Podcast: Play in new window

Subscribe: RSS

Reference: Ellison AM, et al. PERC-Peds rule for bedside exclusion of pulmonary embolism without radiation in children in the USA (BEEPER): a multicentre, prospective, observational, diagnostic accuracy study. Lancet Respir Med. July 2026.

Date: August 13, 2026

Dr. Lauren Westafer

Guest Skeptic: Dr. Lauren Westafer is an Associate Professor in the Department of Emergency Medicine at the University of Massachusetts Medical School, Baystate. She is the co-founder of FOAMcast and a researcher in pulmonary embolism and implementation science.  Dr. Westafer serves as the research methodology editor for Annals of Emergency Medicine.

Check out this FOAMcast episode on PERC-Peds as well!

Case: A 15-year-old previously healthy teenage girl presents to the emergency department (ED) with pleuritic chest pain and feeling short of breath after a recent viral illness. On your exam, she is comfortable, has a heart rate of 88 beats/min and oxygen saturation of 99% throughout the visit. She denies hemoptysis, unilateral leg swelling, recent surgery, estrogen exposure, or previous venous thromboembolism. You are working with a trainee who gives you a fantastic differential diagnosis for the chief complaint, which includes a pulmonary embolism. The trainee thinks pulmonary embolism is unlikely and asks, “I have heard of PERC and Wells in adults, but can we apply these to pediatric patients?”

 Background: Pulmonary embolism (PE) represents a broad spectrum of disease with clinical history and presenting features that overlap with many other disease processes. Historically, venous thromboembolism (VTE) has been thought to be quite uncommon in pediatric patients, absent malignancy or indwelling lines. The gold standard diagnostic modality for PE, CT pulmonary angiography (CTPA), carries additional risk to children due to ionizing radiation. In adult patients, the epidemiology of VTE, validated diagnostic algorithms, and D-dimer testing can safely exclude PE in many patients without the need for CTPA.

Previously, there were no prospectively validated bedside tools to identify children in whom we could feel comfortable safely stopping PE testing. Some clinicians extrapolated from adult Pulmonary Embolism Rule-Out Criteria (PERC). In adults, PERC can be used to exclude PE without need for additional blood work or imaging studies. The use and accuracy of D-dimer test to exclude PE in children is also unclear.

The new PERC-Peds rule is a modified version of PERC that was created using retrospective data. There are 8 criteria.

Many criteria are similar between the adult and pediatric PERC but there are some notable differences.

Obviously, for children, PERC-peds has no age<50 criterion. Also, because of differences in vital signs in pediatric patients, the heart rate is age-adjusted rather than using a fixed HR >100. Rather than relying only on unilateral leg swelling, PERC-Peds asks whether deep vein thrombosis (DVT) is clinically suspected.

There’s also a small change regarding the definition for surgery, which is a procedure requiring endotracheal intubation within the past 30 days.

Clinical Question: In children aged 4-17 years in whom pulmonary embolism is being tested for or strongly considered, can a negative PERC-Peds rule safely exclude PE or proximal DVT within 45 days without laboratory testing or pulmonary vascular imaging?

Reference: Ellison AM, et al. PERC-Peds rule for bedside exclusion of pulmonary embolism without radiation in children in the USA (BEEPER): a multicentre, prospective, observational, diagnostic accuracy study. Lancet Respir Med. July 2026.

  • Population: Children aged 4-17 years at 21 pediatric EDs in the United States who had PE testing ordered (included D-dimer, CTPA, ventilation perfusion scan, magnetic resonance angiography, or other pulmonary vascular imaging) or in whom the senior clinician strongly considered PE (this could be when a D-dimer or CTPA were ordered or cancelled or there was an open verbal discussion about the possibility of ordering one of those tests but decided against it).
    • Exclusion: Known pregnancy, current anticoagulation for VTE, intoxication, incarceration, inability to obtain consent/assent, and inability to complete 45-day contact
  • Intervention: PERC-Peds
  • Comparison: None
  • Outcome: Performance of the PERC-Peds rule
    • Primary: Safe exclusion of PE or proximal DVT, defined a priori as the upper bound of the 95% confidence interval for the false-negative rate not exceeding 1.5%.
    • Secondary: Sensitivity, specificity, predictive values, likelihood ratios, inter-rater reliability, D-dimer performance, and potential reduction in CTPA use
  • Trial: Multicentre, prospective, observational diagnostic accuracy study

Authors’ Conclusions: “In this multicentre, prospective, observational, diagnostic accuracy study of children with suspected pulmonary embolism in the emergency department, we found a 6·3% prevalence of pulmonary embolism or proximal DVT; in this population, the PERC-Peds negative rule can safely rule out pulmonary embolism. Use of PERC-Peds might reduce low-value diagnostic testing for pulmonary embolism in children and adolescents.”

Quality Checklist for A Diagnostic Study:

  1. The clinical problem is well defined. Yes
  2. The study population represents the target population that would normally be tested for the condition (ie no spectrum bias). Yes
  3. The study population included or focused on those in the ED. Yes
  4. The study patients were recruited consecutively (ie no selection bias). Unsure
  5. The diagnostic evaluation was sufficiently comprehensive and applied equally to all patients (ie no evidence of verification bias). No
  6. All diagnostic criteria were explicit, valid and reproducible (ie no incorporation bias). Yes
  7. The reference standard was appropriate (id no imperfect gold-standard bias). Yes
  8. All undiagnosed patients underwent sufficiently long and comprehensive follow-up (ie no double gold-standard bias) Yes
  9. The likelihood ratio(s) of the test(s) in question is presented or can be calculated from the information provided. Yes
  10. The precision of the measure of diagnostic performance is satisfactory. Yes
  11. Who funded the trial? National Institutes of Health.
  12. Did the authors declare conflicts of interest? At least one author, T Charles Casper, reported payments to his institution.

Results: From July 2020 through September 2024, 4,039 children were enrolled. The median age was 15 years (IQR 13-16), 64% were female, and 17% were younger than 12 years. A final VTE outcome was adjudicated for 4,011, and 3,988 had complete PERC-Peds data for the diagnostic accuracy analysis.

Among the adjudicated cohort, 254 children (6%) met the criterion standard:

    • 122 had PE only
    • 56 had PE plus DVT
    • 76 had isolated proximal DVT.

The false negative rate was 0.1% (95% CI 0.0-0.8). Below the threshold of 1.5% they had set at the beginning of the trial.

PERC-Peds was negative in 734 of 3,988 children (18.4%). But there was one false negative.

One child with a negative rule was adjudicated as having PE. This was a 16-year-old female with cough, recent positive SARS-CoV-2 test, chest pain, dyspnea. But she also had normal vital signs and a normal D-dimer test. She had a CTPA ordered that had a filling defect in a peripheral pulmonary arterial branch of the left lower lobe.

Key Results: The PERC-Peds rule had high sensitivity and a very low false-negative rate, but low specificity. Apply it carefully as indiscriminate use could increase downstream testing and potentially unnecessary imaging.

Verification Bias:

This is sometimes called work-up bias and can occur when only some participants receive the definitive reference-standard test (often those who appear higher risk) while the remainder are classified using a less intensive method. This can make a diagnostic test appear more accurate because disease is more likely to be detected in test-positive or clinically concerning patients, while disease in lower-risk, untested patients may be missed.

In this study, pulmonary vascular imaging was performed in only around a third (36.6%) of participants, and all diagnostic testing was left to the treating clinician. Children with concerning symptoms, abnormal vital signs, or other risk factors were more likely to undergo imaging, whereas lower-risk or PERC-Peds-negative children were considered disease-free based on medical-record review and 45-day follow-up rather than direct imaging. An occult PE that caused no return visit or diagnosis during follow-up could therefore have been missed, potentially inflating sensitivity and lowering the observed false-negative rate.

The authors did take meaningful steps to reduce this bias. The PERC-Peds variables were collected prospectively but were not used as a mandated clinical management pathway, all adjudicated participants had complete 45-day follow-up, and an independent committee reviewed imaging, follow-up responses, and outside records. This makes the study stronger than one in which un-imaged patients simply were assumed not to have PE, but follow-up is still not equivalent to applying the same definitive imaging reference standard to everyone. Given the harms of imaging thousands of children solely for research, this was a pragmatic and ethically reasonable compromise, but some residual partial verification bias remains.

PE and Proximal DVT:

The study’s criterion standard was broader than pulmonary embolism alone. A patient was considered positive if they developed either imaging-confirmed PE or proximal DVT within 45 days. Of the 254 patients classified as having venous thromboembolism, 76 (~30%) had proximal DVT without a verified PE. Among those 76, 42 had a negative CTPA and 33 never underwent pulmonary vascular imaging.

Is it appropriate to lump these two things together? PE and proximal DVT are related manifestations of venous thromboembolism, but they are not the same clinical diagnosis. A child with a proximal DVT is at risk of embolization and will often receive anticoagulation, but that does not establish that a clot was present in the pulmonary vasculature at the time of evaluation.

This was an intentionally conservative definition. The authors argued that most PEs originate from proximal DVT, that some clinicians may diagnose DVT by ultrasound and begin anticoagulation without performing CTPA, and that CTPA itself is an imperfect reference test. Those are reasonable justifications, particularly when the goal is to avoid missing clinically important venous thromboembolism. However, including isolated DVT could inflate the apparent prevalence of the target condition and penalize a rule intended specifically to exclude PE. Reassuringly, when the authors repeated the analysis treating isolated DVT as negative, the results barely changed: sensitivity was 99.4% rather than 99.6%, and the false-negative rate remained 0.1%.

Clinical Gestalt:

Gestalt is the clinician’s overall impression or “gut feeling” after considering the history, examination, vital signs, risk factors, and whatever else makes them worried, but it is not calculated from a standardized formula. There’s the funny quip from the movie High Fidelity, “I’ve been thinking with my guts since I was fourteen years old, and frankly speaking, between you and me, I have come to the conclusion that my guts have s** for brains.”

The literature suggests that physicians are good at assigning people to broad risk categories for PE. This should not be conflated with the frequent adage “better or as good as clinical decision tools.” If a patient has a Wells Score <2 that is very very low risk of PE and definitely <15%. It is not to be used in pediatric patients, but you can do the same thing and think about the major risk factors, signs/symptoms and clinical assessment.

One of the trickiest parts of PERC-Peds is the requirement that the clinician’s gestalt pretest probability of PE be less than 15%. It is difficult to know what “15%” feels like at the bedside, and the distinction between 14% and 16% creates an artificial precision that probably does not exist. One clinician’s 10% suspicion might easily be another clinician’s 20% or 30%.

The authors appropriately explored how changing gestalt affected the rule. Raising the threshold to less than 20% allowed more children to be classified as low risk and improved specificity to 23.7%, but sensitivity fell to 98.0% and the false-negative rate rose to 0.6%; its upper 95% confidence limit of 1.3% still remained below the prespecified safety threshold.

Removing gestalt entirely improved specificity to 28.1%, but sensitivity dropped to 91.3% and the false-negative rate increased to 2.1%, with an upper confidence limit of 3.1%. Gestalt still seems to be playing a role here, and we shouldn’t be applying this rule indiscriminately.

Interobserver Agreement and Reproducibility:

A clinical decision tool is only useful if different clinicians apply it in the same way. In a convenience sample of 430 patients, overall agreement on the final PERC-Peds classification was 83%, but the kappa was only 0.32, suggesting fair agreement beyond chance. The weakest component was clinician gestalt below 15%, with 69% agreement and a kappa of 0.38.

Even criteria that seem more objective, like the heart-rate criterion, showed only moderate reproducibility, with 74% agreement and a kappa of 0.47. Some of this variation may depend on when the vital sign assessment is performed during an ED visit and on whether transient abnormalities are counted.

Other components performed better but were not perfectly reproducible. Agreement was 90% for oxygen saturation above 94% and 94% for suspected DVT, with Kappas of 0.65 and 0.64, respectively. Prior VTE, estrogen use, and recent surgery or trauma had stronger agreement, likely because they are more concrete historical variables.

The authors appropriately note that the second assessment could occur later in the visit and that observers might not have had identical information or examined the patient under the same circumstances.

The performance of this rule may degrade as it is implemented more widely, as clinicians have different training, documentation practices, and interpretations of subjective elements. A rule with excellent reported sensitivity may not reproduce that performance unless clinicians consistently apply each criterion.

D-dimer:

This study provides some of the first prospective data on D-dimer performance in children being evaluated for PE. Using a standardized cutoff of 500 ng/mL fibrinogen-equivalent units, D-dimer had a sensitivity of 89.8%, specificity of 61.3%, and a false-negative rate of 1.09%. That sensitivity is lower than most clinicians would want from a stand-alone rule-out test for a potentially serious diagnosis, and the upper confidence limit for the false-negative rate reached 1.7%.

The 1.7% upper bound of the 95% CI falls within the upper limit of the failure rate in some international recommendations . In other words, a negative D-dimer alone should not automatically end the evaluation in every child with suspected PE. The same applies to adults. You need to include probability assessment WITH D-dimer. It’s why we use YEARS or Wells with D-dimer. The more promising finding was its use after PERC-Peds. A sequential strategy in which PERC-Peds-negative children received no further testing and PERC-Peds-positive children then underwent D-dimer testing could have ruled out PE or proximal DVT in 54.3% of the cohort, with a false-negative rate of 0.9%. The authors estimated that this approach might have reduced CTPA use by about 20%.

However, D-dimer testing was not protocolized. Clinicians chose whom to test, several different assays were used, and many patients underwent imaging despite a negative D-dimer or without having a D-dimer first. These results therefore support D-dimer as a potential second step in a structured pediatric PE pathway, not as a fully validated stand-alone.

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We agree that a negative PERC-Peds rule met the study’s prespecified safety threshold in this selected cohort for detecting PE or proximal DVT. The phrase “can safely rule out pulmonary embolism” should still be interpreted within the study limitations. This was a prospective diagnostic accuracy study, not an impact analysis. The variable inter-rater reliability and number of missed or nonconsenting eligible patients raise questions about how consistently reproducible the results will be after implementation.

SGEM Bottom Line: PERC-Peds is a promising and likely clinically useful rule-out tool for carefully selected children; however, it is a one-way tool (rule out only) and cannot be used as an overall screen to indicate the need for PE evaluation.

Case Resolution: You and the trainee estimate the patient’s pretest probability of PE to be pretty low, less than 15%. She has no previous VTE, recent surgery, estrogen use, hemoptysis, tachycardia, hypoxemia, or suspected DVT. Her PERC-Peds rule is negative. After discussing the very low residual risk and the potential harms of unnecessary testing, you do not order a D-dimer or CTPA. You evaluate other causes of chest pain, provide clear return precautions, and arrange follow-up if symptoms persist.

Clinical Application: For an ambulatory child aged 4-17 years in whom PE is being considered, first estimate clinical pretest probability. If gestalt is below 15% and every objective PERC-Peds criterion is negative, stopping the PE workup without D-dimer or imaging is reasonable based on this study.

A positive PERC-Peds result does not diagnose PE and does not automatically mandate CTPA. This study suggests that D-dimer may be the next step for many patients, but the sequential strategy was not prospectively mandated or tested as a management algorithm. Use additional caution in younger children, non-academic or rural settings, hospitalized patients, and populations excluded from the study.

What Do I Tell the Trainee? You are right to consider PE, especially as it may be more common in older children than previously thought. A recent study modified the PERC criteria for pediatric patients. Let’s go over the components. For this patient, we think pretest probability is below 15%, and she meets all the remaining PERC-Peds criteria, so the rule would support stopping the PE workup without a D-dimer or imaging. If she were PERC-Peds positive, a D-dimer may be a reasonable next step, but the study does not support using D-dimer alone as a universal rule-out test in children.

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.

Other SGEM Episodes on VTE:

  • SGEM#51: Home (Discharging Patients with Acute Pulmonary Emboli Home from the Emergency Department)
  • SGEM#118: I Hope you Had a Negative D-dimer (ADJUST PE Study)
  • SGEM#126: Take me to the Rivaroxaban – Outpatient treatment of VTE
  • SGEM#163: Shuffle off to Buffalo to Talk Thrombolysis for Acute Pulmonary Embolism
  • SGEM#184:We Weren’t Born to Follow-Up – The PEITHO Long-Term Follow-up Study
  • SGEM#219: Shout, Shout, PERC Rule Them Out
  • SGEM#276: FOCUS on PE in Patients with Abnormal Vital Signs
  • SGEM#277: In the Pregnant YEARS – Diagnosing Pulmonary Embolism
  • SGEM#282:It’s All ‘bout that Bayes, ‘Bout that Bayes- No Trouble – In Diagnosing Pulmonary Embolism
  • SGEM#295: Teacher, Teacher – Tell Me How to Do It (Diagnose a PE)
  • SGEM#323: Mama I’m Comin’ Home – For Outpatient Treatment of a Pulmonary Embolism
  • SGEM#416: She’s Always A Woman, Query PE?
  • SGEM#469: You Take My Breath Away – D-dimer for Ruling Out PE in High-Risk Patients