Risk Factors for Infectious Complications in Adults with Acute Myeloid Leukemia Undergoing Induction Chemotherapy

Introduction: Febrile neutropenia is a common and potentially fatal consequence of Acute Myeloid Leukemia (AML). We performed a retrospective analysis of 97 adult AML patients (median age 67.6 y (58–78) treated between February 2000 and June 2005. Aim of the study was to evaluate risk factors for infectious complications and to identify underlying causative pathogens during first and second cycle of intensive induction chemotherapy. For statistical evaluation of the results, Students t-test, Fisher’ exact test or Chi square test were performed as appropriate.

Results: Of the 97 episodes 64% occurred in male patients at a median age of 67.6 years (58–78). Patients developed fever not related to cytotoxic therapy or transfusion in 77 (80%) and no fever in 20 (20%) of the episodes. Compared to patients without infectious complications patients with fever had significantly more pretreatment comorbidities (0.80±1.20 vs 1.81±1.41, p=0.0042), a prolonged period of severe neutropenia (<100/μl) (3.3±4.8 vs 13.6±8.4 days, p<0.0001) and neutropenia < 500/μl (5.6±7.2 vs 19.9±12.5 days, p<0.0001). More patients in the fever group had a central venous catheter (15% vs 47%, p=0.02, OR 5.0, CI 1.35–18,38). Clinically, this group presented more grade III/IV symptoms like mucositis (30% vs 66%, p=0.007, OR 4.6, CI 1.58–13.3), nausea (10% vs 66%, p<0.0001, OR 17.65, CI 3.8–81.97), vomiting (5% vs 49%, p=0.0008, OR 18.5, CI 5.38–19.86), diarrhea (15% vs 48%, p=0.01, OR 5.4, CI 1.46–19.9), any other gastroenterologic symptoms (5% vs 31%, p=0.03, OR 8.6, CI 1.1–68.9) and skin affections (40% vs 78%, p=0.002, OR 5.3, CI 1.86–15.0). Consequently, these patients were treated more often with interventionally intended G-CSF (15% vs. 26%, p=0.009, OR 11.7, CI 2.1–65.5) and received a higher number of antibiotic (p<0.0001) and antimycotic regimens (p<0.0001). Altogether twelve patients died within 39.9±31.1 days after the start of induction therapy. Of these, 10 deaths occurred in the fever group) with a clearly increased risk of septic reasons (p=0.04). We furthermore analyzed the effects of concomitant medication on the occurrence and outcome of febrile episodes. No difference was noted for the concomitant treatment as gastric acid suppression (p=0.43), antihistaminics (p=0.45) or corticosteroids (p=0.24) for all patients with fever. In the subgroup with proven bacterial infection the use of steroids was significantly different (p=0.009) and a trend for the use of antihistaminics was found (p=0.06). In the subgroup with an at least probable fungal infection or dead patients the use of steroids was significant (p=0.001 resp. p=0.03). No differences were noted between the first and second cycle of induction or the response to chemotherapy.

Conclusion: Our data support the importance of preexisting features as the presence and number of comorbidities that have a profound impact on the risk of infectious complications during induction chemotherapy in AML while response to chemotherapy does not seem to have an impact in our relatively small cohort. Furthermore, the clinically necessary concomitant medications seem to play a role only in subgroups. As this is a retrospective analysis, confounding influences e.g. as antipyretic medication cannot be excluded. As the outcome of patients was generally not statistically different we recommend treating patients at risk prophylactically with broader spectrum antibiotics and a close clinical and laboratory monitoring.