Prioritize PE in high-risk patients with syncope

When evaluating syncope, don't count out pulmonary embolism (PE) as a primary cause. Rates remain fairly low but could be higher than hospitalists may expect, said Daniel D. Dressler, MD, MSc, FACP.

Dr. Dressler, professor of medicine, director of internal medicine teaching services, and associate program director of the J. Willis Hurst IM Residency Program at Emory University Hospital in Atlanta, reviewed data from three recent studies on this question last October at a Southern Hospital Medicine 2018 talk titled “Syncope: The 24-Hour Hospital Workup.”

The potential game-changer in this area of practice, PESIT (Pulmonary Syncope in Italian Trial), was published in the New England Journal of Medicine in October 2016. PESIT was a cross-sectional study of 2,584 patients at 11 hospitals in Italy who presented to the ED within 24 hours of syncope. “This was a well-designed, prospective, essentially cohort study,” Dr. Dressler said.

The PESIT researchers performed a structured and focused history and physical exam on all eligible patients and discharged 1,867, or approximately 72%. “They were able to say, ‘Oh, this was a vasovagal episode, or reflex-mediated, or this was drug-induced, clearly based on the history,’ … or, ‘You were volume-depleted and we'll just discharge you,’” Dr. Dressler said.

The discharge percentage in PESIT was relatively high, he said, and was based on a detailed, highly structured history and physical that EDs don't always have time to do. “My ED, for syncope, is admitting closer to 50% or 60%, and probably the average in the U.S. is around 40% to 50%,” he noted.

Of the remaining admitted patients in PESIT, 157 were excluded and 560 were evaluated with the simplified Wells score and a D-dimer test. PE was ruled out in 330 of 560 patients (58.9%) due to low pretest probability of PE and negative D-dimer results. This group either underwent further evaluation for other conditions besides PE or were sent home, Dr. Dressler said.

The 230 remaining patients had imaging with CT of the chest or a ventilation-perfusion scan. Dr. Dressler stressed that imaging was done only in patients who had positive results on D-dimer testing and no clear, low-risk cause of syncope. “If it was clear ahead of time, [they] wouldn't even have done this. [They] wouldn't have checked the D-dimer,” he said. Of these 230 patients, 97 (42.2%) were found to have PE. The incidence of PE in the entire cohort of hospitalized patients was 17.3%.

“They found that almost one in five patients who were admitted, who were evaluated for syncope, had PE,” Dr. Dressler said. Of these, two-thirds had large-vessel PE, and one-quarter of those with PE had no clinical symptoms or manifestations. The study was also one of the first to indicate that tachypnea could be a predictive factor for PE, since it was present in 45.4% of patients with PE versus 7.1% of those without, he noted.

The rate of PE in this study was much higher than the rate of less than 1% that was previously considered the standard, and other investigators took notice, Dr. Dressler said. “In essence, some other investigators said, ‘Hmmm … we're not seeing this high PE rate, and we're going to start doing too many imaging studies, and so let's look at all the data out there.’”

A systematic review and meta-analysis published by the American Journal of Emergency Medicine in September 2017 analyzed 12 studies involving 6,608 ED patients and 975 hospitalized patients. The authors found a pooled estimate of PE prevalence of 0.8% in the former and 1.0% in the latter, considerably lower than the PESIT estimates of 3.8% and 17.3%. The authors said that the PESIT results suggest “a possible site effect, accrual bias, or investigation strategy” and called for further study before implementing changes in practice.

In addition, a large retrospective cohort study, published as part of the Less Is More series in JAMA Internal Medicine in March 2018, looked at data from five administrative databases in Canada, Denmark, Italy, and the United States, including more than 1.5 million patients who presented to the ED with syncope from Jan. 1, 2000, to Sept. 30, 2016. PE prevalence ranged from 0.06% to 0.55% among all patients and from 0.15% to 2.10% among hospitalized patients. At 90 days, prevalence of PE ranged from 0.15% to 0.83% in the former and from 0.35% to 0.63% in the latter. The authors of this study said that PE was rare in patients with syncope. “Although PE should be considered in every patient, not all patients should undergo evaluation for PE,” they wrote.

Dr. Dressler pointed out that the data in the American Journal of Emergency Medicine and JAMA Internal Medicine studies were primarily from retrospective cohorts, most of which did not use specific definitions of syncope and did not use the kind of structured prospective evaluation seen in PESIT, “so we're looking at very different populations.” He also noted that some of the patients involved in the JAMA Internal Medicine study were outpatients rather than ED patients and were therefore less likely to be at risk for PE.

All three studies, taken together, offer a muddled message for hospitalists, so Dr. Dressler aimed to provide some perspective. “Is the percentage of patients who come in with syncope, who are admitted to the hospital, who have a PE, 17%? No, it's probably not. … Reconciling these results, I'd say up to 1 in 6 patients admitted with syncope may have a PE, but probably closer to 1 in 15 or 1 in 20, based on a much lower admit rate in the study than we typically see, and some syncope patients—in practice—may not be receiving the same structured ED evaluation as this study did,” he said. “But it's still a much higher rate than we suspected previously, when we thought it was less than 1%.”

Regarding the potential effect of these results on practice, Dr. Dressler said ED physicians should always do a thorough workup on patients who present to the ED with syncope. He noted, however, that imaging for these patients in the ED does not appear to be necessary.

“That is a process that should happen once they come into the hospital, once they get admitted,” he said. “Let's get a second set of eyes and someone who's got potentially a little bit more time to do a workup.”

If PE can't be ruled out in the ED and the patient is admitted, hospitalists should use their own assessments to build on the ED's findings, Dr. Dressler said. In his own hospital, for patients whose workup doesn't unmask a clear cause of syncope, he said he follows the protocol outlined in PESIT. “If I do a structured evaluation after my ED physician and I still can't figure out that there's something low risk, then I am going to go about the same process that [the PESIT investigators] did,” he said: a quick simplified Wells score (“which takes about five seconds”), followed by a D-dimer test.

“And then if it's positive,” he said, “I'm going to order the imaging.”