Hemorrhagic and thrombotic complications in patients with myeloproliferative diseases

Objective: To correlate the incidence of hemorrhage and thrombosis to bleeding time (BT) and platelet aggregation in 27 consecutive patients with myeloproliferative diseases (MPD). Design: Retrospective study. Setting: Public tertiary referral center. Patients: Eighteen patients with chronic myelogenous leukemia (CML), 5 with polycytemia vera (PV), 2 with essential thrombocytemia (ET) and 2 with idiopathic myelofibrosis (MF). Duke's BT and epinephrine-induced platelet aggregation were performed on the patients and on 10 healthy individuais. Results: Eleven patients presented symptoms (41 %): 9 with hemorrhage (33%) and 5 with thrombosis (19%). There were less symptomatic patients in the CML group (28%) than in the other MPD (67%), without statistical significance (Fisher, p=0.06). Duke's BT was longe r in symptomatic patients (Mann-Whitney, p<0.05). Platelet aggregation was abnormal in 7 patients (26%) and 71 % of them were symptomatic (Fisher, p = 0.07). Conclusions: The high incidence of bleeding and thrombosis in patients. with MPD was related to prolonged BT, but not to platelet aggregation abnormalities.


INTRODUCTION B
oth hemorrhagic aod thrombotic episodes account for the morbidity and mortality in myeloproliferative diseases (MPD)'.Either phenomena may be related to acquired platelet abnormalities observed in some patients, such as increased platelet production or destruction, altered platelet secretory granular content, changes in glycoprotein concentrations and abnormal response to aggregating agents, especially epinephrine.The incidence of all these platelet abnormalities in patients with MPD has been published in a cc:>mprehensive review of 124 studies by Holme et a1.The aim of this study was to verify the incidence of thrombotic or hemorrhagic complications in patients with MPD, followed up for a 2 year period, and to correlate them to the bleeding time and platelet aggregation induced byepinephrine.
Platelet aggregation pattem was abnormal in 7 patients (26%): 5 out of 18 with CML (28%), 1 out of 2 with ET (50%), and 1 out of2 with MF(50%).AlI patients with PV had normal platelet aggregation (Table 2).No difference was seen between the frequency of abnormal platelet aggregation in the CML group, compared to the other MPD (Fisher test, p=0.57).
Symptomatic patients had a higher incidence of abnormal platelet aggregation patterns (46%) than

Methods
Blood cell counts were performed automatically in a Coulter counter and platelet counts were made in Neubauer chambers, with phase microscopy.
Bleeding time was performed according to Duke's technique, and consisted of a small incision in the earlobe using a sterile lanceto The cut was blotted every 30 seconds with filter paper until bleeding stopped.Results were recorded in minutes 4 •  Platelet aggregation study was performed on platelet rich plasma (PRP) obtained from citrated whole blood, using a Minigator II-Payton Scientific Aggregometer, according to Bohr5.Platelet count on PRP was adjusted to 300 x 10 9 /L.Epinephrine was used as an aggregating agent at a final concentration of2.2 mM.Normal individuaIs presented two aggregation waves and this was considered to be the normal pattern.Patients were classified as having an abnonnal aggregation pattern when there was either a single wave or a complete absence of aggregation.All patients were evaluated on two occasions, at least one month apart, and were considered abnormal only when they presented the abnormal pattern in both analysis.To evaluate the assay conditions, blood from healthy volunteers was collected daily as a controI.
The frequency of symptomatic and asymptomatic patients, as well as the frequencies of normal and abnormal platelet aggregation patterns, were compared using the Fisher exact testo The Mann-Whitney test was used to compare bleeding time in patients and controls.----------------------------the incidence ofhemorrhage and thrombosis Platelet Aggregation were similar, whereas Zucker et aI.found that thrombosis was more common 11 .
The relatively lower incidence of hemorrhage in patients with CML, compared to other MPD, was also observed by other authors 7 , 15.As pointed out by Schafer l , cutaneous and mucous membrane were the most common sites of bleeding.
Duke's BT was longer in symptomatic patients, compared to asymptomatic ones and controls.In fact, the occurrence of prolonged bleeding time was reported by some authors, but in no case was there a positive correlation between BT and the presence of hemorrhage l3 , 15,16.BT is an "in vivo" measuring of primary hemostasis, that depends not onl y on platelet function, but also on complex interactions among platelets, subendothelium, and von Willebrand factor l7 • The prolonged BT in patients who presented hemostatic complications suggests that these individuaIs may have an altered platelet/ endothelium interaction that could favor the occurrence of thrombosis or hemorrhage.
A prolonged Duke's BT correlates with hemostatic defects that are clinicall y relevant, probably better than does the Ivy BT, which (NO= not done, MPO= myeloproliferative diseases, CML= chronic myelogenous may be prolonged even in asymptomatic  The frequency of abnormal platelet aggregation induced by epinephrine ranged from 32% to 76% in the reported series (Table 3).This great variability could be due to the epinephrine concentrations used in the different assays.
In the present study, abnormal platelet aggregation was observed in only 26% ofthe 27 patients, maybe because the patients were considered abnormal only when they present the alteration on two different occasions.
Some authors reported altered platelet aggregation induced by other agents, such as ADP and collagen, although the incidence of low responsiveness to epinephrine has been the most common finding in these studies 13. 20.
71 % of patients with abnormal and 30% with normal platelet aggregation were symptomatic, but no statistical difference was seen.Therefore platelet aggregation pattern seems not to predict the occurrence of bleeding in patients with MPD, as observed by othersI0.15.19.
In conclusion, the authors observed a high incidence (41 %) of bleeding or thrombotic events in 27 patients with MPD.Duke's BT could discriminate the symptomatic patients, while platelet aggregation induced by epinephrine proved to be useless for this purpose.

table 3 .
The authors observed a higher incidence ofhemorrhage (33%) than thrombosis (19%), in 27 patients with MPD.Similar findings have been reported by Adams et aI.6, Walsh et aI.16,and Bush et aI. 7, who studied 21, 16 and 72 patients with Table 2 MPD, respectively.However, Barbui et aI.I2 Results of platelet counts, bleeding time and platelet aggregation in a large series of 101 patients observed that according to myeloproliferative disease

Table 3
Frequency of hemorrhage, thrombosis and abnormal platelet aggregation in patients with MPD (NS=not stated)