Home > Cardiology > Model to predict bleeding, thrombotic risks may help tailor PCI management

Model to predict bleeding, thrombotic risks may help tailor PCI management

JAMA Cardiology
Reuters Health - 15/01/2021 - Clinical decision-making at the time of percutaneous coronary intervention (PCI) in certain patients can be aided by a model which takes into account both bleeding risk and the likelihood of other untoward events, according to an international group of researchers.

"PCI patients at high bleeding risk (HBR) are known to also have an increased risk of coronary thrombotic events (myocardial infarction and/or stent thrombosis). We believe our model can assist physicians in assessing the absolute and relative risks of bleeding versus thrombosis after PCI in individual HBR patients," Dr. Philip Urban of Hopital de la Tour, in Geneva, told Reuters Health by email.

To develop and validate their model, Dr. Urban and colleagues used individual data from six studies involving more than 6,600 HBR patients with an average age of 80 years.

The studies took place between 2009 and 2017 in more than 200 centers in Europe and the U.S. Patients were followed for one year and 33 baseline variables were used to calculate the risks of myocardial infarction (MI), stent thrombosis and major bleeding events.

During follow-up beyond the procedure, 350 patients (5.3%) experienced MI or stent thrombosis alone or in combination. Major bleeding as defined by the Bleeding Academic Research Consortium (BARC) type 3 to 5 was seen in 381 patients (5.7%).

Four variables, among them being more than 65 years old and having chronic obstructive pulmonary disease, were associated only with BARC types 3 to 5 bleeding. Four variables, including prior MI and diabetes treated with either insulin or oral medication, were the only ones significantly associated with MI and/or stent thrombosis.

Both risk models showed moderate discrimination, the researchers report in JAMA Cardiology, with a C statistic of 0.69 for predicting MI and/or ST and 0.68 for predicting BARC types 3 to 5 bleeding. When employed in a validation cohort of more than 1,400 patients the strength of discrimination was similar.

The researchers note that at one year, compared with patients without such factors, those with elevated BARC had a hazard ratio for mortality of 3.7. For those with MI and/or stent thrombosis, the corresponding value was 6.1. Both risk increases were statistically significant.

Across the six studies, 44.1% of the patients at high bleeding risk had a greater risk of MI and/or stent thrombosis than of BARC 3 to 5 bleeding, while 23.4% had a greater risk of BARC 3 to 5 bleeding than of MI and/or ST. A third of the patients had a comparable risk of both events.

Dr. Urban concluded that the model "is based on readily available clinical and laboratory values and will shortly be available on a smartphone app for use at the bedside. The app should be helpful in tailoring the PCI procedure itself and the post-procedure anti-thrombotic treatment to the individual patient's needs."

Dr. Renato D. Lopes of Duke University Medical Center Durham, in North Carolina, told Reuters Health by email, "With the advance of antithrombotic therapies, there is a need to balance ischemic and bleeding risk when treating patients with coronary disease so we can try to find what I call the 'antithrombotic sweet-spot,' in other words, the right combination of drugs, at the right dose, for the right patients and for the right duration. Finding the sweet-spot is not easy since most of the factors that are associated with the increased risk of an ischemic event are the same factors associated with the risk of bleeding."

Dr. Lopes, a professor of medicine in the Division of Cardiology, who was not involved in the study, added, "Many randomized trials have been performed and tried to test different antithrombotic regimens for different clinical scenarios. I would say that this is the best way to advance the field - through high-quality evidence testing a strategy A versus a strategy B, for example, in different clinical scenarios."

However, "researchers and physicians continue to use different databases to explore different ways to try to identify predictors of bleeding and ischemic event."

While the new study was "well done," he said, it is limited by its retrospective nature, a combination of heterogeneous RCTs with one registry with different endpoint definitions and follow-up and with only modest overall discrimination.

He added that the research "does not change the way clinical decisions are made about the use of antithrombotic therapy or its duration in clinical practice. It does, however, provide information that can be applied during the design of prospective randomized trials testing different antithrombotic strategies in different scenarios, which is the main way that will help physicians in clinical practice to find the antithrombotic sweet-spot."

SOURCE: https://bit.ly/3oJSIRd JAMA Cardiology, online January 6, 2021.

By David Douglas

Posted on