4PEPS

The 4-Level Pulmonary Embolism Clinical Probability Score (4PEPS) sums 13 clinical items (points −2 to +5) for patients already being evaluated for suspected PE in the emergency department. Dyspnea or chest pain alone are not scored because the cohort was restricted to suspected PE presentations. Educational tool—follow local PE pathways and the original publication (Roy et al., JAMA Cardiol 2021).

Age (years)

Scoring uses three age bands from the final model; ages ≥65 are the reference category (0 points).

<50 (−2)50–64 (−1)≥65 (0)

Sex

Female (0)Male (+2)

Items that lower probability

Chronic respiratory disease (−1)Heart rate <80 beats/min (−1)

Items that raise probability

Chest pain and acute dyspnea both present (+1)Hormonal estrogenic treatment (+2)Personal history of VTE (+2)Syncope (+2)Immobility within last 4 weeks: surgery, lower-limb plaster cast, or bedridden >3 days for an acute medical condition (+2)Room air pulse oximetry SpO₂ <95% (+3)Calf pain and/or unilateral lower limb edema (spontaneous leg pain, pain on deep vein palpation, or swelling) (+3)PE is the most likely diagnosis (+5)

Optional: numeric age (moderate band helper)

If the result is moderate (6–12), entering age shows the D-dimer rule-out threshold (<0.5 μg/mL if <50, else <age × 0.01 μg/mL). Leave blank if you only need the total score.

Disclaimer: 4PEPS supports risk stratification in suspected PE; it does not replace clinical judgment, shared decision-making, or institutional protocols. D-dimer units and assays vary—align thresholds with your assay and guideline. Original validation was in emergency department cohorts with suspected PE.

4PEPS scoring explained

What is 4PEPS?

The 4-Level Pulmonary Embolism Clinical Probability Score (4PEPS) is a clinical pretest probability tool derived from merged emergency department cohorts of patients already evaluated for suspected pulmonary embolism (PE). It combines positive and negative predictors so that a single total maps to one of four probability bands, each tied to a specific D-dimer or imaging strategy in the original publication.

Important: Isolated dyspnea or chest pain are not separate score items because study patients already had suspected PE; the score adds both symptoms together (+1) when chest pain and acute dyspnea coexist.

Item points (Table 3, Roy et al. 2021)

Variable	Points
Age <50 years	−2
Age 50–64 years	−1
Age ≥65 years (reference)	0
Chronic respiratory disease	−1
Heart rate <80 beats/min	−1
Chest pain and acute dyspnea	+1
Male sex	+2
Hormonal estrogenic treatment	+2
Personal history of VTE	+2
Syncope	+2
Immobility within last 4 weeks (surgery, lower-limb cast, or bed >3 days for acute illness)	+2
Room air SpO₂ <95%	+3
Calf pain and/or unilateral lower limb edema	+3
PE is the most likely diagnosis	+5

Four clinical probability levels and strategy

Total <0 — Very low CPP

In the 4PEPS strategy, PE may be ruled out without D-dimer or imaging (clinical criteria only), with safety demonstrated in the authors' validation analyses.

Total 0–5 — Low CPP

Rule out PE if high-sensitivity D-dimer <1.0 μg/mL (units as in the primary paper; confirm FEU vs DDU locally).

Total 6–12 — Moderate CPP

Rule out if D-dimer <0.5 μg/mL (age <50) or <(age × 0.01) μg/mL (age ≥50), matching the age-adjusted convention described with ADJUST-PE in the article.

Total ≥13 — High CPP

Do not rely on D-dimer to rule out; proceed to imaging (CTPA or V/Q) per protocol.

Setting and data limitations

• Derived and validated in emergency department patients with suspected PE—not established for inpatients.
• Patients with missing variables were excluded in the primary analysis; sensitivity analyses treated missing items as lowest risk.
• The strategy should be applied as a whole; mixing ad hoc thresholds from other scores may be unsafe.

Source: Roy P-M, Friou E, Germeau B, et al. Derivation and Validation of a 4-Level Clinical Pretest Probability Score for Suspected Pulmonary Embolism to Safely Decrease Imaging Testing. JAMA Cardiol. 2021;6(6):705-713. doi:10.1001/jamacardio.2021.0064. Open access. Official calculator referenced in the paper: peps.shinyapps.io/PEPS.

Overview

Pulmonary embolism (PE) remains a common and potentially life-threatening diagnosis in emergency care, yet its presenting features overlap broadly with more benign cardiopulmonary conditions. As a result, clinicians often face tension between missing PE and overusing CT pulmonary angiography (CTPA) and other imaging. Clinical prediction rules that estimate pretest probability help structure this decision: they align the intensity of testing (for example, high-sensitivity D-dimer versus immediate imaging) with the patient’s baseline risk.

The 4-Level Pulmonary Embolism Clinical Probability Score (4PEPS) is one such rule. It was derived to produce a single additive score that places a patient into one of four pretest probability categories. Each category was linked in the original work to a specific next step in the diagnostic pathway—ranging from clinical rule-out without laboratory testing to direct imaging—so that D-dimer cutoffs and imaging use could be harmonized in one framework rather than mixing incompatible rules.

What problem does 4PEPS address?

Many strategies already exist to reduce unnecessary testing (for example, gestalt, Wells, revised Geneva, PERC, YEARS, age-adjusted D-dimer approaches, and others). In practice, combining rules that were designed separately can increase complexity and the risk of inconsistent application. 4PEPS was developed as an attempt to integrate concepts from prior approaches into one explicit score with four discrete probability levels, each mapped to a prespecified testing strategy in the primary publication.

The score is intended for outpatients in the emergency department setting who are already under evaluation for suspected PE. It is not a screening tool for the general population; it assumes the clinician has already decided that PE is in the differential.

Structure of the score

4PEPS includes 13 clinical variables. Some items subtract points (features associated with lower PE probability in the multivariable model), and others add points. The algebraic sum is the total 4PEPS value. There is no separate weighting step beyond the point assignment in the published model.

Notably, because study participants were enrolled on the basis of suspected PE, isolated dyspnea or isolated chest pain were not entered as standalone predictors. Instead, the score awards points when both chest pain and acute dyspnea are present together, reflecting how the variable was defined in the derivation analysis.

Item-by-item scoring (primary model)

The following table summarizes the published point assignments. When using the calculator, ensure each criterion is applied exactly as defined in the source study (timing windows, definitions of immobility, room air oximetry, and “PE most likely diagnosis”).

Variable	Points
Age <50 years	−2
Age 50–64 years	−1
Age ≥65 years (reference category)	0
Chronic respiratory disease	−1
Heart rate <80 beats per minute	−1
Chest pain and acute dyspnea	+1
Male sex	+2
Hormonal estrogenic treatment	+2
Personal history of venous thromboembolism (VTE)	+2
Syncope	+2
Immobility within the last 4 weeks (surgery, lower-limb plaster cast, or bedridden more than 3 days for an acute medical condition)	+2
Room air pulse oximetry SpO₂ <95%	+3
Calf pain and/or unilateral lower limb edema (including spontaneous leg pain, pain on deep vein palpation, or swelling)	+3
PE is the most likely diagnosis	+5

Four probability bands and the linked testing strategy

In the primary work, total score cutoffs partition patients into four bands. The bands were defined with an explicit goal of aligning pretest probability with D-dimer negative likelihood ratios and a prespecified safety mindset around post-test risk. The operational “4PEPS strategy” described alongside the score ties each band to a next diagnostic action.

Very low clinical probability (total score <0): In the published strategy, PE may be excluded using clinical criteria alone, without D-dimer or imaging, when this category applies—subject to the same inclusion assumptions as the validation cohorts.
Low clinical probability (total score 0–5): The pathway uses a high-sensitivity D-dimer with a higher numerical cutoff than the traditional 0.5 µg/mL rule in many systems: PE may be ruled out if D-dimer is below 1.0 µg/mL in the units and assay context described in the primary publication.
Moderate clinical probability (total score 6–12): The strategy uses a lower D-dimer threshold. For patients younger than 50 years, rule-out is considered if D-dimer is below 0.5 µg/mL. For patients 50 years and older, rule-out uses an age-adjusted cutoff of age × 0.01 µg/mL, consistent with the age-adjustment convention discussed in relation to prior PE management trials in the same article family.
High clinical probability (total score ≥13): PE is not considered safely excluded by D-dimer alone in this framework; imaging (typically CTPA or, when appropriate, ventilation–perfusion scanning) is indicated as the next step without using D-dimer to rule out PE first.

These thresholds assume high-sensitivity D-dimer assays and consistent reporting units. Laboratories differ in units (for example, D-dimer reported as FEU versus DDU) and in local validation; any application should be reconciled with institutional laboratory standards and current specialty society guidance, not inferred from the score alone.

Performance and safety considerations from the derivation work

The score was derived using merged prospective emergency department databases and evaluated in internal and external validation samples with materially different PE prevalences. Discrimination, summarized by the area under the receiver operating characteristic curve, was in a range broadly comparable to other well-known PE clinical models in those analyses. The authors also reported operational consequences of applying the full 4PEPS strategy retrospectively, including reductions in imaging utilization compared with several comparator strategies, alongside false-negative rates that remained within prespecified safety bounds in their reported analyses.

Interpreting those numeric results requires caution: they describe cohort-level performance under retrospective application of a protocol, not guaranteed performance for an individual patient in a new setting. Hemodynamic instability, pregnancy, renal failure, anticoagulation, recent surgery, and other factors may alter pretest probability and acceptable risk tolerance even when the arithmetic total falls into a lower band.

Missing data and practical use

In the primary statistical approach, patients with missing predictor data were excluded from the main analyses; supplementary sensitivity approaches treated missing items as the lowest-risk contribution. In real time, missing vitals or oximetry should be obtained when feasible rather than assumed negative. Electronic calculators and structured documentation can reduce omission errors for a 13-item instrument.

How 4PEPS differs from other PE rules

Compared with binary rules (for example, PERC) or three-level scores (for example, some Wells implementations), 4PEPS emphasizes four explicit tiers and pairs them with graduated D-dimer logic. It also incorporates negative weights for features that reduced PE likelihood in the multivariable model—an design choice that can pull totals downward in patients whose presentations are dominated by alternative explanations, which is one mechanism by which the score can identify a very-low-probability group eligible for no testing in the published pathway.

Limitations and scope

Setting: Evidence chiefly reflects emergency department cohorts with suspected PE; extrapolation to inpatients, perioperative wards, or primary care is not established by the same body of work.
Subjective item: “PE is the most likely diagnosis” is clinician judgment–dependent, similar to analogous items in other scores; inconsistent use can shift totals substantially because of the large point weight.
Assay dependence: D-dimer interpretation is not interchangeable across assays and units; local validation matters.
Implementation science: Retrospective validation of a strategy is not identical to prospective demonstration that workflow adoption reproduces safety and efficiency gains.

Using this calculator responsibly

Use 4PEPS as structured support alongside bedside assessment, electrocardiography when indicated, risk discussion, and institutional algorithms. When the score suggests deferring imaging, document the rationale, ensure appropriate follow-up when symptoms evolve, and revisit the assessment if new findings appear. When the score indicates high probability, coordinate timely imaging and consider concurrent stabilization measures if the patient is unstable.

CalcMD.

4PEPS — 4-Level Pulmonary Embolism Clinical Probability Score

Learn how 4PEPS estimates PE pretest probability in the ED using 13 weighted clinical items and four bands linked to D-dimer and imaging strategy. Educational overview of the Roy et al. clinical prediction model.