Fear and Loathing in D-Dimer

Am I the only one who hates the d-dimer for pulmonary embolism? I can’t imagine that I am. It was supposed to reduce our number of CTs, but the data suggests that it has instead increased them. (Full disclaimer, I never practiced in the days of V/Q scans or the days without d-dimer, but this is what I’m told.)

I find my practice pattern typically using the PERC Rule and/or Well’s Criteria for PE to identify well-appearing people who are “very low risk,” who would likely be harmed more than benefitted by a d-dimer test. And then for low-risk, I’ll end up using a d-dimer.

But when the computer screen blips that the result is back, I get a similar little blip in my chest, hoping it’s going to be negative. Interesting that I feel this way, given that I have no other reaction like this, except occasionally while waiting for the altered patient’s rectal temperature.

On one hand, I wonder, if this is the reaction I’m feeling, hoping and trying to mentally will the number to be negative when I click the “View Results” button, should I have even ordered the test to begin with? And on the other is how atypical, nefarious, and sometimes-weird presentations of pulmonary embolism can be. And then on the third hand: is the pulmonary embolism in the otherwise healthy, young, well-appearing person actually cause for alarm? (Some experts would suggest that our bodies are in a constant state of coagulation/anticoagulation, and that we’re all walking around with occasional, small PEs that our lungs dissolve or filter.) Maybe this is different (“benign PE”) from the PE in the cancer patient, or the hypotensive patient, or the one with the saddle thrombus. And on the fourth hand: there’s not even any good data that anticoagulation is of any benefit in pulmonary embolism (even though it’s the standard of care, and we all still give it).

Maybe I just hate PEs, or ruling them out in seemingly low-risk patients: the time, the money, and most of all, the contrast load and radiation exposure. But for now, I guess we’re stuck with our imperfect tests, clinical gestalt, and bedside evaluations of risk and benefit.

, , , ,

  1. #1 by Lars - February 2nd, 2010 at 07:14

    I’m just a 4th year med student, but at least what I’ve been taught is that a D-dimer is best employed for its negative predictive value. Clinical suspicion tees up a low pretest probablilty and a negative D-dimer will add laboratory data useful in ruling out a PE. Another point I’ve been repeatdly taught is to recall when a D-dimer will be artificially elevated by ageing, trauma, etc.

  2. #2 by blaz - February 3rd, 2010 at 06:40

    Another one here. I must be forced into ordering d-dimer (occasionally a radiologist wants to know the result before agreeing to do a CT angio despite obvious clinical sings, ABG etc).

  3. #3 by a specialist in emergency medicine - February 4th, 2010 at 04:51

    The problem isn’t the D-dimer. The problem is that you, and all of us, are doing tests we believe are unnecessary just to protect ourselves in the unlikely event that our judgment is wrong. You get that little blip in your chest because you fear the positive D-dimer that will then REQUIRE you to do a CT of the chest that you truly think is unjustified. The fault lies with the public and their lawyers (and our colleagues who prostitute themselves to the plaintiff’s bar).

  4. #4 by Aaron Johnston - April 15th, 2010 at 23:06

    The misuse of the d-dimer is perhaps the most glaring example of a fear of litigation based testing strategy in emergency medicine. This is a poor strategy because it is ineffective and because it places our patients at heightened risk of iatrogenic harm.
    In fact the problem is that the d-dimer is only a useful test when employed in an evaluation strategy for possible PE/DVT in low and intermediate risk patients. It is not useful in the high risk patient group. Thus when the radiologist asks you what the d-dimer is in the cancer patient with a PICC line, a swollen arm, breathlessness and an O2 sat of 90%, the correct response is that the d-dimer is irrelevant and therefore was not ordered.
    The same holds true in evaluation of the no risk/extremely low risk patient, in this setting the test provides no useful information. The 20 year old college student who comes to emerg because they had one day of cough, runny nose and slight discomfort with cough and a normal physical exam is a good example. This patient is likely an ‘almost zero risk’ patient. The right tests to do on this patient are none, with a clear explanation to the patient of symptoms that would necessitate return.
    It is folly to think that an aggressive testing strategy in the near zero risk patient is benign or protects from litigation. Needless tests are risky to the patient. What if the d-dimer is high, what if you order a CTA and irradiate your patient, what if you find a small sub-segmental PE in a no risk patient and anti-coagulate them? None of these things are free of risk.
    If we in emergency medicine continue with the insanity of irrelevant testing of well patients to prevent litigation, we should also change the research statistics we use to reflect this; NNPL (number needed to prevent litigation) and NNK (number needed to kill a patient with our tests). We could use a NNPL:NNK ratio to determine how comfortable we were in testing the extremely low risk patients (rather than the more sensible strategy of simply telling them they are well, and that they should return if they have any of the red flag symptoms you will list and write down for them…)
    My 2 cents…
    Dr. J

(will not be published)

  1. No trackbacks yet.