What is the best approach to estimate the risk of pulmonary embolism (PE) in an adult patient with no significant past medical history?

How to Calculate the Risk of Pulmonary Embolism

Use the Wells score or revised Geneva score to calculate pretest probability of PE, as both demonstrate similar diagnostic performance with approximately 12% PE prevalence in the "unlikely" category and 30% in the "likely" category. 1

Primary Validated Clinical Decision Rules

Wells Score (Most Widely Validated)

The Wells score uses 7 clinical variables to stratify patients into risk categories 2:

Clinical Variables and Point Values:

• Clinical signs/symptoms of DVT: 3 points
• Alternative diagnosis less likely than PE: 3 points
• Heart rate >100 bpm: 1.5 points
• Immobilization ≥3 days or surgery in previous 4 weeks: 1.5 points
• Previous DVT/PE: 1.5 points
• Hemoptysis: 1 point
• Malignancy: 1 point

Risk Stratification:

• Low risk (Wells score <2): 3% PE prevalence 2
• Intermediate risk (Wells score 2-6): 13% PE prevalence 2
• High risk (Wells score >6): 36% PE prevalence 2

Alternative Two-Tier System:

• PE unlikely (Wells score ≤4): 12% PE prevalence 1
• PE likely (Wells score >4): 30% PE prevalence 1

When combined with negative D-dimer in low-risk patients, the negative predictive value reaches 99.5% 2, 1. However, a major limitation is the subjective variable "alternative diagnosis less likely than PE," which represents physician judgment override worth 3 points 2.

Revised Geneva Score (Fully Objective Alternative)

The Geneva score provides a completely objective assessment without subjective clinical judgment 1:

Risk Categories:

• Low risk: 7.9% PE prevalence 1
• Intermediate risk: Moderate PE prevalence
• High risk: 73.7% PE prevalence 1

Both Wells and Geneva scores demonstrate comparable diagnostic performance, making either acceptable for clinical use 1.

Additional Risk Stratification Tools

PERC Rule (Rule-Out Criteria)

Use PERC to identify patients whose PE likelihood is so low that no diagnostic workup is needed 1. All 8 criteria must be absent:

• Age <50 years
• Heart rate <100 bpm
• SaO2 ≥95%
• No hemoptysis
• No estrogen use
• No prior DVT/PE
• No unilateral leg swelling
• No surgery/trauma requiring hospitalization within 4 weeks

When all PERC criteria are met, PE prevalence is only 2%, potentially avoiding unnecessary testing 1.

Kline Rule (D-Dimer Safety Assessment)

The Kline rule identifies patients "unsafe" for D-dimer testing alone 2:

Unsafe patients (requiring imaging regardless of D-dimer) have shock index >1.0 OR age >50 years PLUS any of:

• Unexplained hypoxemia (SaO2 <95%, no prior lung disease)
• Unilateral leg swelling
• Recent major surgery
• Hemoptysis

Unsafe patients have 42.1% PE prevalence, while safe patients have only 13.3% prevalence 2.

Pisa Model (Mathematical Prediction)

The Pisa model uses regression coefficients for multiple variables 2:

Key Predictors with Highest Coefficients:

• Sudden onset dyspnea: coefficient 2.00, OR 7.38 2
• Acute cor pulmonale on ECG: coefficient 1.96, OR 7.11 2, 3
• Chest pain: coefficient 1.01, OR 2.74 2
• Age ≥75 years: coefficient 1.14, OR 3.11 2

Protective Factors (Decrease Risk):

• Fever ≥38°C: coefficient -1.47, OR 0.23 2
• Orthopnea: coefficient -1.51, OR 0.22 2
• Preexisting pulmonary disease: coefficient -0.89, OR 0.41 2

Calculate probability: sum all applicable coefficients plus constant (-3.43), then probability = [1 + exp(-sum)]^-1 2.

Gestalt Clinical Assessment

Unstructured clinical judgment based on training and experience remains validated 2:

PIOPED Study Risk Categories:

• Low risk (0-19% pretest probability): 9.2% actual PE rate 2
• Intermediate risk (20-79% pretest probability): 29.9% actual PE rate 2
• High risk (80-100% pretest probability): 67.8% actual PE rate 2

In the Kabrhel study of 7,940 patients, gestalt assessment categorized 68% as low risk (<15% probability) with 3% actual PE rate, 26% as intermediate risk (15-40% probability) with 10% actual PE rate, and 6% as high risk (>40% probability) with 33% actual PE rate 2. Gestalt performs comparably to Wells score but requires significant clinical experience 2.

Clinical Algorithm for Risk Assessment

1. Calculate Wells or Geneva score for all patients with suspected PE 1
2. Apply PERC rule if clinical suspicion is very low to potentially avoid all testing 1
3. For low-risk patients (Wells <2 or Geneva low): obtain D-dimer; if negative, PE excluded 2
4. For intermediate-risk patients (Wells 2-6): obtain D-dimer; if positive, proceed to CT pulmonary angiography 2
5. For high-risk patients (Wells >6 or Geneva high): proceed directly to imaging regardless of D-dimer 2
6. Consider Kline rule to identify patients requiring imaging despite negative D-dimer 2

Common Pitfalls to Avoid

Do not rely on clinical features alone without structured scoring, as PE presentation is highly variable and nonspecific 4, 5. Dyspnea was the only presenting symptom in 29% of confirmed PE cases, and many classical features are often absent 5.

Do not use ECG findings in isolation for diagnosis, as ECG cannot diagnose or exclude PE but serves primarily to exclude alternative diagnoses and assess severity 3. The S1Q3T3 pattern has limited sensitivity despite being a classic finding 3.

Do not skip risk stratification before ordering D-dimer, as D-dimer has poor specificity in moderate-to-high risk patients and should only be used to exclude PE in low-risk patients 2.

Do not assume normal vital signs exclude PE, as only 39% had tachypnea, 35% had hypoxia, and 33% had tachycardia in one large series 5.