The Use of Aspirin and Clinical Implications in Essential Thrombocythemia

The difference in event-free survival of 5% in the Primary Thrombocythaemia 1 (PT1) study[1] is entirely explained and attributable to transient ischemic attacks (TIAs) and transient nasal and gastrointestinal (GI) bleedings (Table [1]). Given that low-dose aspirin is highly effective in the prevention of platelet microvascular disturbances, including atypical and typical TIAs,[2] [3] the excess of a few TIAs (3%) in the low-dose aspirin plus anagrelide arm of the PT1 study remains to be elucidated. In the study, a slight excess (3%) of transient GI and nasal bleeds was seen in the anagrelide plus aspirin arm compared with the hydroxycarbamide (hydroxyurea) plus aspirin arm (Table [1]). There were no differences in major thrombotic events and bleeding complications in the hydroxycarbamide plus aspirin versus the anagrelide plus aspirin arm (Table [1]). At the doses recommended for use in the treatment of essential thrombocythemia (ET), anagrelide may theoretically potentiate the effects of other medications that inhibit or modify platelet function (e.g., aspirin). However, during clinical development, no such effects were observed when used with aspirin.[4] Given that bleeding is not normally seen in ET patients treated solely with anagrelide, this complication is very likely related to aspirin, which is known to be linked to a risk of hemorrhage in ET patients whose platelet levels are excessively high. Spontaneous hemorrhages can be an intrinsic feature of ET and usually occur at platelet counts > 1000 × 10⁹/L due to an acquired von Willebrand factor (vWF) deficiency at increasing platelet levels.[5] [6] GI bleeding can be an outcome of long-term aspirin use alone, even with low doses of aspirin (75 to 325 mg/d).[7] Bleeding may occur at platelet levels of 1000 to 2000 × 10⁹/L, and low-dose aspirin aggravates or may elicit symptoms of bleeding.[5] [6] In addition, it is not indicated in the PT1 study whether patients with GI bleedings may have had a previous history of this condition, and may otherwise have been eligible for treatment with a proton pump inhibitor, for example.

Table 1 Direct Comparison of Hydroxyurea and Anagrelide in the PT1 Study: A Matter of Interpretation Study Arm Hydroxyurea Anagrelide p Number of patients 404 405 Low- + intermediate- + high-risk ET age 62 yr age 62 yr Study follow-up 2.5 yr 2.5 yr Hemoglobin (mmol/L) 8.7 8.8 Platelets (× 109/L) 1011/853 1004/839 Discontinuation of drug 74 (18%) 147 (36%) 0.00001 Side effects 46 109* No control platelet count 15 19 True ET versus false ET CIMF 1/3 versus 2/3 1/3 versus 2/3 Myelofibrosis 5 (1.2%) 16 (4%) 0.01 Leukemia within 2.5 yr 6 4 All major thrombosis: 30 (7.4%) 29 (7.2%) NS Arterial thrombosis 16 26 Venous thrombosis 14 3 TIA while receiving aspirin: not ET 1 14† Major bleedings 4 (1%) 5 (1.2%) NS Transient Gl and nose bleeds 4 (1%) 17 (4.2%) 0.01 Skin side effects 40 (10%) 27 (6.6%) 0.04 There is no ideal drug for high-risk ET; hydroxyurea and anagrelide differ-do not overtreat. PT1, Primary thrombocythaemia 1 study; ET, essential thrombocythemia; CIMF, chronic idiopathic myelofibrosis; NS, not significant; TIA, transient ischemic attack; GI, gastrointestinal. Side effects of anagrelide are very well known. Cardiac rhythm disturbances.

ET patients with platelet counts between 800 and 2000 × 10⁹/L are at high risk of minor and/or major thrombotic complications, and frequently present with the paradoxical occurrence of microvascular thrombosis complications and bleeding manifestations.[8] [9] [10] Aspirin will prevent the serious thrombotic complications, but will elicit or aggravate the bleeding tendency. In this situation, discontinuation of aspirin will be followed by a high risk of serious thrombotic complications and, therefore, it is strongly recommended not to discontinue aspirin but rather to lower platelet counts from above to below 1000 × 10⁹/L or to normal with the administration of hydroxyurea or anagrelide.[5] [6] [7] [8] It is widely agreed that aspirin should be continued in symptomatic ET patients once platelet counts have been lowered. In the PT1 study, both hydroxycarbamide and anagrelide, when used in combination with aspirin, resulted in occurrences of nonfatal hemorrhagic events.[1] Transient nasal and GI bleedings in ET are not to be considered as major problems, because they are easy to treat provided that the underlying platelet and or vWF defects are properly analyzed at time of bleeding.[7] Platelet counts at the time of thrombohemorrhagic events in the PT1 study were not presented in the publication, and the ristocetin cofactor status and bleeding times, which provide useful measures of when to withhold and when to introduce aspirin, also were not reported.

The possibility of a subclinical or overt acquired von Willebrand syndrome (AvWS, platelet type 2B) should also have been considered before initiation of aspirin therapy.[6] [7] ET-associated AvWS usually occurs in the presence of platelet counts > 1000 × 10⁹/L, but, as mentioned, the use of aspirin is known to cause GI bleeding, in particular at platelet counts > 1000 × 10⁹/L.[7] [8] [9] The mechanism is believed to involve increased catabolism of high molecular weight vWF multimers, secondary to increased platelet-vWF protein interaction.[5] [9] Symptomatic patients receiving cytoreductive therapy, which reduces the platelet count to < 900 × 10⁹/L, usually correct both clinical and laboratory abnormalities of the AvWS. In asymptomatic patients, prophylactic cytoreduction is advised only in the absence of a clinically relevant reduction in vWF function (ristocetin cofactor activity and collagen-binding activity [< 60%]) together with a normal bleeding times and Platelet Function Analyser closure time (PFA-100; Dade-Behring, Marburg, Germany). ET patients with platelet counts between 600 and 1000 × 10⁹/L may show some reduction of large vWF multimers with prolonged PFA-100 closure times consistent with a latent bleeding tendency, but generally do not suffer from overt bleedings when receiving low-dose aspirin. It is appropriate to use aspirin therapy in the presence of subclinical ET-associated AvWS as long as the ristocetin cofactor activity and collagen binding activity remains > 60% in the presence of a normal bleeding time and PFA-100 closure times < 300 seconds.[7] It is known that aspirin may elicit upper abdominal complaints and gastritis due to a local irritant effect of aspirin on the intestinal mucosa. This may be a reason to discontinue aspirin and correct platelet counts to normal.

It is important to elucidate whether the thrombohemorrhagic events coincided with instances when platelet levels were controlled. The platelet levels at the time of the events are not stated, despite the fact that these data could be highly significant with regard to how end points were reached. There is still concern about the higher incidence of skin toxicity in the hydroxyurea arm as compared with the anagrelide arm of the PT1 study (Table [1]). This difference may increase after long-term follow-up, which is of particular importance for ET patients younger than 65 years.

REFERENCES

• 1 Harrison C N, Campbell P J, Buck G et al.. for the United Kingdom Medical Research Council Primary Thrombocythaemia 1 Study. Hydroxyurea compared with anagrelide in high-risk essential thrombocythaemia.  N Engl J Med. 2005;  353 33-45
• 2 Koudstaal P J, Koudstaal A. Neurologic and visual symptoms in essential thrombocythemia: efficacy of low-dose aspirin.  Semin Thromb Hemost. 1997;  23 365-370
• 3 Michiels J J, Berneman Z, Schroyens W, Koudstaal P J, Lindemans J, van Vliet H HDM. Platelet-mediated thrombotic complications in patients with ET: reversal by aspirin, platelet reduction, and not by coumadin.  Blood Cells Mol Dis. 2006;  36 199-205
• 4 Fruchtman S M, Petitt R M, Gilbert H S, Fiddler G, Lyne A. Anagrelide: analysis of long-term efficacy, safety and leukemogenic potential in myeloproliferative disorders.  Leuk Res. 2005;  29 481-491
• 5 Michiels J J. Acquired von Willebrand disease due to increasing platelet count can readily explain the paradox of thrombosis and bleeding in thrombocythemia.  Clin Appl Thromb Hemost. 1999;  5 147-151
• 6 Michiels J J. Platelet-mediated microvascular inflammation and thrombosis in thrombocythaemia vera: a distinct aspirin-responsive arterial thrombophilia, which transforms into a bleeding diathesis at increasing platelet counts.  Pathol Biol (Paris). 2003;  51 167-175
• 7 Serrano P, Lanas A, Arroyo M T, Ferreira I L. Risk of upper gastrointestinal bleeding in patients taking low-dose aspirin for the prevention of cardiovascular diseases.  Aliment Pharmacol Ther. 2002;  16 1945-1953
• 8 Michiels J J, Berneman Z, Schroyens W, Finazzi G, Budde U, van Vliet H HDM. The paradox of platelet activation and impaired function, platelet-von Willebrand factor interactions and the molecular etiology of thrombotic and hemorrhagic manifestations in essential thrombocythemia and polycythemia vera.  Semin Thromb Hemost. 2006;  32 589-604
• 9 Budde U, Dent J A, Berkowitz S D, Ruggeri Z M, Zimmerman T S. Subunit composition of plasma von Willebrand factor in patients with the myeloproliferative syndrome.  Blood. 1986;  68 1213-1217
• 10 Fabris F, Casonato A, Grazia del Ben M, De Marco L, Girolami A. Abnormalities of von Willebrand factor in myeloproliferative disease: a relationship with bleeding diathesis.  Br J Haematol. 1986;  63 75-83