BACKGROUND: The hemostatic system is tightly interconnected with cancer. Research has focused predominantly on thrombotic complications, but less is known about bleeding and bleeding risk prediction. Growth differentiation factor (GDF)-15 has previously emerged as a prognostic biomarker for bleeding. OBJECTIVES: The aim of this study was to investigate the association and predictive ability of GDF-15 for bleeding risk in patients with cancer. METHODS: The CAT-BLED (Vienna Cancer, Thrombosis, and Bleeding) study is a prospective, observational cohort study including cancer patients initiating systemic anticancer therapies. Patients were followed for up to 2 years for thrombotic and bleeding events. The primary outcome was major bleeding. GDF-15 was measured at inclusion and at 3 and 6 months of follow-up. RESULTS: A total of 779 patients (48% women, median age 62 years, 15% on therapeutic anticoagulation) were included. During a median follow-up period of 18 months, 79 patients (10.1%) experienced major bleeding (12-month cumulative incidence 8.8%
95% CI: 6.7-10.9). Higher GDF-15 levels were independently associated with increased major bleeding risk (adjusted subdistribution HR per doubling: 1.29
95% CI: 1.04-1.59), and patients with levels greater than the cohort median (1,864 ng/L) had a significantly higher 12-month cumulative incidence (13.1% vs 4.6%
P <
0.001). This association remained robust in follow-up measurements at 3 and 6 months. GDF-15 showed moderate to good discrimination for predicting 6-month major bleeding risk (C statistic = 0.69
95% CI: 0.60-0.77). GDF-15 was not associated with venous thromboembolism but was strongly associated with mortality (adjusted HR: 1.37
95% CI: 1.25-1.50). CONCLUSIONS: GDF-15 levels predict major bleeding risk in cancer patients and are not associated with venous thromboembolism, making GDF-15 a particularly promising biomarker for bleeding risk prediction.