Experience with a Therapeutic Platelet Transfusion Strategy in Acute Myeloid Leukemia: Preliminary Results of a Randomized Multicenter Study After Enrolment of 175 Patients.

We have shown in a monocenter study (ASH 2005, abstract # 428) that routine prophylactic platelet (plt) transfusions are not necessary in patients (pts) with acute myeloid leukemia (AML). Therefore, we started a multicenter randomized study comparing the routine prophylactic (morning trigger: plts ≤ 10/nl, arm P) with a therapeutic platelet transfusion strategy (arm T).

Diagnosis of AML and feasability of intensive chemotherapy within standard protocols was required (M3 only in CR). In arm T transfusions were given only when clinically relevant bleeding (more than petechiae) happened regardless of the morning plt count. For safety reasons prophalyctic plts were transfused in case of invasive aspergillosis, sepsis syndrom and repeated or continous headache. In case of headache pts in arm T received a ct-scan. The experimental transfusion strategy startet at day 8 of induction and at day 1 of consolidation. Primary endpoint was the reduction of plt transfusions by 20%; secondary endpoints were bleeding complications, red blood cell transfusions, side effects, duration of thrombocytopenia and hospital stay.

Until July 2009 175 pts have been randomized. Full data were evaluable for 161 pts (79 arm P and 82 arm T). Median age was 52 years (20-78). We registered 251 induction and 112 consolidation cycles. Mean duration of thrombocytopenia (<20/nl or <10/nl) was 17 and 8 days, respectively. Duration of thrombocytopenia was significantly longer in arm T. Depite the longer duration, the number of platelet transfusions in arm T was significantly reduced, by about 25% (p <0.001). Number of red blood cell transfusions were not different (p = 0.95). Days in hospital and side effects were not different in both groups. Bleeding complications (more than petechial) were significantly increased in arm T showing a 2.3 times higher risk compared to arm P. This is clearly correlated with bleeding being the trigger for plt transfusion in arm T. Majority of hemorrhages were of minor or moderate significance that could all be treated effectively by plt transfusion. Most important was the observation that we registered 5 minor and 2 major cerebral hemorrhages in arm T and none in arm P. Minor cerebral hemorrhages with headache were not registered in arm P since we did not perform ct-scans routinely in arm P in case of headache. Minor cerebral complications were 2 sub-/epidural bleedings after vasovagal collapses of pts at plt counts of 53 and 96/nl, respectively, and 3 spontanuous minor subarachnidal/ intracerebral hemorrhages at plt counts of between 15 and <5/nl. All 5 minor cerebral complications were detected by ct-scan performed due to headache. All pts recovered after plt transfusion without any sequelae.The two other pts died following fatal cerebral bleeding. One pt. had repeated headache since several days. At a plt count of 11/nl fatal bleeding happened. A ct-scan revealed intracerebral mass bleeding. The pt was randomized to the experimental strategy but even following standard transfusion strategy the pt would not have been transfused prophylactically. The only fatal cerebral bleeding that must be clearly attributed to the therapeutic strategy was a female pt with a pulmonary fungal infection that improved under antifungal treatment. She had a morning platelet count of <5/nl. During the following night she woke up with severe headache. Despite timely platelet transfusions she died of progressive mass bleeding. Unfortunately, we could not perform an autopsy to see whether the patient had fungal involvement of the brain. 1-2 % of fatal cerebral bleedings are reported in AML trials despite a prophylactic strategy. Whether these two complications happened by chance in arm T or were the consequence of the experimental strategy cannot be answered by our study. To give proof of this important question one would need about 2000 pts in both arms for statistical reasons. Conclusions A therapeutic platelet transfusion strategy according to our protocol is feasible in AML pts during intensive chemotherapy and can reduce the number of platelet transfusions significantly. An increased risk of fatal cerebral bleeding cannot be excluded by this study. An international collaboration would be necessary to set up a study large enough to determine the final safety.

No relevant conflicts of interest to declare.