A super simplified yet evidence-based approach to using the d-dimer in PE
Q1. Is PE your top differential?
If yes, do a V/Q or CTPA.
Wells et al suggested that if the PE was clinically the most likely diagnosis, then the d-dimer should not be used.
“According to our model, if pulmonary embolism is as likely as or more likely than any other diagnosis to cause the patient’s symptoms, the patient should not be assigned low clinical probability”
Using gestalt over a scoring system may not be such a bad thing, especially for more senior doctors. A study of 1038 patients comparing gestalt and the Wells score showed better determination of high and low risk with gestalt.
If no, go to Q2.
Q2. Is PE really in the differential diagnosis?
If yes, go to Q3.
If no, pause here.
There is a pressure to investigate all sorts of trivial things in A&E. If there is no real suspicion of a PE but you just ‘want to rule it out’ then a d-dimer is a dangerous gamble. Seek senior advice, or if you are the senior do the right thing.
Q3. Would you be happy to pursue definitive investigations and anticoagulant if the d-dimer were positive?
If yes, go to Q4 and get a Wells score.
If no, stop here.
Q4. Is there an alternative cause of a raised d-dimer?
If yes, consider CTPA or V/Q scan instead.
If no, go to Q5.
The D-dimer is a marker of the degradation of cross linked fibrin clots. The coagulation and fibrinolysis cascades are activated in pretty much any cause of inflammation (as in most hospital inpatients to some degree), as well as pregnancy and advancing age.
Do d-dimers still have a role in older patients? Yes, but be prepared for positives. A study in the BMJ looked at the role of age adjusting the d-dimer cut offs and found that it was safe to do so.
Q5. What does your lab say about using the Wells score for d-dimer?
Lab has definied criteria –> Use this
Not sure –> Find out from lab if there are any instructions from the d-dimer assay about Wells scores / pretest probability for interpretation
No lab criteria –> Use the NICE guidelines from 2012, which the BTS now refer to. A Wells score of 4.5 or more is high risk, and 4 or below is low risk.
This is important, as the d-dimer assays vary. Some are very high sensitivity and others just pretty high sensitivity. All d-dimer assays can be used to rule out PE in those at low risk. The very high sensitivity ones can also rule out PE in those at moderate risk of PE.
No assay rules out PE in those at high risk of PE. This is why doing a d-dimer in a high risk patient is not simply a waste of money (d-dimer nearly always positive) but maybe even dangerous.
Note that if the diagnosis of PE is most likely, then this automatically scores 3. It doesn’t take much to push the total score into high risk with this.
How do you define ‘PE most likely’? Isn’t this totally subjective?
Yes, but this may not be a bad thing. Wells said:
“For the final variable [PE most likely], which was not strictly defined, physicians were told to use the clinical information (obtained by history and physical examination), along with results on chest radiography, electrocardiogaphy, and whatever blood tests were considered necessary to diagnose pulmonary embolism.”
There is also evidence that relying on physician gestalt rather than an objective scoring system may be just as valid for deciding if the patient is high or low risk of a PE. The fact the Wells score essentially has a degree of built in gestalt may not be a bad thing.
What do you think? Does this make sense? Anything you’d do differently?