Diagnosing pulmonary embolism

There's no shortage of tools or data for a physician trying to diagnose a pulmonary embolism (PE).

In a single week in December 2021, a systematic review assessing the Wells score and the revised Geneva scores combined with a fixed or modified D-dimer cutoff, as well as the YEARS algorithm, was published by Annals of Internal Medicine, and a study comparing YEARS combined with age-adjusted D-dimer to conventional diagnostic strategies in patients not excluded by the PE Rule-out Criteria was published by JAMA.

“It can give you information overload when you have lot of different potential competing strategies,” said Daniel Brotman, MD, FACP, author of an editorial accompanying the Annals study and a professor of medicine at Johns Hopkins University School of Medicine in Baltimore.

He and others who have researched PE diagnosis offered their advice to help clinicians sort through the latest data and quickly and efficiently diagnose PE without putting patients through unnecessary testing.

Why to score

The avoidance of imaging is a primary reason the scores and algorithms exist.

“Without use of validated clinical prediction rules, clinicians overorder CT pulmonary angiography,” said David Vinson, MD, an adjunct investigator with the Kaiser Permanente Division of Research and emergency medicine specialist with The Permanente Medical Group in Oakland, Calif.

That's a common circumstance in EDs and hospitals, noted Geoffrey Barnes, MD, a cardiologist and an assistant professor at the University of Michigan in Ann Arbor. “‘I have a CT scanner, so why not just do it?’ Especially in the United States, where we're so risk averse because of the potential legal ramifications.”

The answer is that CTs carry risks. “The risks of radiation exposure, the risk of contrast-induced nephropathy, and, of course, the risk of incidental findings on a CT scan,” said Dr. Barnes.

But that last risk can actually be a benefit sometimes, and adverse reactions aren't as likely as they used to be, countered Dr. Brotman. “CT imaging is often helpful for looking for other things. If you've got a patient with unexplained hypoxemia, unexplained dyspnea, unexplained chest pain, a CT may show something else,” he said. “The contrast burden associated with these studies and the risk of renal failure or other reactions, such as anaphylactoid reactions, are much lower now.”

Still, Dr. Barnes is concerned about the scans' costs—not just financially to the patient and health care system, but temporally as well. “As a cardiologist, I get admitted patients to my service all the time who came in with chest pain, they got a PE CT scan, and then we realize they have an elevated troponin. This person's having an acute coronary syndrome, but yet we can't take them to the cath lab for two days because we want to make sure the IV contrast is way out of their system before we give them another contrast load,” he said.

For all these reasons, “There's a need to avoid unnecessary CT scanning,” said Wendy Lim, MD, a professor of medicine at McMaster University in Ontario, Canada, and chair of the American Society of Hematology's guidelines on diagnosis of venous thromboembolism.

That's where the scores come in. “We can look at some objective evidence to help us decide in an evidence-based manner how to best stratify patients,” she said.

But which?

Unfortunately, the evidence won't exactly help a clinician decide what score to apply. “Use a validated tool. There isn't a best scoring system,” Dr. Lim said.

“These rules share a fair amount of commonality,” agreed Dr. Brotman. “You can certainly look up any rule you want, but I don't think that I would recommend to anybody, if they've got a decision rule that they're comfortable using that's been validated, that they should adopt another one because it might be slightly better.”

It can take some time to get comfortable with a score. “Validated clinical prediction rules may not be easier than gestalt on first use, but with experience and growing familiarity, ease of use improves,” said Dr. Vinson.

Health care systems can help their physicians get to that point by choosing a score and building it into the electronic health record's decision support, he added. “This makes them conveniently available at the clinician's fingertips when they are needed for medical decision making.”

Don't get too comfortable with a method just because it's convenient, however, cautioned Dr. Barnes. “Clinicians need to be aware of what other elements are either in different scoring systems or maybe not built into a scoring system at all that could influence how they approach making a diagnosis or thinking about risk stratifying,” he said.

Dr. Brotman put the caveat another way. “One of the cornerstones of diagnosis remains clinical judgment,” he said. “There have been plenty of instances where I've encountered clinical situations where somebody thinks PE is the most likely diagnosis, and I think, ‘Hogwash, it's heart failure.’”

The tools often try to match or incorporate that sense of experienced clinicians. “What these scores do is they're taking that gestalt, formalizing it in a way using variables that we know influence that risk,” said Dr. Lim.

For example, some criteria include the likelihood of PE or another diagnosis as a factor. “That's a very subjective element, but it's really critical when thinking about should you order a PE CT scan or not,” said Dr. Barnes.

Another consideration, which is not in the scores, should be whether a PE diagnosis would change management, he added. “Is this somebody who's already on anticoagulants? Is this somebody who's had bleeding problems and you wouldn't be able to anticoagulate them anyway? Those might be situations in which it's not as useful to perform diagnostic testing.”

The future likely holds only more complexities for PE diagnosis, the experts said. For one, there's a potential for new imaging options. “MRI has been studied in diagnosing [deep venous thrombosis] and PE and looks good,” said Dr. Lim. “If the technology becomes more accessible, that may be a way to go because it will avoid radiation.”

Researchers may also uncover new ways to use the existing tools. “We now recognize that D-dimer is a continuous variable,” said Dr. Brotman. “Thinking about our D-dimer testing the way we think about troponin or BNP levels as opposed to binary positive or negative tests, like an HIV test, is perhaps an important future direction for research.”

Finally, there will remain the challenge of getting the evidence, whether new or focused on the existing algorithms for PE, to be widely applied in clinical care. “There is a big gap to close between what we know to be true and what we do in the real world,” said Dr. Vinson. “Translating existing knowledge into daily practice will go a long way to improving patient care.”