Afdhal NH, Giannini EG, Tayyab G, Mohsin A, Lee J-W, Andriulli A, et al., for the ELEVATE Study Group. Eltrombopag before procedures in patients with cirrhosis and thrombocytopenia. N Engl J Med 2012;367:716–724. (Reprinted with permission.)
Background: Eltrombopag is an oral thrombopoietin-receptor agonist. This study evaluated the efficacy of eltrombopag for increasing platelet counts and reducing the need for platelet transfusions in patients with thrombocytopenia and chronic liver disease who are undergoing an elective invasive procedure.
Methods: We randomly assigned 292 patients with chronic liver disease of diverse causes and platelet counts of less than 50,000 per cubic millimeter to receive eltrombopag, at a dose of 75 mg daily, or placebo for 14 days before a planned elective invasive procedure that was performed within 5 days after the last dose. The primary end point was the avoidance of a platelet transfusion before, during, and up to 7 days after the procedure. A key secondary end point was the occurrence of bleeding (World Health Organization [WHO] grade 2 or higher) during this period. Results: A platelet transfusion was avoided in 104 of 145 patients who received eltrombopag (72%) and in 28 of 147 who received placebo (19%) (P<0.0001). No significant difference between the eltrombopag and placebo groups was observed in bleeding episodes of WHO grade 2 or higher, which were reported in 17% and 23% of patients, respectively. Thrombotic events of the portal venous system were observed in 6 patients who received eltrombopag, as compared with 1 who received placebo, resulting in the early termination of the study. The incidence and severity of other adverse events were similar in the eltrombopag and placebo groups. Conclusions: Eltrombopag reduced the need for platelet transfusions in patients with chronic liver disease who were undergoing elective invasive procedures, but it was associated with an increased incidence of portal-vein thrombosis, as compared with placebo. (Funded by GlaxoSmithKline; ELEVATE ClinicalTrials.gov number, NCT00678587.)
Platelets contribute to hemostasis in three ways. First, they adhere to the subendothelial matrix at the site of vessel wall injury by means of membrane receptors and the adhesive multimeric protein von Willebrand factor (Fig. 1, left). Second, they aggregate one another by means of membrane receptors and von Willebrand factor or fibrinogen (Fig. 1, left). Third, activated platelets help assemble vitamin K-dependent coagulation factors (i.e., tenase and prothrombinase complexes) on their surface by means of negatively charged phospholipids (i.e., phosphatidylserine), thus speeding up thrombin generation (Fig. 1, right) and fibrinogen-to-fibrin conversion.
Patients with chronic liver disease are variably thrombocytopenic and possibly thrombocytopathic and this is considered an index of the bleeding risk, especially during/after invasive procedures. The bleeding time, which was the test of choice to investigate primary hemostasis, has been performed for many years in patients who were about to undergo invasive procedures despite the fact that results of this test were not good predictors of bleeding in these patients. As far as we know the bleeding time test is no longer carried out before invasive procedures, but platelet counts are still assessed and patients with low counts are considered at increased risk of bleeding. Guidelines for liver biopsy suggest platelet transfusion whenever platelet counts are lower than 50 × 109/L and a survey conducted to assess the variation of practice showed that 81% of the respondents would use platelet transfusion before liver biopsy in patients with thrombocytopenia. However, neither the trigger (i.e., the alarming platelet count) nor the target, i.e., the platelet number posttransfusion before invasive procedures, are evidence-based.
Laboratory methods that may help determine the trigger/target platelet count needed before invasive procedures are also lacking. Some years ago Lisman et al, provided in vitro evidence that the adhesion of platelets from patients with cirrhosis to the subendothelial matrix was normal despite the fact that these patients were thrombocytopenic. This is explained by the increased levels of the adhesive protein von Willebrand factor, which is a typical feature of patients with cirrhosis. Recently, we showed in an in vitro study of platelet-rich plasma that the threshold platelet count needed to secure thrombin generation correspondent to the lower limit of the normal reference range amounts to 56 × 109/L. However, the methods employed in the Lisman et al. and in our study, although useful to shed light on the mechanisms of adhesiveness and thrombin generation mediated by platelets in cirrhosis, are not yet applicable to manage patients. Furthermore, in another study we showed that the prophylactic transfusion of one single adult platelet unit (often used regardless of the pretransfusion platelet counts) is barely sufficient to increase platelet counts above 50 × 109/L and that both thrombin generation and thromboelastometry (i.e., a global method that assesses the viscoelastic properties of clotting blood) were barely affected by this transfusion schedule. Accordingly, if a greater platelet count is required, this requires multiple transfusions. Thrombopoietin receptor agonists that are able to increase platelet counts in chronic idiopathic thrombocytopenic purpura may present an alternative to multiple transfusions.
In a recent issue of the New England Journal of Medicine, Afdhal et al. reported interesting results on a randomized/controlled clinical trial of one such oral agent (eltrombopag) for its ability to spare platelet transfusion in patients with cirrhosis undergoing an elective invasive procedure. The study was timely, as platelet transfusions have some limitations (short duration of efficacy, risk of transfusion reactions, and development of antiplatelet antibodies). Patients with platelet counts equal or lower than 50 × 109/L were randomized to receive eltrombopag 75 mg once daily or placebo. A platelet transfusion was avoided in 72% of the patients who received eltrombopag and in 19% of those who received placebo. No significant difference between the eltrombopag (17% of patients) and placebo (23% of patients) groups was observed in bleeding episodes of World Health Organization (WHO) grade 2 or higher. However, thrombotic events of the portal venous system were observed in six patients who received eltrombopag, as compared with one who received placebo. Because of this, the investigators opted for an early termination of the study. We would like to comment on the following aspects.
First, thrombosis as shown by the post-hoc analysis reported by Afdhal et al. occurred more frequently in patients with platelet counts higher than 200 × 109/L. One may argue that such a relatively high platelet count is not needed in these patients to prevent bleeding. Therefore, the study could have aimed at reaching a lower platelet count. Perhaps a lower than 75 mg dose of eltrombopag able to achieve only a moderate increase of the platelet count could be associated with fewer thrombotic events. According to a previous phase I study on eltrombopag in healthy subjects, a ratio of the platelet count corresponding to approximately 1.3 and 1.5 times the baseline value was achieved at an eltrombopag dose of 30 or 50 mg, respectively, with little gain at 75 mg. Whether this is valid also for patients with chronic liver disease is not known, but would deserve attention.
Second, although this was only a secondary endpoint of the study, the occurrence of bleeding during/after invasive procedures was not different in the eltrombopag as opposed to the placebo group. This consideration points to the real need of correcting thrombocytopenia in patients who are moderately thrombocytopenic. If one considers that (albeit in vitro) primary hemostasis (i.e., platelet-vessel wall interaction) in chronic liver disease is apparently rebalanced, owing to the relatively high increase of von Willebrand factor and that thrombin generation (albeit in vitro) is near normal when platelet counts are around 50 × 109/L, the logical consequence should be that correcting thrombocytopenia before invasive procedures would be required only in severe thrombocytopenia. However, how severe should the thrombocytopenia be before being worried about clinical bleeding should be determined with appropriate clinical trials, which (we are afraid) will never be carried out because of their complexity. The same considerations apply to the question of whether or not one has to correct hypocoagulability (based on abnormal prothrombin time, international normalized ratio [PT-INR]) in cirrhosis before invasive procedures. The rebalanced coagulation[12, 13] due to the concomitant reduction of procoagulant and anticoagulant factors, which is a typical feature of patients with stable cirrhosis, would support the concept that infusion of plasma or coagulation factor concentrates are not useful and should not be used indiscriminately or based solely on the prolongation of the PT-INR. In this respect, the current literature provides strong evidence that, at variance with platelets, there are many randomized/controlled clinical trials that failed to show significant benefit of recombinant activated factor VII (one of the most potent procoagulant agents) to stop esophageal variceal bleeding or to affect intraoperative blood loss in patients undergoing liver transplantation.
Third, the reason(s) why the thrombotic risk was higher in patients taking eltrombopag as compared to those taking placebo is unclear. The authors considered a number of interesting hypotheses. Among them the most plausible includes the fact that some of the patients may have had a subclinical partial portal vein thrombosis at the study entry, which was undetected because Doppler ultrasonography of the abdomen was not performed at enrollment. Perhaps the sustained platelet increase following treatment with eltrombopag, risk factors such as an imbalance of coagulation, portal hypertension and reduced blood flow, local inflammation or endothelial injury could have acted in combination to exacerbate subclinical portal vein thrombosis in these patients. Another interesting hypothesis could be that platelets in cirrhosis are overactivated, as shown by the increased urinary excretion of markers of in vivo platelet activation observed in these patients. It is possible that the relatively rapid increase of the number of overactivated platelets may have acted as a trigger for thrombosis.
In conclusion, the study of Afdhal et al. shows that the strategy of using eltrombopag in patients with cirrhosis undergoing elective invasive procedures is effective in increasing the platelet count and thus avoiding platelet transfusion, but carries the risk of increasing the rate of thrombotic events. On the other hand, the benefit of increasing platelet counts in this population in order to prevent hemorrhagic events has not yet been established. Hence, the benefit of improving hemostasis at the expenses of increasing thrombotic risk should be carefully evaluated in individual patients.
Armando Tripodi, Ph.D.
Massimo Primignani, M.D.
Università degli Studi di Milano and
IRCCS Cà Granda Maggiore Hospital Foundation