Wells Score for PE Calculator — Pulmonary Embolism Pre-Test Probability

About

The Wells clinical prediction rule for pulmonary embolism, developed by Dr. Philip Wells and his research group at the Ottawa Hospital Research Institute, was first published in Thrombosis and Haemostasis in 2000 and subsequently validated in multiple large prospective cohort studies. It is one of the most extensively validated clinical decision rules in emergency medicine, with over 50,000 patients included in validation studies across diverse clinical settings. The rule incorporates 7 clinical variables: signs and symptoms of DVT (rated as objective leg swelling, pain, and tenderness on deep vein palpation), heart rate >100 bpm, immobilization for ≥3 days or surgery within 4 weeks, previous objectively diagnosed DVT or PE, hemoptysis, active malignancy (treatment within 6 months or palliative), and the clinician's judgment that PE is the most likely diagnosis (the gestalt item). The rule can be used as either a three-tier (low 0-4, moderate 4.5-6, high >6) or a two-tier scoring system (PE unlikely ≤4, PE likely >4). The two-tier system was introduced in 2001 and has become the preferred approach in most contemporary guidelines because it provides a clear threshold for D-dimer testing: PE-unlikely patients first receive D-dimer testing, while PE-likely patients proceed directly to imaging. The original derivation study reported a sensitivity of 100% and specificity of 44% for the two-tier system combined with D-dimer testing, and subsequent meta-analyses confirmed a negative predictive value exceeding 99% for PE-unlikely patients with a negative D-dimer. The Wells score has been compared prospectively with other clinical decision rules including the Revised Geneva Score, the PERC rule, and the YEARS algorithm, and demonstrates comparable sensitivity with slightly variable specificity across populations. The score's versatility — particularly the gestalt item — is both a strength (allowing clinician judgment to override algorithmic limitations) and a limitation (introducing inter-rater variability).

Formula

Clinical DVT Symptoms (+3) + PE is #1 Diagnosis (+3) + Heart Rate > 100 (+1.5) + Immobilization/Surgery (+1.5) + Previous DVT/PE (+1.5) + Hemoptysis (+1) + Malignancy (+1)

The Wells PE score uses a weighted scoring system with seven clinical variables, each assigned points proportional to its diagnostic weight. Clinical symptoms of DVT (objective leg swelling, pain, and tenderness on palpation) is assigned 3 points because it is the strongest single predictor of PE — approximately 70% of patients with documented PE have evidence of DVT on imaging. The clinician's judgment that PE is the most likely diagnosis (the "gestalt" or PE is #1 diagnosis item) is also assigned 3 points, reflecting the importance of overall clinical assessment. Heart rate >100 bpm is assigned 1.5 points because isolated tachycardia has moderate diagnostic value (likelihood ratio approximately 1.5-2.5). Immobilization or surgery within 4 weeks and prior DVT/PE are each assigned 1.5 points because they represent significant predisposing risk factors. Hemoptysis (1 point) and active malignancy (1 point) are less common but highly specific findings. The total score ranges from 0 to 12.5. The three-tier interpretation: low probability (0-4 points) corresponds to a PE prevalence of 5-10%; moderate probability (4.5-6 points) corresponds to 15-30% prevalence; high probability (>6 points) corresponds to 40-65% prevalence. The two-tier interpretation is commonly used: PE unlikely (≤4 points), PE likely (>4 points). In the PE-unlikely group (which includes both the low and moderate probability categories), a negative high-sensitivity D-dimer assay effectively excludes PE (negative predictive value 99-100%). In the PE-likely group (high probability), D-dimer has limited value because its sensitivity is insufficient to overcome the high pre-test probability, and direct imaging (CTPA or V/Q scan) is indicated. Alternative algorithms may also be used: the Revised Geneva Score (8 variables, no gestalt item) may be preferred in settings where clinician gestalt is less valued, and the YEARS algorithm (3 variables from Wells plus D-dimer threshold) reduces the need for CTPA by 14% compared to the traditional Wells approach.

Score Interpretation

The Wells score for PE is embedded in the diagnostic algorithms of major international guidelines, including the 2019 European Society of Cardiology (ESC) guidelines on acute pulmonary embolism, the 2016 AHA/ACC Scientific Statement on the Evaluation of Patients with Suspected Acute PE, and the 2020 NICE Guideline on Venous Thromboembolic Diseases (NG158). The integration of the Wells score into these guidelines reflects the shift from imaging-first strategies to clinical-probability-first strategies, which reduces CTPA utilization by approximately 30-40% without increasing the rate of missed PE diagnoses. This reduction in imaging is clinically important because CTPA involves ionizing radiation (approximately 7-15 mSv per study), contrast-induced nephropathy risk (incidence 3-5% in patients with CKD), and incidental findings requiring further investigation. The health economic impact is substantial: a strategy of Wells score plus D-dimer followed by selective CTPA costs approximately $1,000-1,500 per PE diagnosis, compared to $2,500-3,500 for a CTPA-first strategy. The Wells score also facilitates early anticoagulation decision-making, as patients in the PE-likely category are often started on empirical anticoagulation pending imaging. Observational studies suggest that this approach reduces time to therapeutic anticoagulation by 1-2 hours. In low-resource settings, the Wells score with D-dimer testing provides a practical alternative when CTPA is unavailable or cost-prohibitive. However, the score's performance varies across populations: it performs best in emergency department patients with cardiopulmonary symptoms, less well in hospitalized patients, and may underestimate PE probability in patients with prior PE who have residual thrombi. Alternative strategies include the YEARS algorithm (which incorporates three Wells items with a D-dimer threshold adjusted to 1,000 µg/L for patients with 0 YEARS items) and the PEGeD algorithm (which uses a tiered approach based on revised Geneva score and age-adjusted D-dimer). Both alternatives demonstrate similar safety profiles with further reductions in CTPA utilization.

Reference Ranges