Wells score interpretation for pulmonary embolism

What does a Wells score mean?

A Wells score for PE is a pretest probability: it estimates how likely pulmonary embolism is in a patient already suspected of having it, before any D-dimer or imaging result is known. Wells and colleagues derived the rule in Thrombosis and Haemostasis in 2000 by analyzing 40 clinical variables in a random 80% of a prospective cohort, then checked it in the remaining 20%. Seven items survived: clinical symptoms of deep vein thrombosis (DVT) 3 points, no alternative diagnosis as likely as PE 3, heart rate over 100 beats per minute 1.5, immobilization or surgery in the previous 4 weeks 1.5, previous DVT or PE 1.5, hemoptysis 1 and malignancy 1. The total therefore runs from 0 to 12.5 in half-point steps, although 0.5 and 12 can never occur. The same paper set two ways to read the total, a three-level scale and a two-level scale, so the meaning of a number depends on which scale your pathway uses.

Because the moderate band crosses the two-level cut-off, totals of 2 to 4 are moderate probability and PE unlikely, while 4.5 to 6 are moderate probability and PE likely. Say which scale you used whenever you report a result.

Three-level Wells score: low, moderate and high probability

On the three-level scale, a Wells score under 2 means low clinical probability of PE, 2 to 6 means moderate probability and over 6 means high probability. Wells and colleagues tested this scale prospectively in the Annals of Internal Medicine in 2001, in 930 consecutive patients with suspected PE at the emergency departments of four Canadian tertiary care hospitals. Pretest probability was low in 527 patients, moderate in 339 and high in 64, and PE was found in 1.3%, 16.2% and 37.5% of them respectively. A 2010 meta-analysis by Ceriani and colleagues, covering 29 studies of PE prediction rules with 31,215 patients, pooled the PE prevalence for the three-level Wells score at 6% for low, 23% for intermediate and 49% for high probability. The 2019 European Society of Cardiology (ESC) guideline gives rounded figures for three-level scores in general: about 10% of patients with low, 30% with moderate and 65% with high probability turn out to have PE.

The figures differ because the populations differ. Ceriani and colleagues concluded that the rules are not interchangeable and that the choice of rule and scale should follow local PE prevalence, whether patients are outpatients or inpatients, and the D-dimer assay in use.

Two-level Wells score: PE unlikely or PE likely

On the two-level scale, a Wells score of 4 or less means PE unlikely and a score over 4 means PE likely. Wells and colleagues designed this split in their 2000 paper so that PE unlikely patients with a negative D-dimer would have PE in under 2% of cases. In the derivation set, 7.8% of patients scoring 4 or less had PE, but only 2.2% (95% CI 1.0% to 4.0%) of those who also had a negative SimpliRED D-dimer, and 1.7% in the validation set; that combination covered 46% of the study patients. The three-level low band with a negative D-dimer had a similar PE rate, 1.5% in the derivation set, but covered only 29% of patients, which is why the two-level scale excludes PE in more people. Ceriani's meta-analysis pooled the PE prevalence for the two-level Wells score at 8% for PE unlikely and 34% for PE likely, and the 2019 ESC guideline expects about 12% and 30% for two-level scores.

Which test follows each Wells range?

The Wells range decides whether a D-dimer or imaging comes first. In the Christopher study, published in JAMA in 2006 by van Belle and colleagues, 3,306 consecutive patients with suspected PE at 12 Dutch centers were scored on the two-level Wells rule. PE unlikely patients had a D-dimer test, and PE was considered excluded when it was normal; every other patient had computed tomography (CT). PE unlikely with a normal D-dimer occurred in 1,057 patients (32.0%), and of the 1,028 left untreated, 5 (0.5%) had nonfatal venous thromboembolism within 3 months. CT showed PE in 674 patients (20.4%). The 2019 ESC guideline follows the same logic: a D-dimer, preferably with a highly sensitive assay, for patients with low or intermediate clinical probability or PE unlikely, and CT pulmonary angiography (CTPA) without a D-dimer at high clinical probability, because a normal D-dimer does not safely exclude PE in that group.

The ESC guideline restricts moderately sensitive and point-of-care D-dimer assays to low probability or PE unlikely patients. For patients over 50 it suggests an age-adjusted D-dimer cut-off as an alternative to the fixed one. In the emergency department, a patient whose gestalt probability is already low can be checked with the PERC rule calculator before any blood test. The earlier Wells 2001 study used only the low band this way: of 437 patients with low probability and a negative D-dimer, 1 developed PE in 3 months, a negative predictive value of 99.5%, and all others went to ventilation-perfusion scanning.

What each Wells total means, point by point

Because the Wells items are worth 3, 1.5 or 1 point, a Wells score for PE can be 0, 1, 1.5, any half-point from 2 to 11.5, or 12.5. Totals of 0, 1 and 1.5 are low probability and PE unlikely; a heart rate over 100 on its own gives 1.5. Totals of 2 to 4 are moderate probability but still PE unlikely, so a patient with clinical signs of DVT and hemoptysis scores 4 and has a D-dimer first on a two-level pathway. Totals of 4.5 to 6 are moderate probability and PE likely: the calculator's default example, a heart rate over 100 with no alternative diagnosis as likely as PE, gives 4.5 and points to CTPA. Totals of 6.5 to 12.5 are high probability, such as DVT signs plus no alternative diagnosis plus tachycardia, which makes 7.5. Half a point separates 4 from 4.5, so a single item can change the whole pathway.

The 3-point no alternative diagnosis item is a judgment, so two clinicians can land on opposite sides of 4. If you want a rule without it, see Wells vs revised Geneva score. To check a total, tick the items in the Wells score for PE calculator, which shows both scales at once.

How the Wells DVT score's ranges differ

The Wells score for DVT is a separate rule with its own items and cut-offs, so a number from one cannot be read on the other's scale. The DVT score adds 1 point for each clinical feature present and subtracts 2 when another diagnosis is at least as likely as DVT, so it uses whole numbers from minus 2 upward instead of half-points from 0 to 12.5; in the PE score the same judgment works the other way, adding 3 when no alternative is as likely. The 1997 three-level DVT model in the Lancet calls 0 or less low, 1 to 2 moderate and 3 or more high pretest probability, and in 593 outpatients DVT was found in 3%, 17% and 75% of these groups. The 2003 two-level model in the New England Journal of Medicine calls DVT likely at 2 or more and unlikely below 2. The imaging test also differs: compression ultrasound of the leg, not CTPA.

In the 2003 trial of 1,096 outpatients, a DVT unlikely patient with a negative D-dimer skipped ultrasound, and 39% of the D-dimer group needed no ultrasound at all. Score a suspected leg clot on the Wells score for DVT calculator.