Incidence and Risk Factors for Central Nervous System Relapse in Adult Patients with Acute Myeloid Leukemia: A Single Center Experience.

Introduction: There is scarce information concerning incidence and risk factors for central nervous system (CNS) relapse in adult patients with acute myeloid leukemia (AML). In acute lymphoblastic leukemia CNS relapse occurs in up to 30% of patients without prophylactic intrathecal chemotherapy (ITC). This has lead to establish its prophylactic use during induction and post-remission phase. Due to the lack of information about incidence of CNS relapse in adult patients with AML, the usefulness of ITC prophylaxis is not clear.

Objectives: Analyze incidence and risk factors for CNS relapse in a large cohort of adult patients with newly diagnosed AML.

Material y methods: Between 1976 y 2005, 747 adult patients (median 54 years, range 16–81) were diagnosed of de novo AML in our institution. All of them received induction with intensive chemotherapy. Prophylactic ITC was not administered, and cerebrospinal fluid was analyzed only if CNS infiltration was suspected. We analyzed the incidence and risk factors for CNS relapse in patients who reached a complete remission. To calculate the Kaplan-Meier estimates of event-free survival (EFS) we considered as an event the first relapse in CNS, censoring patients at the moment of death or at first relapse in a site different than CNS.

Results: 432 patients (58%) obtained complete remission. Between 1976 and 1989 (period 1) 12 of 136 patients (9%) were submitted to autologous or allogeneic stem-cell transplantation (SCT), whereas 129 of 296 (44%) received SCT between 1990 and 2005 (period 2). Overall, 8 of 432 patients (2%) had a CNS relapse, 3 isolated in CNS and 5 in bone marrow plus CNS. Of them, only 1 presented CNS infiltration at diagnosis. In univariate analysis, CNS relapse was associated with high LDH (3% vs 0%, p=0.06), lisozyme >30 (8% vs 1%, p=0.06), FAB M4–M5 (5% vs 1%, p=0.04) and period 1 (5% vs 0.3%, p<0.01). The median follow-up of the cohort was 85 months. CNS relapses occurred at a median of 10 months after complete remission (range, 3 to 84 months). The EFS at 10 years was 95%, and it was lower in patients with elevated LDH (91% vs 100%, p=0.02), FAB M4–M5 (88% vs 97%, p<0.01), leukocytes >10 ×10⁹/L (92% vs 98%, p=0.07), no SCT (92% vs 100%, p<0.01), and period 1 (80% vs 99%, p<0.01). In multivariate analysis, AML M4–M5 remained as the only independent prognostic factor for EFS (HR 6.4, p=0.01). Only 1 of 8 patients with CNS relapse is alive (AML M3, 11 years after a second complete remission). Median survival after CNS relapse was 168 days (range, 16 to 3821 days).

Conclusion: In adults with de novo AML CNS relapse is an infrequent event. Intensification of post-remission therapy, especially with SCT, during the last decades may have contributed to reduce its incidence. Therefore, administration of prophylactic intrathecal chemotherapy should not be recommended, even in patients with high risk of CNS relapse, such as monocytic AML.