Yoshihiro Tanaka, MD, PhD, on Predicting Bleeding Risk in Patients With AFib Taking DOACs


In this podcast, Yoshihiro Tanaka, MD, PhD, answers our questions about his team's study on predicting bleeding risk among patients with atrial fibrillation who are taking direct oral anticoagulants instead of warfarin. He also presented these findings at the American Heart Association’s Scientific Sessions 2021.

Additional Resource:


Yoshihiro Tanaka, MD, PhD, is a research fellow at the Center for Arrhythmia Research at the Feinberg School of Medicine at Northwestern University.



Amanda Balbi: Hello, and welcome to another installment of Podcasts360—your go-to resource for medical news and clinical updates. I’m your moderator, Amanda Balbi with Consultant360.

Today I’m speaking with Dr Yoshihiro Tanaka, MD, PhD, who is a research fellow at the Center for Arrhythmia Research at the Feinberg School of Medicine at Northwestern University. He recently presented his team’s study, “New Bleeding Risk Prediction Model for Patients with Atrial Fibrillation on Direct Oral Anticoagulants,” at the American Heart Association’s Scientific Sessions 2021, which he will be speaking with us about today.

To start, can you give us an overview of your study and its findings?

Yoshihiro Tanaka: Direct oral anticoagulants (DOACs) are more commonly prescribed in patients with atrial fibrillation, and because I think DOACs for stroke prevention is equivalent to warfarin with less bleeding complications.

Several risk prediction models for bleeding least are currently available. However, these risk scores were developed in patients on warfarin, the conventional type of anticoagulant. Therefore, the development of new risk prediction models for bleeding specific to DOACs is essential for accurate prediction of bleeding risk and the prevention of serious complications.

Therefore, we examined whether a DOAC-specific risk prediction model will have better prediction ability compared with the conventional one for 1-year and 3-year major bleeding risk events.

Amanda Balbi: So, can you tell us a little bit more about the prediction model, what factors are included, and how clinicians might be able to use it?

Yoshihiro Tanaka: To make a prediction model, we used the Cox regression model to identify a significant predictor in a univariate model. Given we found a significant association between some factors with major breeding risk, all those factors are incorporated into the multivariate Cox proportional hazards model. Harrel’s c-statistics were useful for the discrimination and the Hosmer-Lemeshow test for calibration.

Amanda Balbi: Perfect. So how can clinicians use your findings in clinical practice? How can they implement them?

Yoshihiro Tanaka: In this study, the final model involves age, race, body mass index, in addition to all the components of the CH2AD2-VASc score, this is used for bleeding risk prediction in patient on anticoagulants.

It is reported that there is a huge racial disparity in the cardiovascular event rate in patients with atrial fibrillation. So, adding race and the body mass index to the conventional risk model can improve the prediction of bleed ability. This really makes sense.

Race and BMI information are easily available in daily clinical practice, so clinicians should pay more attention to their racial or BMI information when you see patients with atrial fibrillation on DOACs.

Amanda Balbi: Perfect. So, what is the importance of race in your prediction model, and how is that a new thing compared to other prediction models?

Yoshihiro Tanaka: Generally speaking, White patients are more likely to have incident atrial fibrillation compared with Black patients, but cardiovascular complications are more likely happening in Black patients compared with White patients.

This is probably because Black patients have more cardiovascular comorbidities compared with White patients. Educational or socioeconomic disparities may cause these huge disparities. So, race has a lot of meaning in patient control.

Amanda Balbi: What is the next step in your research in this area?

Yoshihiro Tanaka: In this study, we compared the risk prediction ability with a conventional reported prediction ability. So, we have to validate our results using other populations, such as the Asian population or more general population. Because, in this study we mainly included patients who visited our institutions, which means this study included a very high-risk population. So, we don't know the efficacy of this prediction model on the general population or low-risk patients.

Amanda Balbi: Great. Thank you so much for speaking with me and answering my questions about your research.

Yoshihiro Tanaka: Thank you for providing a precious opportunity to introduce our research.