SAEM: Clinical Protocol Can Rule Out Pulmonary Embolism

SAEM: Clinical Protocol Can Rule Out Pulmonary Embolism


CHICAGO, May 17 -- Simple clinical criteria to rule out a pulmonary embolism can substitute for blood tests and CT scanning, according to researchers here.

The criteria, making up the so-called the PERC rule, consider eight clinical factors. If all eight are absent (PERC-negative) the probability of pulmonary embolism is quite low, said Jeff Kline, M.D., of the Carolinas Medical Center in Charlotte, N.C., and colleagues, at a Society for Academic Emergency Medicine meeting here.

The PERC rule includes clinical criteria such as age less than 50, pulse rate less than 100, 95% oxygen saturation level, no prior pulmonary embolism or deep vein thrombosis, and no recent surgery.

Ninety percent of the blood tests and ultrafast CT scans to detect pulmonary embolism turn out to be negative, the investigators said.

Over-testing for pulmonary embolism has emerged as a contentious issue, the researchers said. The CT scan may cost $2,500, has the potential for kidney damage in one of 12 patients, and has a heavy radiation dose. Yet physicians are ordering these tests for 2% to 3% of all emergency department patients, or 3.5 million patients a year, Dr. Kline said. Furthermore, the fear of medical malpractice has led to doing tests rather than using the data available through clinical evaluation.

To evaluate PERC, the authors from 2004 to 2006 enrolled 8,138 consecutive or random patients from 13 U.S. academic and community emergency departments staffed by emergency physicians. The patients' chief complaints were chest pain (52%), dyspnea (30%), cough (3%), syncope (2%), or other (12%).

The physicians' evaluations preceded the test results. Tests for pulmonary embolism were D-dimer (72%), CT (54%) and VQ scans (8%).

In the best case, PERC could reduce testing for pulmonary embolism by about 20% to 25%, the researchers said. Fully two-thirds of the testing was done for patients where the physician believed the probability of pulmonary embolism was less than 15%.

The authors found that the emergency physicians thought another diagnosis was more likely than pulmonary embolism in 83% of the cases, and that 67% were at low risk.

Of the patients, 524 (6.9%) were positive for image proven venous thromboembolism (pulmonary embolism or deep vein thrombosis). The median prevalence of positive venous thromboembolism across the 13 emergency departments was 6.7%.

From the entire cohort, 25% were PERC-negative, and only 1.2% of these patients were positive for venous thromboembolism. This equaled a diagnostic sensitivity of 95.6% (93.5 to 97.2%) and a specificity of 25.6% (24.6 to 26.6%).

Among low-risk patients, 1,519 (20% of the cohort) were PERC-negative, 14 (0.9%, 0.5 to 1.5%) were positive for venous thromboembolism, but none died.

Thus, low risk and PERC-negative had a sensitivity of 97.3% and specificity of 21.5%.

The combination of an emergency physician's impression that a patient is at low risk for pulmonary embolism, together with a negative PERC, provides compelling rationale to not order a test for pulmonary embolism, the researchers concluded.

This multicenter collaborative project shows that a careful history and physical examination can achieve the same degree of certainty without the cost and side effects of diagnostic testing.

The ability to document "PERC Rule negative" on the chart will provide physicians with the scientific and medicolegal backstop they need to justify not ordering a test on every patient with even a hint of pulmonary embolism, the investigators said.

Primary source:

Society for Academic Emergency Medicine 2007 MeetingSource reference: Kline J et al "Prospective, Multicenter Validation of the Pulmonary Embolism Rule-Out Criteria," presented May 16.

Abstracts are published in Vol. 14, Issue 5S, the May 2007 supplement of the official journal of the SAEM, Academic Emergency Medicine