Background. Patients with acute myeloid leukemia (AML) undergo high-dose chemotherapy which induce profound and prolonged episodes of neutropenia. Infections during this period are both frequent and severe and are often recognised only by the presence of fever. Corticosteroids (CS) are occasionally administered to these patients before or during chemotherapy. The potent immunosuppressive and antipyretic effect of corticosteroids are well known and could possibly alter fever development in AML patients. The purpose of this study was to evaluate the effect of corticosteroid administration on temperature and fever onset in neutropenic AML patients.
Methods. This was a retrospective analysis of fever profiles in the cohort of patients hospitalized at the infectious disease isolation Unit of CHUV. Specifically, all adult patients diagnosed with AML between 2008 and 2019 were included. A cumulative dose of all systemic corticosteroids was calculated for each stay, and converted into a prednisone equivalent dose using a glucocorticoid potency conversion table. Temperature data, corticosteroid doses and first episodes of bacteremia for each stay were extracted from an electronic database. Patients were then stratified into 3 groups according to their cumulated corticosteroid dose at the time of the first bacteremia: low-dose (0-10 mg), intermediate –dose (10-150 mg) and high-dose (>150 mg). Temperature curves were then created for each group of patients depending on their received cumulative dose of CS.
Results. The study included 615 patients with a total of 863 hospital stays, among which 217 comprised at least one episode of bacteremia. The highest temperatures measured within 5 to 2 days before the first bacteremia were lower among patients receiving a high dose of corticosteroids (median 37.2 °C, IQR 0.7) compared to those receiving an intermediate (median 37.5 °C, IQR 0.6) or a low dose (median 37.4, IQR 0.7, overall P=0.01). The peak temperatures measured within the first day of bacteremia tended to be higher among patients receiving a low dose of corticosteroids (median 38.95 °C, IQR 01.1) compared to those receiving an intermediate (median 38.8 °C, IQR 0.5) or a high dose (median 38.8, IQR 1.1, overall P=0.2).
Conclusion. In patients undergoing intensive chemotherapy, the administration of >150 mg of corticosteroids has a notable effect on basal and peak temperature during infectious episodes. Further research has to be done to determine if corticosteroids significantly delay infection diagnosis and increase risk of severe infections.