Maria de Winter, MD/PhD student , UMC Utrecht
Thrombosis and Haemostasis
Author Interview
Stroke, Systemic or Venous Thromboembolism
T de Winter et al.
Thromb Haemost 2022; 122(05): 818-829
DOI: 10.1055/s-0041-1735251
T&H
Why did you (and your colleagues) write this paper? What was its main purpose?
MARIA DE WINTER
Estimating bleeding risk plays an important role when deciding on type and duration of treatment for cancer-associated thrombosis (CAT). Although a multitude of risk scores for bleeding during anticoagulation exist, assessing bleeding risk among patients with cancer remains challenging for multiple reasons. First, having cancer often automatically categorizes patients as having a high risk of bleeding. Second, the predictive performance of existing risk scores has never been assessed in this population specifically. Furthermore, in addition to well-known risk factors, factors related to the cancer itself and cancer treatment likely have a profound impact on risk of bleeding which are not taken into account by previous scores. The aim of this project was to investigate how best to assess bleeding risk in patients with CAT. Therefore, we first set out to investigate the performance of existing risk scores in CAT. Then, we looked into two novel risk classification methods: a pragmatic classification based on cancer type, and a newly derived prediction model.
T&H
What are the main conclusions?
MARIA DE WINTER
Existing risk scores for bleeding during anticoagulation developed in a broader population of patients with venous thromboembolism (VTE) perform insufficiently in patients with CAT. Estimates of clinically relevant bleeding risk may be obtained by a pragmatic classification based on cancer type (gastrointestinal, genitourinary or other cancer) and provides slightly better estimates of clinically relevant bleeding risk. Further improvement may be achieved with the newly derived CAT-BLEED model, consisting of age, creatinine clearance and 4 cancer-related predictors. However, external validation is warranted to assess its clinical value.
T&H
What are the paper's implications? - to the public? -to medical professionals?
MARIA DE WINTER
Our paper provides guidance on how to estimate risk of bleeding in patients with cancer-associated thrombosis. Existing risk scores discriminate insufficiently in patients with CAT, because they were not developed for this purpose, because all patients are classified as high risk or because factors related to cancer or cancer treatment are not included in these scores. Our results emphasize the importance of type of cancer (i.e. gastrointestinal, genitourinary or other cancer) and other cancer-related aspects when looking at risk of bleeding.
T&H
Are the findings clinically significant? Should the findings change practice?
MARIA DE WINTER
As venous thromboembolism is common among cancer patients and anticoagulation is the cornerstone of treatment for all of them, our results apply to a large number of patients. Our results indicate that existing risk scores for bleeding during anticoagulation developed in the total population of patients with VTE insufficiently discriminate in the subpopulation of patients with cancer. Therefore, we believe that they are of limited value in clinical practice for this population. While awaiting the results of external validation of CAT-BLEED in other populations, we suggest to use a pragmatic classification based on type of cancer (i.e. gastrointestinal, genitourinary or other) and well-established risk factors, such as history of bleeding, severe thrombocytopenia or anemia to assess risk of bleeding for individual patients in clinical practice.
