From Wiki Journal Club
Jump to: navigation, search
Kline JA, et al. "Prospective multicenter evaluation of the pulmonary embolism rule-out criteria". J Thromb Haemost. 2008. 6(5):772-780.
PubMedFull textPDF

Clinical Question

In low-risk outpatients with suspected pulmonary embolism (PE), can the application of the pulmonary embolism rule-out criteria (PERC) predict venous thromboembolism (VTE) and all-cause mortality rates at 45 days of less than 2%?

Bottom Line

By combining low pretest probability for PE using gestalt and PERC-negative, the probability of VTE and all-cause mortality at 45 days is reduced to below 2%. This can be applied to approximately 20% of outpatients suspected for PE.

Major Points

Excessive investigation for pulmonary embolism in low risk patients has resulted in increased radiation exposure from computed tomography angiograms (CTA). This was a large prospective multi-center validation trial demonstrating that in patients who are determined low risk for PE by clinical gestalt and is PERC-negative, the probability of venous thromboembolism and all-cause mortality at 45 days was less than 2%. The authors chose 2% as the point of equipoise - the point at which the risk/benefit of investigating for PE is balanced with the risk/benefit of not investigating.


  • Multi-center, prospective, cohort study
  • N=8138 patients with suspected PE
  • Setting: Twelve emergency departments in the US and one in New Zealand
  • Enrolment: 2003-2006
  • Follow-up: 45 days
  • Primary outcome: DVT or PE or all-cause mortality at 45 days


Inclusion Criteria

  • Initiation of diagnostic testing for suspected PE (CTA, VQ lung scanning or a D-Dimer assay); DVT ultrasound did not trigger enrollment

Exclusion Criteria

  • Recent imaging diagnosis of PE within last 7 days
  • Patient preferred to follow-up at another hospital
  • High risk circumstances in which patient may be lost to followup (such as being homeless, severe psychiatric disorders, no reliable telephone number provided, international travelers, inmates, arrested for felonies)

Original PERC Study Exclusion Criteria

The original PERC study excluded patients with the following characteristics:

  • Shortness of breath is not the most important or equal most important presenting complaint
  • Active malignancy
  • Known thrombophilia
  • Strong family history of thrombosis
  • Concurrent beta-blocker use (that could blunt reflex tachycardia)
  • Transient tachycardia
  • Patients with amputations
  • Massively obese patients in which unilateral leg swelling cannot be assessed
  • Patients with baseline hypoxemia of less than 95% on pulse oximetry

Baseline Characteristics

  • N = 8138
  • Mean age: 49
  • Gender: 67% female
  • Race: Black (32%), Caucasian (59%), Latino or Hispanic descent (6%)
  • Symptoms
    • Pleuritic chest pain: 44%
    • Substernal chest pain: 34%
    • Dyspnea: 51%
    • Cough: 29%
    • Syncope: 6%
    • Hemoptysis: 3%
  • Comorbidities
    • Current smoker: 35%
    • Active malignancy: 15%
    • Immobility: 25%
    • Recent surgery: 7%
    • Pregnant or post-partum (<4 weeks): 10%
    • Prior PE or DVT: 11%
    • Congestive heart failure: 10%
    • Chronic obstructive pulmonary disease: 6%
    • Coronary artery disease: 13%
    • Taking warfarin: 7%
    • Hematologic thrombophilia: 5%
    • Connective tissue disease: 7%
    • Exogenous estrogen: 11%
  • Physical Findings
    • Highest pulse rate (beats per minute): 92
    • Highest respiratory rate (breaths/min): 21
    • Lowest SBP (mmHg): 131
    • Lowest pulse oximetry at room air (%): 96
    • Mean temperature (degrees celsius): 37
    • Wheezing: 10%
    • Unilateral leg swelling: 9%


  • This was a prospective non-interventional trial. Patients enrolled as per inclusion and exclusion criteria were followed through their emergency department and inpatient care as well as 45-day follow-up.

Applying PERC

PERC-negative requires the clinician to answer "no" for all eight of the following criteria:

  • Patient's age over 49?
  • Heart rate above 99 beats per minute?
  • Pulse oximetry reading less than 95% while on room air?
  • Present history of hemoptysis?
  • Patient taking exogenous estrogen?
  • Prior history of venous thromboembolism?
  • Recent history of surgery or trauma (requiring intubation or hospitalization within the last 4 weeks)?
  • Present unilateral leg swelling?

Online PERC calculator:


Primary Outcomes

VTE and all-cause mortality at 45 days was assessed in patients considered low risk for PE using clinical gestalt and were PERC-negative.
  • 1666 out of 8138 patients were considered low risk for PE and were PERC-negative.
  • 15 had VTE and 1 died out of the 1666.
  • False negative rate of 16/1666 = 1.0%
  • When combining low pretest probability and PERC-negative, the sensitivity is 97.4%


  • Clinical gestalt was required to determine pre-test probability.
  • Since this was not a management study, safety cannot be fully assessed.
    • Clinicians did not use PERC rule and low-risk assessment instead of diagnostic testing
  • Prevalence of PE in the study population was relatively low - 5.9% within 45 days. In centres with higher prevalence of PE in their population, this strategy may not be safe enough.


The study received government funding through the Grants from the National Institutes of Health. Additional funding included a Medical Student Award from the Emergency Medicine Foundation.

J. A. Kline owns stock in CP Diagnostics LLC and Studymaker LLC. C. L. Johnson owns stock in CPDiagnostics LLC and Studymaker LLC.

Further Reading

The original PERC derivation study: