Clinical Decision-Making

Wells Score for PE: Keep the Probability Rule and D-Dimer Pathway Together

Quick Medical Calculator Editorial Team · Reading label 3 September 2026 · 3 min read

Retrospective study sequence, not publication date

The Wells score estimates pretest probability in a patient already being assessed for suspected pulmonary embolism. It is not a population screening test, a diagnosis, or a severity score. Hemodynamic instability or another emergency requires immediate clinical management, not completion of an outpatient rule-out exercise.

Half Points Matter

The original weighted score has two legitimate classification systems:

Writing the three-level categories as "0-1, 2-6, and at least 7" loses possible totals: 1.5 is low and 6.5 is high. "PE unlikely" also does not mean "PE excluded." The original study combined the score with a particular D-dimer assay, and its residual event estimates were not zero. Original Wells PE model.

A Worked Decision, With the Assumptions Visible

Consider a fictional 70-year-old outpatient with suspected PE, pulse 110/min and a previous PE. Assume no other Wells findings, a reliable assessment, and suitability for the local two-level diagnostic pathway.

• Pulse over 100 contributes 1.5 points.
• Previous DVT or PE contributes 1.5 points.
• Total: 3, within the PE-unlikely branch.

NICE pairs a score of four or less with D-dimer testing; a score above four leads to imaging. The subjective item "an alternative diagnosis is less likely than PE" contributes three points when present, so it must be explicitly considered rather than silently omitted. NICE diagnostic pathway and score.

For an appropriate high-sensitivity assay reported in ng/mL FEU, an age-adjusted cutoff for someone over 50 is commonly age multiplied by ten. At age 70 this is 700 ng/mL FEU. A result of 650 would be below that cutoff in this eligible, non-high-probability example. The prospective ADJUST-PE study evaluated age adjustment within a structured clinical-probability pathway, not as a stand-alone blood test. ADJUST-PE primary study.

Now add clinical signs of DVT, worth three points. The total becomes 6, placing the patient in the PE-likely branch of the two-level system. The earlier negative-threshold argument no longer supplies the intended rule-out pathway. Use the selected system consistently; do not switch category definitions midway through assessment.

Three Checks Before Interpreting a Negative Result

1. Confirm the assay and units. FEU and DDU are not interchangeable labels. A value without its units and laboratory threshold cannot be safely inserted into this example.
2. Confirm population and protocol eligibility. Pregnancy needs a pregnancy-specific diagnostic strategy; do not raise the D-dimer cutoff arbitrarily because pregnancy increases D-dimer. Pregnancy-adapted YEARS was evaluated prospectively as its own pathway. Primary pregnancy-adapted YEARS study.
3. Confirm that the pathway is actually completed. When testing is delayed, NICE includes assessment for interim therapeutic anticoagulation. That is a clinical prescribing decision with bleeding-risk considerations, not an automatic calculator instruction. NICE recommendations.

What the Evidence Does Not Promise

In the CHRISTOPHER study, the combination of PE-unlikely classification and a normal D-dimer occurred in 1,057 of 3,306 patients, about 32%, not 85%. Among 1,028 such patients left untreated, five developed nonfatal VTE during follow-up, about 0.5%. The denominators describe different questions: how often imaging could be avoided and the observed follow-up event rate. Neither is a guarantee for an individual patient. CHRISTOPHER primary study.

This page addresses accurate use of Wells with testing. Comparing entire alternative PE algorithms is a separate question; mixing individual thresholds from several algorithms creates a pathway that was not the one validated.