Objective: The purpose of this study was to determine whether a systematic strategy of multiple microbiological samples for all adult patients with neutropenia admitted to an intensive care unit could document sepsis despite prior empiric antibiotic therapy.
Patients and methods: All adults admitted to the intensive care unit with chemotherapy-induced neutropenia (WBC < 500/mm3) were included in the study. Microbiological samples were systematically obtained in all patients: 6 blood, 1 stool, 1 urine, 1 nasal swab. Data were analyzed for 54 consecutive patients: 22 men, 32 women, 42 with hematology diseases and 12 with solid tumors, mean age 46.6 +/- 14.4 years. Assessment of disease severity showed: simplified gravity index = 256 +/- 17; organ system failure score on day 1 = 2.42 +/- 1.1; 26 cases of septic shock among 47 cases of sepsis. Mortality was 50%. Within 48 h, bronchioalveolar lavage samples were obtained in 48 patients, and there were 12 spinal taps, 3 laparotomies, 3 skin biopsies and 6 sinus punctures.
Results: A microorganism was isolated in 39 patients (70%): 12 Gram negative, 12 Gram positive, 11 mycoses, 3 cases of herpes. Positive cultures were found for 31 blood, 1 bronchioalveolar lavage, 2 surgical, 5 serous and 22 oral or digestive tract samples. Documented microbial infection did not have prognostic value. Septic shock at admission was predictive of poor outcome.
Discussion: Despite prior antibiotic therapy in 46 patients, with a theoretically active drug in 37 cases, a systematic strategy of multiple deep samples increased the number of documented cases of infection in neutropenic patients. Oral and digestive tract infections predominated despite selective decontamination of the digestive tract.