Efficacy and safety of prolonged rivaroxaban treatment for high-bleeding risk cancer patients with acute low-risk pulmonary embolism: a post-hoc subgroup analysis of the ONCO PE trial

Abstract Background Cancer patients often have a high-bleeding risk characteristic. The ONCO PE trial revealed that 18-month rivaroxaban treatment for cancer patients with acute low-risk pulmonary embolism (PE) was superior to 6-month rivaroxaban treatment in terms of thrombotic events. However, the efficacy and safety of prolonged rivaroxaban treatment in cancer patients with high-bleeding risk (HBR) remains unknown. Purpose In this post-hoc subgroup analysis of the ONCO PE trial, we assessed the efficacy and safety of 18-month rivaroxaban treatment in HBR cancer patients with acute low-risk PE. Methods Patients were stratified according to their HBR status using two established scoring systems for venous thromboembolism (VTE): the VTE-BLEED and RIETE scores. In the ONCO PE trial, cancer patients with acute low-risk PE of the simplified pulmonary embolism severity index score of 1 were assigned to 18-month or 6-month rivaroxaban treatment. Because all patients were considered HBR by the VTE-BLEED score due to their cancer status, we defined the HBR subgroup as those with a VTE-BLEED score exceeding 4.5, which represented the optimal cut-off on the ROC curve for predicting the primary endpoint. Patients with a RIETE score greater than 4 were classified as HBR. The primary endpoint was recurrent VTE at 18 months. The major secondary endpoint was major bleeding at 18 months. Results Of the 178 participants, 106 patients were classified as the HBR by VTE-BLEED score, whereas 22 patients were classified as HBR by RIETE score. The mean age was 65.7±10.4 years and 47% were male. The mean VTE-BLEED score was 4.6±1.3 and the mean RIETE score was 3.1±2.3. In the HBR subgroup of the VTE-BLEED score, the primary endpoint rate of 18-month rivaroxaban treatment was significantly lower than that of 6-month rivaroxaban treatment (3.5% vs. 16.3%, P=0.02; odds ratio OR, 0.19; 95% confidence interval CI, 0.03–0.79), and a similar trend was observed in the non-HBR group (9.4% vs. 22.5%, P=0.13; OR, 0.36; 95% CI, 0.07–1.32). No significant differences in major bleeding rates were observed between the two treatments in either the HBR group (8.8% vs. 6.1%, P = 0.60; OR, 1.47; 95% CI, 0.34–7.50) or non-HBR group (6.3% vs. 5.0%, P = 0.82; OR, 1.13; 95% CI, 0.15–11.1). In the HBR subgroup of RIETE score, there was no significant difference in the primary endpoint (0.0% vs. 11.1%, P=0.17); however, in the non-HBR subgroup, the primary endpoint rate of 18-month rivaroxaban treatment was lower than that of 6-month rivaroxaban treatment (6.6% vs. 20.0%, P=0.01; OR, 0.28; 95% CI, 0.09–0.76). The incidence of major bleeding did not differ both in the HBR group (15.4% vs. 11.1%, P = 0.63; OR, 1.45; 95% CI, 0.12–34.5) and non-HBR group (6.6% vs. 5.0%, P = 0.67; OR, 1.34; 95% CI, 0.33–5.59). Conclusions Prolonged rivaroxaban treatment for cancer patients with acute low-risk PE did not show a signal of an increased risk of bleeding even in HBR patients.Primary endpoint Major secondary endpoint