The clinical and biological correlates of coagulopathy in children with acute leukemia.

Abstract

We determined the clinical and biological correlates of coagulopathy in a large series of patients with untreated childhood acute leukemia. Twenty-five of 805 children with acute lymphoblastic leukemia (ALL) (3.1%) and 27 of 195 with acute myeloid leukemia (AML) (13.8%) met any two of three requirements for a coagulation disorder: fibrinogen level less than 150 mg/dL; fibrin degradation products greater than 10 micrograms/mL; and prolongation of prothrombin time (PT) greater than 12 seconds, activated partial thromboplastin time (PTT) greater than 45 seconds, or thrombin time (TT) greater than 18 seconds. Patients with ALL complicated by abnormal coagulation were more likely to be boys and to have a T-cell immunophenotype, a high leukocyte count, a mediastinal mass, leukemic involvement of the CNS, hepatosplenomegaly, and L2 blast cell morphology. These features were highly interrelated, with only T-cell markers and CNS involvement achieving independent significance in a multivariate logistic regression model. Hepatomegaly, blast cell morphological subtype (French-American-British [FAB] M3, M4, and M5) and age less than 2 years were each associated with coagulopathy in patients with AML, although age failed to retain importance after logistic regression analysis. The presence of coagulopathy at diagnosis of ALL did not influence the rate of remission induction (P = .55). By contrast, only 14 of 27 children with coagulopathy at diagnosis of AML achieved a complete remission (CR), compared with 129 of 168 other patients who lacked this complication (P = .003). After multivariate analysis, coagulopathy remained independently associated with failure to attain remission in AML (P = .02). Fatal hemorrhagic complications arising in the CNS or lungs accounted for nine of the 13 induction failures in this group. The presence or absence of coagulopathy had no discernible influence on treatment outcome among patients with either ALL or AML who attained a CR. Laboratory evidence of a coagulation defect may be useful in identifying patients with AML who have a greater risk of induction failure and, hence, require close surveillance and intensive replacement therapy to prevent fatal hemorrhagic complications.