Performance Status, Cancer-Associated Thrombosis, and Bleeding: A Post-Hoc Analysis of the Avert Trial

Background: The Eastern Cooperative Oncology Group Performance Status (ECOG-PS) is a reliable prognostic determinant for cancer-related outcomes, and a widely used tool to guide cancer management. However, the association between ECOG-PS and clinical outcomes in the setting of cancer-associated venous thromboembolism (VTE) prophylaxis is unexplored to date.

Methods: To evaluate this association, we leveraged data from the AVERT trial: a randomized, placebo-controlled, double-blind clinical trial assessing the efficacy and safety of apixaban 2.5 mg twice daily for primary VTE prevention in intermediate-to-high risk ambulatory cancer patients (Khorana score ≥2) initiating chemotherapy. The primary efficacy outcome was VTE, comprising deep vein thrombosis and pulmonary embolism, occurring from randomization through end of follow-up (month 7). The primary safety outcome was clinically relevant bleeding, comprising major and clinically relevant non-major bleeding (CRNMB), occurring during the treatment period. Secondary outcomes included individual components of the primary outcomes and all-cause mortality. Hazard ratios (HRs) with corresponding 95% confidence intervals (CIs) for the study outcomes in patients with ECOG-PS score of 0-1 vs ≥2 were estimated using a Cox regression model controlling for age, sex, body mass index (BMI), creatinine clearance (CrCl) and study treatment.

Results: Overall, 428 (76.0%) of 563 trial participants had an available baseline ECOG-PS score and were included in the modified intention-to-treat analysis: 368 (86.0%) and 60 (14.0%) had an ECOG-PS score of 0-1 and ≥2, respectively. Compared with subjects with ECOG-PS score of 0-1, those with ECOG-PS score ≥2 had significantly lower BMI and CrCl, and older age. The two ECOG-PS groups were balanced with respect to other major baseline characteristics. Overall, VTE occurred in 10% (N=6) and 6.3% (N=23) of subjects with ECOG-PS score ≥2 and 0-1, respectively (HR, 1.63; 95% CI, 0.62-4.33; p = 0.32), with similar findings observed for deep vein thrombosis (HR, 1.63; 95% CI, 0.40-6.67; p = 0.50) and pulmonary embolism (HR, 1.90; 95% CI, 0.74-4.87; p = 0.18). Clinically relevant bleeding occurred in 10% (N=6) and 5.7% (N=21) of subjects with ECOG-PS score ≥2 and 0-1, respectively (HR, 2.43; 95% CI, 2.35-4.33; p < 0.0001). Similarly, the risks for major bleeding (HR, 8.51; 95% CI, 1.16-62.52; p = 0.035), CRNMB (HR, 1.74; 95% CI, 1.68-1.80; p < 0.0001), and all-cause mortality (HR, 11.10; 95% CI, 1.52-80.90; p = 0.018) were significantly higher among patients with ECOG-PS score ≥2 vs 0-1.

Conclusions: Worse performance status was independently associated with substantially increased risks for bleeding complications and mortality among ambulatory patients with active cancer enrolled in the AVERT trial. Our findings identify ECOG-PS as a multidimensional and clinically relevant factor that might be considered for the development of more efficient risk stratification models in the field of cancer-associated VTE. If adequately validated, ECOG-PS may potentially serve as a novel convenient tool aiding clinical decision making in subjects with or at risk for cancer-associated VTE.