Bleeding Events in Patients with Cancer: Incidence, Risk Factors, and Impact on Prognosis in a Prospective Cohort Study

Oncology Hematology Medicine
cancer bleeding events risk factors prognosis prospective study anticoagulation mortality

The Incidence, Risk Factors, and Prognostic Impact of Bleeding Risk in Cancer Patients: A Review of the Vienna Cancer, Thrombosis, and Bleeding Study

Academic Background and Research Motivation

Cancer patients frequently experience hemostatic dysregulation, yet existing research has primarily focused on cancer-associated venous thromboembolism (VTE). Data on bleeding events in cancer patients remain sparse, despite their significant clinical impact, including an increased mortality risk. Understanding bleeding risks and clinical patterns is critical for informed clinical decision-making. Additionally, anticoagulant therapies, integral for preventing and treating VTE, increase bleeding risk, further necessitating comprehensive evaluations of both thrombotic and bleeding risks. Existing risk assessment tools, primarily developed for non-cancer populations, are not accurate for cancer patients.

To address this gap, a prospective observational study was conducted within the Vienna Cancer, Thrombosis, and Bleeding Study (Cat-Bled Study), aiming to systematically analyze the patterns, incidence, risk factors, and prognostic impact of bleeding events in cancer patients.

Source and Author Information

This study, authored by Cornelia Englisch, Florian Moik, and colleagues, was conducted at the Medical University of Vienna’s departments of Hematology, Hemostasis, and Oncology in Austria. It was published in the November 28, 2024, issue of Blood (Volume 144, Issue 22).

Research Design and Methods

Study Outline

This study was a single-center prospective cohort study based on data from the Cat-Bled Study, spanning July 2019 to December 2022. The study population consisted of cancer patients initiating systemic anticancer therapies, including newly diagnosed and recurrent or progressing cases, with a total enrollment of 791 patients.

Data were collected using standardized interviews, questionnaires, and electronic medical record reviews, with a maximum follow-up time of two years. The primary focus was on first clinically relevant bleeding (CRB), encompassing major bleeding (MB) and clinically relevant non-major bleeding (CRNMB). All events were classified using the International Society on Thrombosis and Hemostasis (ISTH) criteria and adjudicated by an independent expert committee.

Statistical Analysis

Competing risk analyses were performed to calculate cumulative bleeding incidence, and Fine-Gray proportional subdistribution hazard models were applied to identify risk factors. Cox multistate modeling assessed the impact of CRB events on mortality risk.

Key Findings

Incidence and Patient Characteristics

The study included 791 patients, with a median age of 63 years, 48% female. Of these, 65.5% had stage IV cancer, with the most common cancer types being lung cancer (23.6%), head and neck cancer (11.1%), and pancreatic cancer (10.4%). A total of 120 patients (15.2%) received therapeutic anticoagulation, while 124 (15.7%) were on antiplatelet therapy.

The cumulative CRB incidence for the entire cohort was 9.8% at 6 months, 16.6% at 12 months, and 23.9% at 24 months. For MB, the 12-month cumulative incidence was 9.1%, and for CRNMB, it was 10.0%. In patients not receiving anticoagulation, the 12-month cumulative CRB incidence was 14.4%, higher than in those on therapeutic anticoagulation (11.6%). Patients with head and neck cancers had the highest CRB risk.

Clinical Characteristics of Bleeding Events

Among 194 CRB events, 30.0% were tumor-related (i.e., the bleeding site coincided with the tumor site), 46% were gastrointestinal, 11.5% were urogenital, and 5.0% were intracranial. Notably, “tumor-related bleeding” is a novel category identified by the study, with oropharyngeal and pulmonary sites being most common. The study calls for standardized definitions for this unique form of bleeding.

Risk Factors

For patients not receiving anticoagulation, the following factors were associated with higher bleeding risks:

  1. Cancer Type: Patients with head and neck cancers had significantly higher CRB risk (HR=2.38).
  2. Biomarkers: Low hemoglobin (<10 g/dL, HR=0.88) and low albumin (<35 g/L, HR=0.95) were strongly associated with increased bleeding risk. Patients with both low biomarkers had a 6-month cumulative CRB incidence of 25.5%.

Interestingly, stage IV disease was not associated with an elevated bleeding risk, a finding that differs from prior studies.

Prognostic Impact of Bleeding Events

CRB events were associated with significantly increased overall mortality (HR=5.80). Among 7 fatal bleeding events, 6 occurred in patients not receiving anticoagulation (CRB case fatality rate: 6.3%; MB case fatality rate: 13.9%).

Clinical and Research Implications

This study offers the most comprehensive insights yet into bleeding risks among cancer patients:

  1. Scientific Contribution: The identification and quantification of “tumor-related bleeding” introduce a new dimension to research in this field, warranting further exploration and standardization.
  2. Clinical Applications: Hemoglobin and albumin levels are proposed as potential biomarkers for determining baseline bleeding risks and guiding personalized treatment plans.
  3. Guidance for Personalized Therapies: The findings emphasize careful monitoring of at-risk populations, particularly those with head and neck cancers or low hemoglobin and albumin levels.

Study Limitations and Future Directions

This study had some limitations. It was single-center-based, which restricts generalizability to patients undergoing other cancer treatments (e.g., surgery, radiotherapy). Additionally, smaller sample sizes for certain cancer subtypes, such as hematological malignancies, may have limited subgroup analyses.

Future research should focus on validating hemoglobin and albumin as independent predictive factors and developing risk stratification tools for optimizing treatment while balancing thrombotic and bleeding risks in cancer patients.

Conclusions

This study reveals the complexity of bleeding risk and its significant impact on prognosis in cancer patients. It not only provides a foundation for translational research in this area but also offers critical insights to improve clinical management for cancer patients.