Evidence-Based Medicine

Best Evidence Says: The PERC Score Can Help EPs Avoid Unnecessary Tests
The Setup:
You are working a busy critical care shift that has provided you with a variety of chest pain complaints. Your last patient chart is handed to you and the chief complaint is…chest pain. You interview a pleasant 30 year old female who complains of an episode of acute pleuritic chest pain while picking up her 4 year old son two hours ago. She had some mild dyspnea at that time which has since resolved. She is otherwise healthy and takes no medications. Her vital signs and physical exam findings are unremarkable. You think that her symptoms are most likely attributable to musculoskeletal strain and that this is a very low risk presentation for pulmonary embolus (PE).

Despite your risk stratification, you wonder if you should order a d-dimer to rule out PE?

The Study:
Kline JA, Courtney DM, Kabrhel C, Moore CL, Smithline HA, Plewa MC, Richman PB, O’Neil BJ, Nordenholz K. Prospective multicenter evaluation of the pulmonary embolism rule-out criteria. Journal of Thrombosis and Haemostasis. 2008 May;6(5):772-80.

The Big Idea:
PE is the second leading cause of unexplained death in the US (acute coronary syndrome is first).  Because of its vague presenting symptoms and potential for significant morbidity and mortality, over-testing to exclude this disorder is commonplace in the ED. Aside from resource overutilization and cost, diagnostic testing for PE is not without complications. Most significantly, chest CTs expose patients to harmful radiation as well as places them at risk for contrast induced nephropathy. D-dimer testing in low-risk individuals has received Level B recommendations from the American College of Emergency Physicians. Unfortunately, the d-dimer is frequently false-positive with multiple confounding variables including advancing age. Furthermore, EP’s rarely compute pre-test probabilities prior to d-dimer ordering despite the availability of validated risk stratification tools. Application of clinical gestalt and the PERC rule offer the potential to minimize diagnostic testing for PE in low risk ED populations. If your clinical suspicion is low and the eight PERC characteristics (Table 1) are absent, you can forgo testing for PE (d-dimer or chest CT).

Table 1
The 8 Components of the PERC Rule
1. Age ≥ 50
2. Pulse ≥ 100
3. RA pulse Ox ≤ 94%
4. Current history of hemoptysis
5. Estrogen use
6. Prior diagnosis of venous thromboembolis
7. Recent surgery or trauma in the last 4 weeks
8. Unilateral leg swelling

The Evidence:
A large randomized controlled trial recently validated the PERC rule which was derived in 2004. This trial enrolled over 8000 patients from 12 EDs in the US and 1 in New Zealand including both urban and community based centers. Patients were eligible if the ED MD ordered a diagnostic test to exclude PE. Physicians were also asked to fill out a standardized data collection form to include their clinical probability for PE (low, moderate, or high risk) before test results were obtained. All patients were followed up at 45 days to determine if they had suffered a PE. In patients that were considered low risk for PE by clinician gestalt (1666 patients), application of the PERC rule would have safely eliminated diagnostic testing in 99%. In other words, applying the PERC criteria to a low risk population of ED patients in whom PE is suspected would result in missing the diagnosis only 1% of the time.

The diagnostic characteristics of the PERC rule are summarized in Table 2 (below).

TABLE 2                                         Result               95% Confidence Interval
Sensitivity                                        97.4%                95.8% to 98.5%
Specificity                                        21.9%                21.0% to 22.9%
Likelihood Ratio Negative                  0.12                    0.07 to 0.19

The Caveats: 
Deciding your comfort level with missing a particular disease is important in how decision aids such as the PERC rule are interpreted and implemented. Most would argue that a miss rate of 0% is neither attainable nor reasonable. For disorders that carry significant morbidity and mortality, many physicians would consider a miss rate of 2% or less to be acceptable. While the PERC rule meets this cut off, be sure that your definition of low risk mirrors that of the practicing MDs in this trial.

The Outcome:
After reviewing the evidence, you are relieved to note that your patient is entirely PERC negative. Combined with your initial clinical gestalt the PERC rule supports your decision to not pursue an evaluation for PE. Two months later, your decision is validated when the patient returns for an unrelated minor injury and has remained healthy.

Rawle A. Seupaul, MD, is an associate clinical professor of EM at the Indiana University School of Medicine is on Best Evidence in Emergency Medicine (BEEM) faculty. BEEMCourse.com
Continue NEXT for Dr. Kevin Klauer's article  "PE Clinical Decision Rules: Worth the Risk?" 
PE Clinical Decision Rules: Worth the Risk? 

by Kevin Klauer, DO
Clinical decision rules for the diagnosis of pulmonary embolus have been significantly improving over the past several years. From Well’s, not so good, to Charlotte and the modified Charlotte criteria and now the PERC rule which are relatively good rules. Despite the function and accuracy of these decision rules, you have to ask yourself, “What is the ultimate goal of these rules?”

These decision rules are designed to decrease utilization of CT for pulmonary embolus evaluations, not to identify more PEs. In a low prevalence disease, this might be an acceptable concept. When no one has the disease, you don’t have to worry so much about missing it. Although no one really knows the true prevalence of pulmonary embolism, it is believed to be around 1% in hospitalized patients, Stein, et al., Chest, Oct 1995.  If this number holds true for the general population, PE is not an uncommon phenomena.

Based on what it is designed to do, PERC is a good decision rule. However, they admittedly report a miss rate of around 2%. Which side of the fence are you on? Are you a medical economist or do you practice defensively? If you are medical economist, 2% is probably an acceptable miss rate. If you are worried about missing PEs, 2% may not be. 

Whenever I talk about this topic, there are many advocates for these decision rules, and the discussions are passionate and their support is vigilant. However, you can hear a pin drop when I ask, “OK, Who’s willing to step up and miss the first PE?” Guaranteed, by the design of these rules, you will miss PEs. How long would it take you to evaluate 50 patients with suspected PE? This may be the same amount of time it may take you to miss one PE.

The reality is that utilization management in such a serious disease with potentially disastrous consequences is a very risky business. You may reduce the number of CTs you’ll order, but is that worth missing a PE? For me, it’s not. Missed PE cases almost universally have bad outcomes and frequently result in death. They are frequently in patients that were either inappropriately risk stratified to a low-risk group or in patients that these types of rules would just plain miss. 

Here is the reality check. There are two important points to remember.

1. The patient who’s PE is missed won’t care about PERC, Charlotte or Well’s.

2. The savings to the healthcare system gained by following these rules won’t be factored in during your trial.  “Jury, Dr. X performed 40 less CT scans in the year 2009 by following these rules, which has resulted in a cost savings of nearly $40,000. He should be applauded for his careful ordering practices.”

Tell it to the dead lady’s family!

This type of argument doesn’t save lives and it won’t impress juries. Sure, reduce cost and be more efficient. They’ll just expect you to order the right test on the right patients. When you think about it, this is healthcare rationing. Our patients and our medical legal climate are not prepared to save cost at the expense of a missed PE.  Without tort and healthcare reform providing protections for such behaviors, why are we accepting all the risk? If the government wants meaningful healthcare reform and to reduce healthcare expenditures, I am all for it. However, they cannot expect physicians to bear the brunt of the consequences. 

OK, maybe you are not a medical economist. Maybe you think of the CT scanner as a small “Three mile island.” I agree with you. Unnecessary ionizing radiation exposure should be a serious consideration when you are ordering these tests.  Although we will never have a randomized, controlled trial answering this question, (Who would sign up for that one?) intuitively it is reasonable to assume that these exposures, particularly that of CT, are not benign.

So, what is the answer? Most make a judgment call for the patient, using decision rules like PERC to justify not doing the test and reduce radiation exposure? Hey, whose decision is it anyway?

The best approach is an informed discussion about what pulmonary embolism is, their risk for PE (based on PERC), the risk of radiation exposure to investigate it and the likelihood of missing a PE and the potential outcomes associated from such a miss.  It’s their health, and it’s their choice. Document the discussion and their decision and you have met your obligation.


# Number Needed to Test for PE Protocol CTChris Carpenter 2009-05-01 04:40
Kevin & Tony:

Based upon the PIOPED-II data, what is the Number Needed to Test to diagnose one PE with PE protocol CT in a low-risk subset? Concurrently, what is the Number Needed to Harm? In your opinion, what is the no further testing vs. continued testing threshold whereby most physicians would feel uncomfortable exposing patients to the risks of further testing to exclude one PE? Obviously, the threshold cannot be zero since that would mean testing everybody (symptomatic or not, you & I included) for PE everytime they come into the ED for anything. So what is the threshold and how do we logically define and compute the threshold?
# NYS Office of Professsional Medical Conduct Boardalexander kuehl MD. MPH 2009-05-05 14:05
In my experience.. if one thinks about PE in a differtential diagnosis of an ED patient..the next step is a high reliability D-dimer. If it is positive you have bought a full scale workup..if it is negative and there is a logical clinical explanation for the (usually) pleuritic chest pain... the patient can be discharged. If you are sued, call me..I 'll be your trial expert for free.

Those physicians that demean the value of a negative d-dimer in in ruling out a clotting disorder(in the absense of any symptoms of such)..are doing emergency
medicine and the health system a great disservice.
# MDRawle Seupaul 2009-05-07 23:51
For any potentially life threatening condition, EM MDs have to ask the question: At what cost are we willing to not miss the diagnosis. Most (myself included) would argue that it is impossible to catch every clinically significant PE. So what miss rate should we be comfortable with and what are the considerations of harm from over-testing? In Kline’s validation of the PERC rule, he defines 2% as the cut point (derived from a fairly sophisticated statistical calculation). This seems reasonable and should be considered with the fact that 1. contrast induced nephropathy now accounts for more than 10% of hospital acquired renal failure and1 and 2. there is a real risk of radiation induced malignancy2. Perhaps the best population to limit over testing is in the low risk population as suggested by Kline (moderate to high risk patients should be tested and would fail the PERC rule). This is especially important when the performance of the diagnostic gold standard, chest CT, is considered in low risk populations. In the PIOPED II trial, the sensitivity of chest CT in the diagnosis of PE in low risk individuals was only 79%! Therefore, if you wanted to capture all PEs in this population, you would not only have a large number of false positive scans, but would need another test to capture the false negatives (when was the last time you ordered a pulmonary angiogram?). This would not only unnecessarily expose a large number of patients to harmful contrast material and radiation but require more invasive and potentially harmful testing. In the end, judicious use of clinical gestalt with decision aids, like the PERC rule, can help prevent this predicament.

1.Merten GJ, Burgess WP, Gray LV, Holleman JH, Roush TS, Kowalchuk GJ, Bersin RM, Van Moore A, Simonton CA 3rd, Rittase RA, Norton HJ, Kennedy TP. Prevention of contrast-induce d nephropathy with sodium bicarbonate: a randomized controlled trial. JAMA. 2004;291:2328-2334.
2.Klevens M, Carpenter C. CT Radiation: Today’s Asbestos? http://www.epmonthly.com/index.php?option=com_content&task=view&id=292&Itemid=15. (Accessed May 5, 2009)
# PE's in MexicoAdan Atriham 2009-06-03 22:50
After almost a decade of practice in the US, I have returned to my home country. In this side of the world where no fancy tests are available, the clinical skills need to be sharp to catch these cases. Having practiced now with out the high tech stuff available in the neighbor country of the north, if the clinical suspicious is there, we look at the legs first. If doppler is ( ) then we go ahead and treat. If doppler is (-) we scan. Using this practice model, I don't think I have ordered more CTPA's that I would have if still practicing in the US.

Add comment

Security code

Popular Authors

  • Greg Henry
  • Rick Bukata
  • Mark Plaster
  • Kevin Klauer
  • Jesse Pines
  • David Newman
  • Rich Levitan
  • Ghazala Sharieff
  • Nicholas Genes
  • Jeannette Wolfe
  • William Sullivan
  • Michael Silverman

Earn CME Credit