Avatrombopag, an Alternate Treatment Option to Reduce Platelet Transfusions in Patients with Thrombocytopenia and Chronic Liver Disease-Integrated Analyses of 2 Phase 3 Studies

Abstract

Aims: Thrombocytopenia complicates the management of patients with chronic liver disease (CLD) undergoing invasive procedures with a bleeding risk. Until recently, prophylactic platelet transfusion was the only treatment option, but has significant safety and efficacy limitations. Phase 3 data demonstrated the superiority of avatrombopag to placebo in reducing platelet transfusions for bleeding, supporting its recent approval.

Methods: Integrated analyses of pooled data (N = 435) from two randomized, double-blind, placebo-controlled, phase 3 studies assessed the original efficacy endpoints. Additional analyses included subgroup analyses, alternate Baseline platelet count definitions, and another efficacy endpoint.

Results: Avatrombopag was superior to placebo in increasing patients not requiring a platelet transfusion or rescue procedure, those achieving a platelet count ≥50 × 109/L on Procedure Day, and the change in platelet counts from Baseline. The avatrombopag treatment effect was consistently positive across clinically important disease and Baseline clinical characteristic subgroups, and using alternate Baseline platelet count cohort definitions. Similarly, more avatrombopag-treated patients achieved ≥50 × 109/L platelets with an increase of ≥20 × 109/L from Baseline. The incidence and severity of adverse events were similar between avatrombopag and placebo. Further, safety data demonstrated a low risk for thromboembolic events and hepatotoxicity.

Conclusion: These integrated analyses confirmed the superiority of avatrombopag to placebo in reducing platelet transfusions or rescue procedures for bleeding in patients with thrombocytopenia and CLD scheduled to undergo an invasive procedure, and its tolerable safety profile. Importantly, these data warrant reconsideration of clinical decision making regarding the need to treat thrombocytopenia in patients with CLD. This trial was registered with NCT01972529 and NCT01976104.

Conflict of interest statement

Fred Poordad, Norah A. Terrault, Naim Alkhouri, and Mordechai Rabinovitz have had consulting agreements in the past with Dova Pharmaceuticals Inc. Wei Tian and Lee F. Allen are employed by Dova Pharmaceuticals Inc.

Copyright © 2020 Fred Poordad et al.

Figures

Figure 1

Patients' disposition and primary reason for discontinuation (pooled data from ADAPT-1 and ADAPT-2, all randomized patients). AE = adverse event; EC = entry criteria; LTFU = lost to follow-up; SC = patients' choice; WC = withdrawn consent.

Figure 2

Forest plot of proportion of patients not requiring a platelet transfusion or any rescue procedure, combined Baseline platelet count cohorts—pooled data from ADAPT-1 and ADAPT-2 (full analysis set). CI = confidence interval, MELD = model for end-stage liver disease. aTreatment difference = proportion of responders for avatrombopag−proportion of responders for placebo; 95% confidence interval is calculated based on normal approximation.

Figure 3

(a) Proportion of patients who did not require a platelet transfusion or rescue for bleeding 7 days post-procedure by individual Baseline platelet count subgroup—pooled data from ADAPT-1 and ADAPT-2 (full analysis set). aResponders were defined as patients who did not require a platelet transfusion or any rescue procedure for bleeding after randomization and up to 7 days following a scheduled procedure. Note: P-values are based on Fisher's exact test. (b) Proportion of patients achieving the target platelet count of ≥50 × 109/L on Procedure Day by individual Baseline platelet count subgroup—pooled data from ADAPT-1 and ADAPT-2. aResponders were defined as proportion of patients achieving the target platelet count of ≥50 × 109/L on Procedure Day. Note: P-values are based on Fisher's exact test. (c) Mean change in platelet count from Baseline to Procedure Day by individual Baseline platelet count subgroup—pooled data from ADAPT-1 and ADAPT-2. Note: P-values are based on Wilcoxon rank-sum test.

Figure 4

(a) Alternate Baseline Platelet Count Cohorts—Proportion of patients not requiring a platelet transfusion or rescue procedure for bleeding—Pooled data from ADAPT-1 and ADAPT-2 (Full analysis set). aResponders were defined as patients not requiring a platelet transfusion or any rescue procedure for bleeding after randomization and up to 7 days following a scheduled procedure. bAlternate Low Baseline Platelet Count Cohort included patients with a platelet count of <35 × 109/L. cAlternate High Baseline Platelet Count Cohort included patients with a platelet count of 35 to ≤50 × 109/L. dCombined Baseline Platelet Count Cohort included patients with a platelet count of <50 × 109/L. ∆ value represents the difference of proportion versus placebo = proportion of Responders for avatrombopag minus the proportion of Responders for placebo. P-value is based on Cochran–Mantel–Haenszel Test stratified by the risk of bleeding associated with the scheduled procedure. CI = confidence interval. (b) Alternate secondary efficacy endpoint analysis—Summary of proportion of patients that achieved platelet count ≥50 × 109/L and an increase of ≥20 × 109/L on Procedure Day—Pooled data from ADAPT-1 and ADAPT-2 (Full Analysis Set). aLow Baseline Platelet Count Cohort included patients with a platelet count of <40 × 109/L. bHigh Baseline Platelet Count Cohort included patients with a platelet count of ≥40 to <50 × 109/L. cCombined Baseline Platelet Count Cohort included patients with a platelet count of <50 × 109/L. ∆ value represents the difference of proportion versus placebo = proportion for avatrombopag minus the proportion for placebo. 95% CI is calculated based on normal approximation. P-value is based on Cochran–Mantel–Haenszel test stratified by the risk of bleeding associated with the scheduled procedure. CI = confidence interval.