Signal for a Different Risk of Thrombosis Between Eltrombopag and Romiplostim

Introduction:Eltrombopag and romiplostim are two thrombopoietin receptor agonists (TPO-RAs) marketed for the treatment of immune thrombocytopenia (ITP). Thrombotic events have been reported with both drugs. Whether there exists a differential risk of thrombosis between the two TPO-RAs is unknown. Indeed, there exist pharmacological arguments to sustain a differential risk: i) the two drugs have different structures and different sites of action: romiplostim is an Fc-peptide fusion protein binding to the extracellular TPO receptor domain, while eltrombopag is a small non-peptide molecule that binds to a transmembrane site of the TPO receptor; ii) consequently the intracellular signaling pathways activated by TPO-RAs are not strictly identical in vitro; iii) in clinical practice, a patient who stops a given TPO-RA due to inefficacy, adverse drug reaction (ADR) or important platelet fluctuations can be efficiently rechallenged with the other TPO-RA; and iv) a disproportionality study carried out in the French PharmacoVigilance Database suggested a signal for a different pattern of ADRs between the two drugs.

The aim of this study was to assess whether there exists a signal for a different risk of thrombosis, arterial thrombosis, venous thrombosis, ischemic stroke (IS) and myocardial infarction (MI) between romiplostim and eltrombopag.

Methods:We carried out a disproportionality analysis (case/non-case method) in the World Health Organization PharmacoVigilance Database (VigiBase^®). We selected all the adverse drug reaction (ADR) reports with exposure to eltrombopag or romiplostim between 01 January 2011 and 31 December 2013. Cases were all reports of thrombosis, identified thanks to validated standardized queries and checked by two independent investigators. Non-cases were all other reports. We searched an exposure to eltrombopag (versus romiplostim) in cases and in non-cases, allowing the calculation of Reporting Odds Ratios (RORs). All analyses were adjusted on thrombotic risk factors recorded in the database: age, gender, indication, exposure to oral contraceptive, danazol, polyvalent immunoglobulins, and other pro-thrombotic drugs. We conducted two sensitivity analyses: one restricted to cases and non-cases reported by physicians, and one including the reports of "cerebrovascular accident" among the cases.

Results: Out of 2 733 224 ADR reports colligated in VigiBase^®during the study period, 4714 were included in the study. TPO-RA indication was ITP in 96.6% of the reports. We found 598 cases of thrombosis: 203 arterial thromboses (94 IS and 110 MI) and 324 venous thrombosis. Among the 324 venous thromboses, 131 (40.4%) were pulmonary embolisms, 45 (13.9%) portal thromboses, and 17 (5.2%) cerebral venous thromboses were found. In multivariate analyses, there was an increased risk of thrombosis (ROR=1.46; 95% CI [1.18-1.82]) and venous thrombosis (ROR=1.46; 95% CI [1.10-1.94]) with eltrombopag. There was no statistically significant difference between the two TPO-RAs regarding arterial thrombosis (ROR=1.29; 95% CI [0.89-1.87]), IS (ROR=1.28; 95% CI [0.79-2.05]) and MI (ROR=1.53; 95% CI [0.96-2.43]). Age was independently associated with the risk of thrombosis. Female gender was a risk factor for venous thrombosis (ROR=1.55; 95% CI [1.16-2.06]); in contrast, it was a protective factor for arterial thrombosis (ROR=0.65; 95% CI [0.44-0.94]) and MI (ROR=0.55; 95% CI [0.34-0.90]). Exposure to oral contraceptive was also an independent risk factor for venous thrombosis (ROR=3.45; 95% CI [1.20-9.90]). Exposure to IVIg was associated with MI (ROR=2.13; 95% CI [1.23-3.68]).

When we restricted to ADRs reported by physicians, all ROR measures were slightly increased (overall thrombosis: ROR=2.07; 95% CI [1.60-2.69]; venous thrombosis: ROR=2.02; 95% CI [1.44-2.83]), reaching significance for arterial thrombosis (ROR=1.58; 95% CI [1.01-2.48]) and IS (ROR=2.00; 95% CI [1.16-3.44]). Inclusion of the 45 reports of "cerebrovascular accident" did not influence the results (data not shown).

Conclusions: This study suggests a signal for an increased risk of thrombosis with eltrombopag compared to romiplostim, that must be confirmed by population-based pharmacoepidemiological studies.