Role of serial microbiologic surveillance and clinical evaluation in the management of cancer patients with fever and granulocytopenia

Abstract

Blood and surveillance body cultures are frequently used to monitor potential infectious complications in febrile cancer patients. However, such cultures are costly and their value unclear when patients are already being treated with antibiotics. We prospectively studied all patients who became febrile (oral temperature > 38 °C three times over a 24-hour period or > 38.5 °C once) and granulocytopenic (less than 500 polymorphonuclear leukocytes per mm 3) during a four-year period, and evaluated the value of clinical assessment and microbiologic surveillance in patient management. Two hundred and seventy-one patients who experienced 652 episodes were entered into study. All patients had routine examination, chest radiograph, and surveillance cultures of nose, throat, urine, stool and at least two preantibtotic blood cultures. According to their initial evaluation, patients were classified as having an infectious etiology for their fever or as having unexplained fever, and were treated accordingly with specific or empiric antibiotic therapy. Patients remained on study until their polymorphonuclear leukocytes were greater than 500 per mm 3, during which time they had weekly surveillance cultures and daily blood cultures when febrile. Although 62 percent of the patients who became septic were colonized with the infecting organism(s), these cultures were not helpful in initial diagnosis or antibiotic management. Similarly, routine surveillance cultures were not useful in guiding or altering subsequent therapy in patients already receiving antibiotics. Although initially negative, blood cultures became positive in 11 patients who had received antibiotics for a median of 12 days (range three to 37 days). As a result of these blood cultures obtained at follow-up, modification of the antimicrobial regimen was made in seven patients, and six patients survived the episode. The mortality rate for patients in whom positive blood cultures developed while receiving antibiotics was 54 percent compared with a 15 percent mortality rate for patients prior to starting antibiotics (p = 0.01). Although routine body surveillance cultures are costly, insensitive, and unlikely to influence antimicrobial management, serial blood cultures, even in patients receiving antibiotic therapy, can specifically influence therapy and outcome, and may be justifiable for patients receiving maximum supportive care.