Wells' Criteria

Wells’ Criteria

Reference: Wells, P. S., Anderson, D. R., Rodger, M., Stiell, I., Dreyer, J. F., Barnes, D., ... & Kovacs, M. J. (2001). Excluding pulmonary embolism at the bedside without diagnostic imaging: management of patients with suspected pulmonary embolism presenting to the emergency department by using a simple clinical model and d-dimer. Annals of internal medicine, 135(2), 98-107.

Clinical Question: Can the Wells’ Criteria and negative D-dimer be used to safely manage patients with suspected pulmonary embolism (PE)?

PICOT:

  • Population: Adult patients from 4 hospitals throughout Canada from 1998-1999 with suspicion of PE with symptoms for less than 30 days AND acute new or worsening shortness of breath OR chest pain.

    • Exclusion Criteria: 1) suspected deep venous thrombosis of the upper extremity as a likely source of pulmonary embolism, 2) no symptoms of pulmonary embolism within 3 days of presenta=on, 3) an=coagulant therapy for more than 24 hours, 4) expected survival =me less than 3 months, 5) contraindica=on to contrast media, 6) pregnancy, 7) geographic inaccessibility precluding follow-up, or 8) age younger than 18 years.

  • Intervention: Diagnostic algorithm + D-dimer

  • Comparison: Low pretest probability (score < 2) versus moderate and high pretest probability (score 2<)

  • Outcome: The proportion of patients with venous thromboembolism (VTE) within 3 months after PE was excluded by low probability score and negative D-dimer

  • Type of Study: Prospective cohort study

Results: One patient stratified to the low-risk and negative D-dimer group in which imaging was not obtained had a VTE during the 3-month follow-up period.

  • The negative predictive value of a negative D-dimer in the low probability group (score < 2) was 99.5% (95% CI 98.4-99.9%)

  • The negative predictive value of a negative D-dimer in the entire population (any score) was 97.3% (95% CI 95.8-98.4%)

Author Conclusion: “Managing patients for suspected pulmonary embolism on the basis of pretest probability and D-dimer result is safe and decreases the need for diagnostic imaging.”

Hot take(s):

  • The prevalence of PE in this population was low. As technology has advanced and computerized tomography (CT) images identify more PEs, the prevalence of PE has increased. Technically the pretest probability changes with prevalence, should we be taking this into consideration?

  • The ‘PE is #1 diagnosis OR equally likely’ component of the Wells’ criteria is subjective and requires gestalt. At what level of training can one determine if PE is the #1 diagnosis or equally likely?

Take Away: Considering the more recent validations studies and introduction of age-adjusted d-dimer, use of the three or two-tiered Wells’ Criteria model can be used to manage patients with a low pretest probability for PE and a negative age-adjusted d-dimer without CT imaging.

Evidence Based Medicine

Pre-test Probability: The probability of a patient having a disease. Pre-test probability multiplied by likelihood ratio gives us the post-test probability. Since probability cannot be divided or multiplied, we need to convert to odds to utilize a likelihood ratio. Pre-test probability can be determined by:

  • Clinician experience

  • Prevalence of the disease

  • Clinical decision rules

Example: If the prevalence of disease X among a population is 25%, the pre-test probability of the disease is 0.25. From this we can convert to pre-test odds and calculate post-test probability:

  • Pre-test probability = 0.25

  • Pre-test odds = 0.25 ÷ (1 - 0.25) = 0.25 ÷ 0.75 = 0.33

  • Pretend a diagnostic test has a positive likelihood ratio of 10

  • Post-test odds = 0.33 x 10 = 3.3

  • Post-test probability = 3.3 ÷ (3.3 + 1) = 0.76

  • If the test result was positive, the probability of the patient having the disease went from 0.25 (25%) to 0.76 (76%) which should substantially increase the disease on your differential diagnosis or in some cases may be enough to make the diagnosis

Summary: You can use pretest probability coupled with the likelihood ratio of a diagnostic text (i.e. labs, imaging) to increase or decrease your posttest probability which can help raise or lower your suspicion for a disease process in your differential diagnosis.