus (33.3%), andCorynebacterium spp. (16.7%). Ampicillin-resistantEnterococci spp. was isolated from only in one patient and only one patient had methicillin-resistant Staphylococcus aureus(MRSA). Most of the causative organisms were isolated from blood, urine, or sputum (27.8%, 12.2%, and 8.9%, respectively). The most common sources of infection were primary bacteremia, urinary tract infection, and pneumonia (23.3%, 13.3%, and 10%, respectively). However, the causative organisms were not isolated in nearly half of patients.
Overall, antibiotic appropriateness in the pharmacist-driven ASP group was significantly higher than control group (88.9% versus 51.1%,P <0.001) (Table 2). In providing empirical therapy, the pharmacist-driven ASP group was more appropriate than the control group (97.8% versus 77.8%, P =0.007). Appropriate dosage regimen in the pharmacist-driven ASP group was significantly higher than the control group (97.8% versus 88.7%, P =0.049) as well as appropriate antibiotic coverage (100% versus 91.1%, P =0.041) while appropriate indications was similar in both groups. When providing therapy for definitive infections, the overall appropriateness was greater in the pharmacist-driven ASP group than in the control group (88.9% versus 64.4%, P =0.004) as was the duration of therapy (91.1% versus 75.6%, P =0.039). For therapy if the source of infection was unknown, overall appropriateness in the pharmacist-driven ASP group also significantly greater than the control group (90% versus 54.4%, P =0.011). Furthermore, appropriateness of duration of therapy in the pharmacist-driven group was significantly greater than in control group (93.2% versus 75.6%, P =0.022). However, antibiotic appropriateness in cases of known causative pathogens were not significantly greater than the control group, but there was a trend of improved appropriateness in the pharmacist-driven ASP group. Total antibiotic duration between two groups were similar (P =0.948) (Table 2). The compliance rate to the pharmacist suggestion was 93.8% in pharmacist-driven ASP group. The most common pharmacist interventions were de-escalation (31.3%), adding additional antimicrobials (18.8%), and avoiding serious drug interaction (18.1%).
The 30-day infectious diseases-related mortality and length of stay were similar in both groups (Table 2). In univariate analysis, neither pharmacist-driven ASP nor ID consultation showed a significant impact on 30-day infectious diseases-related mortality (P =0.810 and 0.267, respectively). However, in multivariate analysis, the pharmacist-driven ASP group and infectious diseases consultation significantly reduced the 30-day infectious diseases mortality in patient with cytotoxic chemotherapy-induce febrile neutropenia (OR 0.082, 95%CI: 0.009-0.762,P =0.028). A MASCC score of less than 21 and a history of febrile neutropenia were associated with an increased 30-day infectious diseases mortality as described in Table 3. The utilization rate of target antibiotics in pharmacist-driven ASP group tended to be higher than control group (882 DDD/1000 patient-day versus 705.1 DDD/1000 patient-day). The trend of overall target antibiotic seemed to be higher in both groups (supplementary data, table 2S). The trend of ceftazidime, cefepime, and meropenem utilization was lower in pharmacist-driven ASP group while piperacillin/tazobactam utilization was higher. In the control group, ceftazidime utilization tended to be decreased, but other target antibiotics utilization including cefepime, piperacillin/tazobactam, and meropenem were increased. Overall intravenous antibiotic utilization in the pharmacist-driven ASP group declined while amount of utilization in the control group increased.