Exchange Transfusion and Leukapheresis in Pediatric AML Patients with High Risk of Early Death for Bleeding and Leukostasis

Background. The risk of early death (ED) due to bleeding and/or leukostasis is high in AML patients with initial hyperleukocytosis (white blood cell count [WBC] > 100 000/µl) and highest in those with hyperleukocytosis and mono- or myelomonocytic leukemia (FAB M4/M5). (Creutzig, et al 1987) Within the AML-BFM studies, emergency strategies for children with AML and high risk for bleeding and leukostasis included exchange transfusion (ET) or leukapheresis (LPh). In order to determine whether these interventions reduced the rate of ED, 1251 AML-BFM patients from the trials AML-BFM 98 and 04 were analyzed. Risk factors for ED and interventions performed were verified focusing on patients with hyperleukocytosis.

Patients. 238 of 1251 (19%) AML-patients <18 years of age (FAB M3 excluded) presented with hyperleukocytosis. Twenty-three out of 1251 (1.8%) patients died by bleeding and leukostasis within 15 days from diagnosis, 18 (78%) of these 23 ED patients had hyperleukocytosis. Seventy-two patients received ET and 17 LPh (including 14 patients with WBC counts < 100 000/µl). 149 patients with hyperleukocytosis did not receive ET/LPh. The median age of patients receiving ET was significantly lower compared to those with LPh (3.5 years vs 12.6 years, p = 0.015). WBC counts were similar in both treatment groups (ET median 224 000/µl vs LPh 218 000/µl, p = 0.20).

Results. The percentage of ED by bleeding/leukostasis increased with higher WBC counts and was highest in 105 patients with WBC > 200 000/µl (14.3%). The ED rates were even higher in patients with FAB M4/M5 and hyperleukocytosis >200 000/µl compared to others with WBC >200 000/µl (M4/M5 13/65 [20%] vs. others 2/40 [5%], p=0.04). Patients with WBC counts >200 000/µl did slightly better with ET or LPh compared to those without ET/LPh (ED rate 7.5 % vs 21.2 %, p=0.055). Patients with WBC between 100 000/µl and 200 000/µl received ET/LPh less frequently compared to those with WBC >200 000/µl (22/133 [17%] vs. 53/105 [50%]). ET/LPh was mainly given in case of clinical symptoms of bleeding or leukostasis or coagulopathies or insufficient reduction (or increase) of WBC counts despite low dose chemotherapy. 15/80 (19%) patients with FAB M4/5 and WBC 100 000 -200 000/µl received ET/LPh and none of these patients died early. ET/LPh was even given in 14 patients with WBC <100 000/µl because of clinical symptoms of bleeding or leukostasis or coagulopathies or rising WBC counts. In multivariate analysis WBC >200 000/µl was the strongest independent risk factor for ED (hazard ratio =15.0, 95% confidence interval 4.9-46.3, p(chi)<0.0001). FAB M4/M5 subtypes, general condition grade 4 and initial bleeding were also significant risk factors. Application of ET/LPh seems to have a non-significant benefit for a reduced ED rate by bleeding/leukostasis (p=0.13). The assessment of the general clinical condition of the patients plays a major role for the decision for ET/LPh. However, this possibly selective cofactor could not be included completely in our calculation because of lack of standardized assessments and documentation of clinical reasons for decision making in particular in patients without ET/LPh. There was no difference in ED rates between ET and LPh. (2/17 vs 5/72, p =0.61). Compared to LPh, ET can be given without time delay. ET is easier to perform especially in young children who need smaller exchange volumes, as well as in adolescents where it can be applied also as partial ET. ET also corrects metabolic disturbances and avoids deterioration of coagulation.

Conclusion. Our data confirm the high risk of bleeding/leukostasis in patients with hyperleukocytosis. Although we could only disclose a trend for a clinical benefit of ET/LPh in this retrospective analysis - probably due to some negative selection bias in patients with the intervention – we strongly advocate ET/LPh in AML patients with WBC >200 000/µl, and in particular in those with FAB M4/M5 subtypes or with clinical symptoms of bleeding or leukostasis or coagulopathies even with lower WBC (100 000/µl - 200 000/µl).

Creutzig, U. et al. (1987) Early deaths due to hemorrhage and leukostasis in childhood acute myelogenous leukemia: Associations with hyperleukocytosis and acute monocytic leukemia. Cancer, 60, 3071-3079.