PROPER

Published

Freund Y, et al. "Effect of the Pulmonary Embolism Rule-Out Criteria on Subsequent Thromboembolic Events Among Low-Risk Emergency Department Patients: The PROPER Randomized Clinical Trial". Journal of the American Medical Association. 2018. 319(6):559-566.
PubMed • Full text

Clinical Question

In adult patients being evaluated in the emergency department, if the pretest probability of of pulmonary embolism (PE) is low, does the use of the pulmonary embolism rule-out criteria (PERC) safely exclude the diagnosis of PE?

Bottom Line

Among adult patients evaluated in the emergency department who have low pretest probability of PE, a PERC score of zero safely excluded PE and was noninferior to usual care at 3 months.

Major Points

Diagnostic testing for PE is well established and utilizes clinicians' assessment of pretest probability, followed by D-dimer and computed tomographic pulmonary angiography (CTPA). Several scoring systems have been developed to help guide the use of CTPA. Among these is the 8-item pulmonary embolism rule-out criteria (PERC) rule which assigns one point for each of patient age ≥50 years, hemoptysis, recent trauma or surgery, prior VTE, exogenous estrogen use, pulse ≥100/min, arterial SpO2 ≤94%, or unilateral leg swelling. In the US, a PERC score of zero is associated with a low false-negative rate of <2%. Retrospective analyses suggested that the PERC rule may not perform adequately in Europe however. PROPER prospectively evaluated the use of the PERC rule versus usual care.

This cluster-randomized non-inferiority trial conducted in 14 French emergency departments enrolled nearly 2,000 patients in whom PE was considered very low probability. Centers were randomly assigned to a 6-month control period, and a 6-month intervention period, in either order, separated by 2 months as a washout. In the PERC group, patients with a PERC score of zero were discharged without subsequent testing. Patients in both groups were followed for 3 months by telephone interview and chart review. The primary endpoint was death from any cause or symptomatic VTE during the follow-up period among patients not diagnosed at study entry. The investigators reported that the PERC-based strategy was noninferior to usual care in excluding PE. In addition, the PERC strategy was associated with a benefit in terms of reduced CTPA use (by about 10%), ED length of stay (by about half an hour), and likelihood of initial admission into hospital (by about 3%).

The PE prevalence was lower than previous trials and thus the sample size calculation was inaccurate. Bias may have been introduced in the control periods that was not recorded and with 3% of the population lost to follow-up, the non-inferiority margin could have been lost. This trial demonstrated a cost-free strategy for decreasing use of CTPA and length of stay in the emergency department. Similar results were seen in the PERCEPIC study which was also conducted in Europe.

Guidelines

ESC Guidelines (2019)

The results…suggested safe exclusion of PE in patients with low clinical probability who, in addition, met all criteria of the PERC rule. However, the low overall prevalence of PE in these studies does not support the generalizability of the results.

Design

• Multicenter, non-inferiority, crossover cluster–randomized clinical trial
• N=1,916 patients evaluated in French emergency departments
  • PERC (n=962)
  • Usual care (n=954)
• Setting: 14 emergency departments in France
• Enrollment: 2015-2016
• Mean follow-up: 3 months
• Analysis: Intention-to-treat, 1.5% non-inferiority margin
• Primary Outcome: Failure of diagnostic criteria

Population

Inclusion Criteria

• New-onset presence or worsening of shortness of breath or chest pain
• Low clinical probability of PE estimated by the treating physician's gestalt as an expectation below 15%

Exclusion Criteria

• Obvious etiology to the acute presentation other than PE
• Acute severe presentation (hypotension, SpO2<90%, respiratory distress)
• Contraindication to CTPA (impaired renal function with an estimated creatinine clearance <30 mL/min; known allergy to intravenous radioopaque contrast)
• Pregnancy
• Inability to be followed up
• Receiving any anticoagulant therapy for other indication

Baseline Characteristics

From the PERC group.

• Demographics: Mean age 44 years, female 51%
• Comorbidities: Respiratory 3%, Heart failure 2%, Stroke 1%
• Emerg presentation: Chest pain 91%, Shortness of Breath 32%, Syncope 1%
• Physiologic parameters: mean heart rate 82, HR >100 beats/min 13%, mean respiratory rate 18, median SpO2 99, mean systolic blood pressure 136, mean temperature 36.7 degC
• Anthropomorphics: Weight
• PE Risks: exogenous estrogens 7%, DVT 4%, history of VTE 3%, surgery or trauma within 1 month 2%, hemoptysis 1%, active malignancy 1%
  • Simplified Revised Geneva Score: low risk 86%, intermediate risk 14%
  • Wells score: low 91%, Intermediate risk 8%, high 1%
  • PERC score: Zero 48%, positive 52%

Interventions

• 6 months PERC-based strategy; if PERC score positive, usual diagnostic criteria applied
  • PERC criteria:
    • SpO2 <94%
    • HR ≥100/min
    • Patient age ≥50 years
    • Unilateral leg swelling
    • Hemoptysis
    • Recent trauma or surgery
    • Prior VTE
    • Exogenous estrogen use
• 6 months usual care
  • if low clinician gestalt of PE then D-dimer
  • if D-dimer positive then CTPA

Outcomes

Comparisons are PERC vs. usual care.

Primary Outcomes

VTE at 3 months

3% vs. 3% (P=0.12)

Secondary Outcomes

CTPA

13% vs. 23% (difference 9.7%, P<0.001, NNT=10)

Length of ED stay, hour

min, median

4:36 vs. 5:14 (difference -00:36, P<0.001)

Hospital admission

13% vs. 16% (difference 3.3%, P=0.04, NNT=30)

Anticoagulation therapy initiated

2% vs. 3% (difference 1.3%, P=0.09)

Hospital readmission at 3 months

4% vs. 7% (difference 2.1%, P=0.051, NNT=48)

All-cause mortality at 3 months

0.3% vs. 0.2% (difference 0.1%, P>0.99)

Subgroup Analysis

PERC negative

48% vs. 38%

Tested with D-dimer

55% vs. 99%

D-dimer <0.5 mcg/mL

35% vs. 50%

PE diagnosed in emergency department

1.5% vs. 2.7%

Adverse Events

None observed

Criticisms

• PE prevalence was lower than previous trials
• CTPA was considered positive if it showed an isolated subsegmental PE and could over-estimate prevalence as these PEs could be left untreated
• Sample size estimation was inaccurate, diagnostic failure in control group was less than estimated
• Non-patient level randomization was utilized
• With 54 patients (2.8%) lost to follow-up, a few events occurring in this group could have changed outcomes. Worst-case scenario analysis led to rejection of non-inferiority

Funding

• Sponsor was Assistance Publique-Hôpitaux de Paris
• Grant from Programme Hospitalier de Recherche Clinique-PHRC 2014 (Ministère de la Santé, Paris, France)