Skip to content

When the PERC Rule Fails

2016-11-11T19:19:14+00:00

PEKline et al developed a clinical decision tool based on parameters that could be obtained from a brief initial assessment to reasonably exclude the diagnosis of pulmonary embolism (PE) without the use of D-dimer in order to prevent unnecessary cost and the use of medical resources. 1 Many of us have used the Pulmonary Embolism Rule-out Criteria (PERC) rule by now, but we should be clear on what it includes. Are we using it appropriately?

PERC Rule Criteria

  1. Age < 50 years
  2. HR < 100 bpm
  3. Room air oxygen saturation > 94%
  4. No prior history of DVT or PE
  5. No recent trauma or surgery
  6. No hemoptysis
  7. No exogenous estrogen
  8. No clinical signs suggestive of DVT

If ALL criteria are met, then the patient can be called “PERC ruled out” or “PERC negative.”

What are the performance characteristics of the PERC rule?

A review and meta-analysis published in Annals of EM in 2012 found 12 qualifying studies evaluating the PERC rule and ultimately determined the following 2 :

  • The pooled sensitivity of the PERC rule is 97.2%.
  • The pooled negative LR was 0.17.
  • The overall proportion of missed PEs was 0.32% (44 of 13,855 total cases).

When does the PERC rule fail?

So who are the CT-PE or V/Q positive patients who could have been falsely “PERC ruled out?” Is there data on false PERC rule negative patients as determined by a positive CT-PE or V/Q scan?

Kline et al reworked the data from a previous paper showing the outcomes of patients who presented to the ED and were diagnosed with PE 3 and used it as a dataset to retrospectively determine the characteristics of known PE patients who had zero of 8 PERC criteria. Out of 1,880 patients with PE, 114 were PERC-negative if it would have been applied. 4 Of these patients, they found only 3 variables demonstrating a significant difference in proportions between the PERC-negative and PERC-positive groups. These variables were:

  1. Pleuritic chest pain (56% of 114 PERC negative patients vs 38% of 1,766 PERC positive patients)
  2. Pregnancy (4% of 114 PERC negative patients vs 1% of 1,766 PERC positive patients)
  3. Postpartum status (4% of 114 PERC negative patients vs 1% of 1,766 PERC positive patients)

PE and pleuritic chest pain

Pleuritic chest pain may be more predictive of PE than you think, as noted in the above statistics for false PERC negative cases. Variables commonly believed to modify the pretest probability of PE have been compared to those already within the existing pretest probability scores. To compare and quantify the predictive value of pleuritic chest pain, substernal chest pain, dyspnea, estrogen use, family history of PE, and patient history of thrombophilic condition (such as protein C deficiency or Factor V Leiden) to aid in the diagnosis of PE, Courtney et al 5 conducted a large prospective study of a heterogenous cohort of nearly 8,000 patients evaluated for PE. Of these patients, 7.2% were found to have a PE.

The group found that the odds ratio (OR) for pleuritic chest pain was 1.53, which is higher than the OR for both hemoptysis and tachycardia. Of note, pleuritic chest pain IS NOT, and hemoptysis and tachycardia ARE variables included in the Wells score.

Don’t forget thrombophilia and family history of PE

Consistent with known decision rules, the variables with the strongest associations with PE were a patient history of venous thromboembolism, unilateral lower extremity swelling, recent surgery, estrogen use, oxygen saturation less than 95%, active cancer, and patient history of thrombophilia. 5  Besides pleuritic chest pain, the authors found that the other two variables not included in clinical decision rules with useful ORs were a personal history of non-cancer related thrombophilia (OR 1.99) and a family history of PE (OR 1.51). Failing to take into account the personal and family history may lead to missing the diagnosis of PE.

Using the PERC rule inappropriately: Pregnancy

No clinical prediction rule has been validated for use in pregnant patients. Although the N numbers were small for pregnancy and postpartum status, they concluded that the PERC rule should not be used in isolation to rule out PE in patients who are either pregnant or postpartum.

Using the PERC rule inappropriately: High-prevalence PE populations

It is important to note that the PERC rule was never intended to be applied to anything but a low risk group of patients determined either by clinical gestalt or by the Wells PE score. 6 In fact, the meta-analysis 2  found some heterogeneity in the PERC rule sensitivity to exclude PE. Two studies from European populations with a prevalence of PE ranging from 21-30% found that a negative PERC rule combined with the low risk Revised Geneva Score only reduced the prevalence of PE in the studied patients to 6%. 7,8  Only in one of these studies did the PERC rule combined with clinical gestalt reduce the prevalence of PE down to nearly zero.8

The prevalence of PE in your community will determine the negative predictive value of the PERC rule where you are practicing. It is suggested that the PERC rule only be utilized where the prevalence of PE is <7%. 9 Most of the well-designed PE literature indicates that the PE prevalence in the U.S. is around 6%. 1

Using the PERC rule inappropriately: Abnormal vital signs

Although intended to be a triage tool, the Wells score doesn’t specify whether you should include an isolated tachycardic pulse rate. One study demonstrated that normalization of vital signs does NOT reduce the probability of a PE in ED patients. So we should use the most abnormal vital signs when risk stratifying patients for PE. 10 Furthermore, remember that beta blockers may mask tachycardia and can theoretically alter your Wells score or inclusion within the PERC rule. So for the PERC rule, hypoxemia or tachycardia at ANY point during the evaluation is a positive point.

Teaching Points

  1. Gestalt or some form of risk stratification should be employed first before using the PERC rule, which is reserved for low pretest probability cases.
  2. Because the negative LR of the PERC rule is 0.17, this allows you to have a maximum pretest probability of about 10.7% to apply the PERC rule to risk stratify your patient down to the standard risk of 2% (see below).
  3. The maximum suggested prevalence for PE in order to use the PERC rule is 7%.
  4. Pleuritic chest pain may increase the likelihood of PE more than some variables within existing decision rules.
  5. The PERC rule should not be used in isolation to rule out PE in pregnant or postpartum patients.
  6. The PERC rule includes hypoxemia or tachycardia at any point during the evaluation.

Bayes-PERC

1.
Kline J, Mitchell A, Kabrhel C, Richman P, Courtney D. Clinical criteria to prevent unnecessary diagnostic testing in emergency department patients with suspected pulmonary embolism. J Thromb Haemost. 2004;2(8):1247-1255. [PubMed]
2.
Singh B, Parsaik A, Agarwal D, Surana A, Mascarenhas S, Chandra S. Diagnostic accuracy of pulmonary embolism rule-out criteria: a systematic review and meta-analysis. Ann Emerg Med. 2012;59(6):517-20.e1-4. [PubMed]
3.
Pollack C, Schreiber D, Goldhaber S, et al. Clinical characteristics, management, and outcomes of patients diagnosed with acute pulmonary embolism in the emergency department: initial report of EMPEROR (Multicenter Emergency Medicine Pulmonary Embolism in the Real World Registry). J Am Coll Cardiol. 2011;57(6):700-706. [PubMed]
4.
Kline J, Slattery D, O’Neil B, et al. Clinical features of patients with pulmonary embolism and a negative PERC rule result. Ann Emerg Med. 2013;61(1):122-124. [PubMed]
5.
Courtney D, Kline J, Kabrhel C, et al. Clinical features from the history and physical examination that predict the presence or absence of pulmonary embolism in symptomatic emergency department patients: results of a prospective, multicenter study. Ann Emerg Med. 2010;55(4):307-315.e1. [PubMed]
6.
Bossart P. Misuse of the pulmonary embolism rule-out criteria. Ann Emerg Med. 2012;60(6):820; author reply 820. [PubMed]
7.
Hugli O, Righini M, Le G, et al. The pulmonary embolism rule-out criteria (PERC) rule does not safely exclude pulmonary embolism. J Thromb Haemost. 2011;9(2):300-304. [PubMed]
8.
Penaloza A, Verschuren F, Dambrine S, Zech F, Thys F, Roy P. Performance of the Pulmonary Embolism Rule-out Criteria (the PERC rule) combined with low clinical probability in high prevalence population. Thromb Res. 2012;129(5):e189-93. [PubMed]
9.
Rehnberg J, Vondy A. Towards evidence-based emergency medicine: Best BETs from the Manchester Royal Infirmary. BET 3: Pulmonary embolism rule-out criteria (PERC) for excluding pulmonary embolism. Emerg Med J. 2014;31(1):81-82. [PubMed]
10.
Kline J, Corredor D, Hogg M, Hernandez J, Jones A. Normalization of vital signs does not reduce the probability of acute pulmonary embolism in symptomatic emergency department patients. Acad Emerg Med. 2012;19(1):11-17. [PubMed]

Expert Peer Review

Q: What do you consider to be the limitations of the PERC rule?

Dr. Kline: Physicians have to use it correctly (meaning look at the vital signs and ask all the questions). Do not try to memorize the components. Look them up-- memorizing decision rules, PERC, NEXUS, PECARN, or any of them is not necessary. This is not like a breath holding contest or a hotdog eating contest. And PERC is not validated in pregnant patients.

Q: Is there any more data on the PERC negative patients who will actually end up getting a diagnosis of PE?

Dr. Kline: Not beyond the letter I published in Annals last year: Kline JA, et al. Clinical features of patients with pulmonary embolism and a negative PERC rule result. Ann Emerg Med. 2013 Jan;61(1):122-4. Pubmed

Q: What about using a modified PERC for pregnant or post-partum patients?

Dr. Kline: Trying to validate in pregnancy, and working on it.

Q: Are there any future directions you are taking the PERC rule, such as modifying it for certain populations? Or do you feel that the PERC rule has been validated enough (excluding pregnant and post-partum patients)?

Dr. Kline: We are trying to validate the PERC rule in pregnancy. The PERC rule has been validated enough. See the meta-analysis by Singh et al (Pubmed). Here are excerpts from that study:

\"We found that when the pretest probability is low, PERC are highly sensitive in predicting pulmonary embolism, and D-dimer testing is thus unnecessary. These findings are at a confidence of what is considered “level 2 evidence,” ie, demonstrated accuracy in either 1 large prospective study including a broad spectrum of patients and clinicians or validated in several smaller settings that differed from one another.\"

\"Our meta-analysis reports consistent high sensitivity and negative predictive value of PERC, with missed pulmonary embolism in just 0.5% of patients. Two of the included studies (Hugli et al and Righini et al) report a higher frequency of missed pulmonary embolism and have raised concern about the reliability of PERC. However, their higher failure rate likely results from the higher pulmonary embolism prevalence observed in their European settings. The threshold for pulmonary embolism diagnostic imaging in the United States is substantially lower than that in Europe, presumably because of the higher litigation risk. The PERC rule was developed for use in low-probability settings.\"

Jeffrey A. Kline, MD
Professor of Emergency Medicine and Vice Chair of Research, Department of Emergency Medicine; Indiana University School of Medicine
Jason West, MD

Jason West, MD

Senior Resident
Jacobi/Montefiore Emergency Medicine Residency Program
Albert Einstein School of Medicine
Jason West, MD

Latest posts by Jason West, MD (see all)

  • Guru

    I have always thought and believe that its good to understand the red flags but its the clinical gestalt of a fairly senior clinician that trumps all the scorings and rules. Its also interesting to note that there are studies to suggest that may be non significant PEs can be managed without any interventions

    • Jason West

      Although I agree with you in point, where is there good modern literature that suggests we don’t need to treat “non-significant” PEs? All of the evidence we have to not treat some PEs is based on the idea that our CT scanner is too good at detecting lung clots. We would have to do a prospective study on confirmed, untreated PEs to actually support not treating them. Instead, we could make the argument that we are scanning too many people; and the PERC may get you out scanning a patient.

      And if I can summarize your first point by saying that experience matters, I think resident education set up to work within that framework. Also, the PERC and the Wells score are supposed to aid, not replace, your gestalt.

      • Michelle Lin

        Interesting debate about “non significant PEs”. Only time will tell, but the preliminary evidence for subsegmental PEs seems to show that they ARE simiarly significant in terms of thrombotic recurrence rates and mortality. Salim wrote about the publication in 2013 issue of Blood.

        http://academiclifeinem.com/subsegmental-pulmonary-embolisms-sspe-important/

        • DD

          Absolutely. There is a misconception amongst our specialty that because a single occurrence of a disease may not itself result in harm, it doesn’t warrant a diagnosis. The truth is, those with pulmonary emboli, no matter how distal, warrant a hypercoag workup, ‘may’ benefit from therapy, and at a very minimum need to understand the symptoms and likely employ prophylactic measures or approach future life events (surgery, long airline flights, smoking, drug choice and more) that may place them at higher risk of thrombosis. The diagnosis is important.

          ‘Not harmful this time,’ and then trying to extrapolate treatment recommendations referencing “good modern literature” in which there are no recent RCT’s – is short sighted, potentially injurious to patients and is of the mindset contributing to our specialties’ lack of respect in the house of medicine.

          Just because someone says it at a CME show, or even suggests it in a time trend analysis, doesn’t mean it’s truth, and we shouldn’t allow it to spread across the EM ecosystem so relentlessly.

  • Natalie Desouza

    How many weeks postpartum counts as “postpartum status” ?

    • Michelle Lin

      Great question. In the original 2008 [reference #1 above], it was up to 4 weeks.

    • Jason West

      In the EM literature (like in reference 1, 4, and 5) postpartum status is defined as 4 weeks after delivery.

      However, there is probably no distinct cutoff for when the risk of VTE ends. There is a new NJEM paper (link below) reporting a retrospective review of 1.6 million women who delivered in California who were found to have a total of 720 VTEs. They used the patients as their own controls to determine her individual risk of VTE after delivery. While the odds ratio for VTE during the first 6 weeks after delivery was 12.1, the odds ratio for weeks 6-12 was 2.2. The odds ratio didn’t approximate zero until about 20 weeks post-partum. So, maybe we should consider postpartum status to be 6 weeks after delivery instead.
      http://www.ncbi.nlm.nih.gov/pubmed/24524551

      There is some OB/Gyn literature suggesting that we do studies using various kinds of modified Wells scores, but the existing literature is not very good. As you can see, there are some things in the works. In the future, we might also look to using a higher D-Dimer cutoff.

      • DD

        Even a low d-dimer isn’t of value in pregnant women. The test is not accurate in this population and can not be used as a screening test reliably. This was proven several years ago.

        • Jason West

          Thanks for your comments.

          I think the literature on the subject is rather poor, actually. So instead of saying we have proven the test to be unreliable, you could say that we don’t have enough evidence to support its use. It is an understudied topic.

          And if you we are going to be talking about D-dimer, you should know that not are created equal. Semi-quantative latex, Vidas, SimpliRed, and the automated quantitative tests all have different properties reported in the literature.

          Another good question is whether pregnancy is an isolated risk factor for PE? Not everyone will agree on this.

          • Andywebster

            This is a complex minefield of a subject and the stakes are high. When tests or decision rules are inappropriately applied then bad outcomes may occur.
            So when applying PERC are people using gestalt? or are they combining with a formal risk stratification?
            If using Wells in the three level rule I presume the pretest probability is only low enough in the low group. What is the pretest probability in the unlikely group in the 2 level Wells group? my quick attempt at a search this morning could not identify the source paper.
            However going to MDCalc, a score of <2 puts the incidence of PE at 1.3% and 3% in the unlikely (<4). This suggests that you could use the 2 level Wells and the PERC rule together (unlikely + PERT negative + sensible clinician=no d-dimer and no further PE work up)….but this I am not sure has been validated.

          • Jason West

            Thanks for your comments, Andy. This is a really nice discussion.

            If I may distill your thoughts, you are saying that the combination of the PERC rule and Wells score together could be used as a decision rule that doesn’t utilize D-dimer testing. I think this is a perfectly reasonable strategy. The max pre-test probability for D-dimer is a little higher (provided you have a good one), but the difference may not be wide enough for some clinicians to think it’s worth it. If you want more support for your idea, check out the ACEP clinical policy document on PE.

            I don’t think the PERC and Wells score-only model has been studied, and it would be really interesting to see.

            Another reason to prefer the PERC over D-dimer is reproducibility:

            Although D-dimer assays may be different at different institutions (potentially giving different results), the PERC rule will be the same anywhere as long as you apply it to a low prevalence population.

            You also bring up a good point about using the bi-level Wells score model instead of the 3 levels. To be honest, I don’t really use the Wells score to make decisions; I just make a decision about whether or not I think they probably do or do not have it…and order the dimer or CT accordingly if I can’t use the PERC to support my decision in the chart.

            Also, keep in mind that the maximum pretest probability I calculated for this post was based on accepting a 2% posterior probability. Some clinicians want the posterior probability to be lower than that; Some will want it to close to zero.

            Again, thanks.

  • Matt Kriese

    I know this post is 3 years old but I’m wondering what people think about saying someone “fails the perc rule” if they have a measured tachycardia at any point during their ED stay? The study referenced provided some information on the lack of utility of the normalization of vital signs for ruling out PE.

    However, it is my understanding that if we are to use a clinical rule, we need to use it the same way it was used in the validation study and from reading the perc rule validation study, it appears that they measured a single pulse rate at the time of enrollment. If someone presents to the ED with a normal pulse and then subsequently develops tachycardia, I don’t think the perc rule applies. Am I missing something?