Skip to Main Content


Skip Nav Destination

More Than a Guessing Game: Risk Prediction in Thrombosis

December 9, 2023
Hilary Whitworth, MD, MSCE (@HilaryWhitworth)  
Division of Hematology, Children’s Hospital of Philadelphia, Philadelphia, PA 

Identifying patients at high risk for thrombosis should be more than just gazing into a crystal ball to see the future. Quantifying risk and putting it in the context of individual bleeding risk is the optimal game plan for intervening with primary thromboprophylaxis and preventing the potentially devastating consequences of blood clots in our patients. As we have learned more about the epidemiology of thrombosis, there has been increased need and enthusiasm for the development of prediction models for thrombosis risk. These remain at the forefront of easily scalable tools and promise to be of high impact for clinical practice and research.  

This year’s presenters tackle challenging situations, not the least of which are thrombotic complications in adults with cancer, for whom venous thromboembolism (VTE) remains the second leading cause of death. The session Thrombosis and Anticoagulation: Clinical and Epidemiological: Clonal Hematopoiesis of Indeterminate Potential (CHIP) and Cancer Associated Thrombosis (CAT): Understanding Risk as a Potential Means to Improve Patient Care (Sunday, 4:30 p.m. to 6:00 p.m., Convention Center, Room 28 A-D) focuses on multiple relevant studies striving to overcome the limitations of currently available models.

During the session, Shengling Ma, MD, will discuss a disease-specific model for improving prediction in patients with lymphoma. As a pediatric hematologist, I cannot imagine having enough patients to derive and validate a risk-assessment model to this degree of specificity and can’t help but dream of a time when we will have the ability to be this accurate for every patient population. Simon Mantha, MD, MPH, and colleagues will share their work on dynamic modelling, which moves beyond typical risk prediction models to allow clinicians to have an updated assessment of VTE risk as the patient’s clinical status changes. This sort of “next-generation” prediction modeling incorporates time-dependent covariates that change as the patient does and is particularly important for patients receiving chemotherapy that may result in thrombocytopenia and those who may develop an increased risk of bleeding over time. 

Of course, preventing thrombosis with anticoagulation is not always straightforward. Often, the patients at the highest risk of thrombosis are also at a high risk of bleeding. Thankfully, Kristen Sanfilippo, MD, will help us get to the bottom of this precarious balance in her oral presentation “Discriminating Risk of Anticoagulant-related Bleeding in Ambulatory Cancer Patients on Thromboprophylaxis.”

Dr. Sanfilippo and colleagues conducted a post hoc analysis of anticoagulant-related bleeding prediction scores using the data from a prior trial assessing thromboprophylaxis in patients with cancer. They found that the RIETE score performed best, but all scores had opportunities for improvement in predicting anticoagulant-associated bleeding. They identified anemia and antiplatelet therapy as associated with increased bleeding risk during prophylactic anticoagulation. “The presence of these findings … may prompt reconsideration of starting prophylaxis, while the absence may provide some reassurance,” Dr. Sanfilippo said. However, she cautioned that future work is needed to better quantify the bleeding risk in this population. Her team is working to do this by adding cancer-specific variables to prediction models. Their work proves that while we have a myriad of prediction scores, we are still not done improving care for our patients.  

Pediatric hematologists are also working hard to predict and prevent thrombosis. While thrombosis is much less common in children than adults, the rate of thrombosis in hospitalized children is rising, and the outcomes can be significant. Julie Jaffray, MD, MS, will share ongoing efforts from the Children’s Hospital-Acquired Thrombosis (CHAT) Consortium to develop prediction models for pediatric thrombosis in the session Thrombosis and Anticoagulation: Clinical and Epidemiological: Genetics, Pregnancy, and Pediatrics: Advances in Venous Thromboembolism (Monday, 2:45 p.m., Convention Center, Room 33). Dr. Jaffray will present the abstract “Multicenter Study of a Risk Prediction Model for Critically Ill Children at High-Risk for Hospital-Acquired Venous Thromboembolism: Findings from the Children’s Hospital-Acquired Thrombosis (CHAT) Consortium.” This validation study included critically ill children from 32 centers in the U.S. and identified five risk factors for VTE in critically ill, hospitalized children. These data will hopefully help us decrease rates of hospital-acquired (HA) VTE in children by identifying those that most benefit from thromboprophylaxis. With less clarity on the risk of bleeding in pediatric patients receiving pharmacologic prophylaxis, I know I will be at Dr. Sanfilippo’s talk tomorrow to learn some pointers on risk prediction in adults, which will hopefully pave the way for similar models in children that balance bleeding and clotting risks.         

Thanks to the outstanding work of our colleagues — whether it be strictly pinpointing disease-specific thrombotic risk, adjusting our scales for dynamic changes in patient conditions, or finely balancing on that tightrope of bleeding/clotting — we need not rely on clairvoyance or mystical orbs to make the most informed clinical decisions for our patients. State-of-the-art models will soon help to demystify the prediction of thrombosis risk in both pediatric and adult patients. 



Close Modal

or Create an Account

Close Modal
Close Modal