AMSTERDAM – Two-thirds of patients who presented to the emergency department of a U.S. tertiary care hospital with an acute pulmonary embolism had no acute deterioration and required no short-term hospital-based interventions, in an analysis of 298 patients seen over a 2-year period.

The finding "supports the assertion that outpatient treatment of patients with pulmonary embolism [PE] is safe," said Dr. Christopher Kabrhel at the 24th Congress of the International Society on Thrombosis and Haemostasis.

"We want to identify patients for whom nothing bad will happen. We showed that two-thirds of patients did well and didn’t need anything from the hospital and didn’t benefit from being in the hospital. We need to identify some of these patients," soon after they present in the emergency department, said Dr. Kabrhel, a surgeon and emergency-medicine physician at Massachusetts General Hospital and Harvard Medical School in Boston. If reliable risk markers can be found with further research, "perhaps we can identify half of the two-thirds—a third of all patients who come to the emergency department with a PE—who we know will be safe with outpatient treatment so we can send those patients home from the emergency department and not admit them."

Most symptomatic U.S. patients who come to an emergency department, and are diagnosed with a PE are immediately admitted to the hospital. In the current study, the hospitalization rate was 92% with a median length of stay of 3 days. "We need a better rule to decide whether a patient needs hospitalization. We need to find which patients benefit from hospitalization," Dr. Kabrhel said in an interview.

He and his associates reviewed 298 adults 18 years or older who presented to the Massachusetts General Hospital emergency department during October 2009 through December 2011 with a radiographically proven PE diagnosed within 24 hours of arrival. They averaged 59 years old, half were women, and the most common comorbidity was malignancy in 107 patients (36%).

The study’s primary outcome was any clinical deterioration or need for hospital-based intervention during the 5 days following presentation at the emergency department, including the need for advanced cardiac life support, the development of a new cardiac dysrhythmia, the development of hypoxia or hypotension, the need for thrombolysis or thrombectomy, recurrent PE, or death. These events occurred in 99 patients (33%); of these, 28 patients (9% of the total group) had "severe" deterioration or required a "major" intervention. Twelve patients (4%) died within 30 days of their initial emergency presentation. The most common acute complication was the need for respiratory support, in about 55 patients, followed by hypotension, in about 34.

A multivariate analysis identified five baseline factors that significantly correlated with the primary outcome. Patients who had normal vital signs at baseline had a 79% reduced rate of significant deterioration or need for hospital-based intervention. The other four factors were linked with increased rates of deterioration and need for intervention: Right heart strain caused by the PE and identified by an echocardiogram boosted the risk of a bad outcome more than fourfold, coronary disease and cerebrovascular disease each were tied to a more than threefold increased rate, and residual deep vein thrombosis was linked with a more than doubled rate of bad outcomes.

The subset of patients with the most severe outcomes had only one direct correlation with bad outcomes, right heart strain on echo. This subset of patients also showed a protective link against bad outcomes when their systolic blood pressure never fell below 90 mm Hg.

In contrast to these factors linked to 5-day outcomes, two different types of patient factors were significantly linked with 30-day mortality: having a malignancy and having chronic lung disease.

"Previously validated clinical prediction rules that looked at outcomes after PE were primarily validated based on 30-day mortality or recurrent PE, and included factors like having cancer, heart failure, or chronic lung disease. But these scores are only able to predict the outcomes we examined with 70% sensitivity," Dr. Kabrhel said. He found this out by running the numbers he collected through three validated scores for predicting PE outcome: the Geneva Prediction Score (Ann. Intern. Med. 2006;144:165-71), the Severity Index (Am. J. Respir. Crit. Care Med. 2005; 172:1041-6), and the Simplified Pulmonary Embolism Severity Index (Arch. Intern. Med. 2010;170:1383-9). "Predictors of all-cause 30-day mortality are different than predictors of short-term outcomes" in PE patients, he said.

"We found that echo is a very good predictor of short-term outcomes, and also abnormal vital signs. The key point is we need to look at outcomes that are relevant to the decisions made" in the emergency department, Dr. Kabrhel said. "Looking at 30-day mortality in patients who are only hospitalized for 3 days doesn’t really inform the decision on who should be in the hospital. I would suggest caution on using [prediction] tools validated against 30-day mortality and recurrent PE to determine what to do acutely. We need better rules to decide which PE patients need hospitalization."

Dr. Kabrhel said he has been a consultant to Diagnostica Stago, and is an officer for LitPulse.

mzoler@frontlinemedcom.com

On Twitter @mitchelzoler

Acute pulmonary embolism doesn’t always require hospitalization