Background: How accurate are clinical prediction rules when combined with D-dimer in the primary care setting?
Systematic review followed by a validation study
Five rules studied: original Wells, modified Wells, simplified Wells, revised Geneva, simplified revised Geneva
n = 600 patients with suspected PE in 300 Dutch practices
Wells rules sens with POC D-dimer were higher than Geneva rules (95-96% vs 88-90%)
All rules with specificity of approx 50%, PPV approx 20%
NPV for Wells 99%, Geneva 97%
Failure rates: 1.2% simplified Wells (min) to 3.1% for simplified revised Geneva (max)
Three pts initially “low prob” eventually diagnosed with PE, regardless of rule applied.
Conclusion: Wells rule, esp simplified Wells, performed the best combined with POC D-dimer.
Simplified Wells = 1 point each for clinical signs of DVT, HR > 100, recent surg or immobilization, previous PE/DVT, hemoptysis, cancer, alt diagnosis less likely than PE.
Of the 3 who were “false negatives” of the rules+D-dimer, each has clinical predictors putting them at high risk. So, these rules should supplement, not replace, clinical judgment.
BMJ 2015 Sep 8; 351:h4438