Date: January 31st, 2020

Reference: Kearon C et al. Diagnosis of Pulmonary Embolism with D-dimer Adjusted to Clinical Probability. NEJM 2019.

Guest Skeptic: Dr. Nour Khatib is an emergency physician at Lakeridge Health and Sunnybrook Hospital. She is also one of the organizers of the EM Vision 2020 conference.

This was recording this LIVE at the EM Vision 2020 conference in Oshawa. You can watch the video of this presentation on The SGEM Facebook page

Case: You are caring for a 33-year-old female who comes to your emergency department with chest pain. She was attending the incredible EM Vision 2020 conference and while listening to Dr. Alan Drummond talk about over over-crowding, she developed sudden onset of pleuritic chest pain.

She is not on any hormone replacement therapy, no history of venous thromboembolism, has not been immobile, has no hemoptysis and has no history of malignancy. Her vital signs are as follows: heart rate 105 beats/minute, blood pressure 110/70, respiratory rate 18 breaths/minute, oxygen saturation 96% on room air. You order a d-dimer test to rule out pulmonary embolism (PE) and it comes back at 850 and your first reaction is DOH!!

You think to yourself, “well my hands are tied” and you are about to order the CT angiogram when you remember hearing about a PEGeD study. She is Well’s score low, but how are you going to handle that elevated d-dimer?

Background: There is a reason why pulmonary embolism is covered so often on The SGEM and other podcasts (SGEM#51, #126, #163, and #219); it’s a diagnosis we are appropriately trained to always think of but it presents itself with symptoms that are common in a variety of other conditions.

Of course, how can we forget the DDD threat – the Dreaded D-Dimer. It is a highly sensitive test but with low specificity. For the longest time we treated it like a binary test. If the DDD is positive (>500ng/ml), then further imaging (CTA or VQ). If the DDD is negative (<500ng/ml) we do are happy dance and no further testing.

We need to talk a little bit about Bayesian thinking. Thomas Bayes was an English statistician, philosopher and Presbyterian minister, a real renaissance man. He is best known for a theorem that he never published. Bayes was trying to mathematically prove divinity but kept finding that observed miracles could be explained by random chance. It was Richard Price who went through his notes and published the theorem after his death. The main thing about Bayesian thinking is that posterior probability is dependent on prior probability. It is really how we think as clinicians. We take a look at some information (history, physical examination and tests) and we interpret it, based on our clinical experience (pre-test probability). All information needs to be interpreted. Bayes can help us with that part.

There have been a few changes to the DDD over the last few years. One thing that has changed my practice is the ADJUST PE Study (Righini et al JAMA 2014). That is where you take the patient’s age over 50 and times it by 10. The result is the new upper limit for the DDD. If the patient is 68 years of age, a d-dimer less than 680 is considered negative or rules out a PE. It is still binary but the cut-off is variable.

We the ADJUST PE Study with Dr. Kirsten de Witt out of McMaster University on SGEM#112. She is one of the authors on this study we are reviewing today.

Drs. de Wit and Kearon

Another change is using the YEARS criteria in diagnosing PE.  Just to remind everyone about the three items of the YEARS criteria are:

  1. Clinical signs of DVT
  2. Hemoptysis
  3. PE most likely diagnosis

If your YEARS score is zero you can use a d-dimer cut-off of 1,000ng/ml which is double the usual cut-off of 500ng/ml. If the YEARS score is 1 or greater than you use the traditional d-dimer of 500ng/ml.

We recently covered the study looking at pregnancy-adapting the YEARS algorithm on SGEM#277. That was another LIVE episode that was recorded at the Kewartha EM conference in Peterborough with the wonderful Dr. Theresa Robertson. We felt this new algorithm was not ready for prime time until externally validated.

Now there is the PEGed study or the Pulmonary Embolism Graduated D-dimer study looking at clinical pre-test probability (C-PTP).

Clinical Question: Can a clinical pretest probability-based D-dimer safely rule out the diagnosis of pulmonary embolism without imaging?

Reference: Kearon C et al. Diagnosis of Pulmonary Embolism with D-dimer Adjusted to Clinical Probability. NEJM 2019.

  • Population: Adults (18 years of age or older) from the emergency department or outpatient clinic with signs or symptoms of possible pulmonary embolism
    • Exclusions:
      • < 18 years of age
      • Received full-dose anticoagulant therapy for 24 hours
      • Undergone major surgery in the past 21 days
      • had a D-dimer level that was known before the C-PTP was assessed
      • had undergone chest imaging contrary to the protocol (i.e., before the C-PTP was documented, or despite having a D-dimer level of <1,000 ng/ml for a low C-PTP or <500 ng/ml for a moderate C-PTP)
      • had undergone contrast-enhanced CT of the chest for another reason
      • had an ongoing need for anticoagulant therapy (example diagnosed with A fib)
      • had a life expectancy of less than 3 months,
      • Pregnant
      • geographically inaccessible for follow-up
  • Intervention: This was a diagnostic accuracy study of the risk of PEusing the Well’s criteria to assess clinical pre-test probability
    • Low Risk: Well’s score 0 to 4 (d-dimer of <1,000 ng/ml was used to rule out patients with low C-PTP)
    • Moderate Risk: Well’s Score 4.5 to 6 (d-dimer of <500 ng/ml was used to rule out patients with moderate risk)
    • High risk: Well’s Score 6.5 or greater and would receive diagnostic imaging (generally, CTPA) with no d-dimer testing
  • Comparison:
    • Standard strategy where d-dimer cut-off is 500 ng/ml, YEARS protocol and an age-adjusted D-dimer
  • Outcome:
    • Primary Outcome: The incidence of venous thromboembolism (VTE) at three-month follow up among the low and moderate clinical pre-test probability groups who had negative (adjusted) d-dimers and did not receive any anticoagulation.
    • Secondary Outcomes: The percentage of patients with VTE in predefined subgroups, bleeding events and deaths and the percentage of patients who avoided diagnostic imaging and had a low C-PTP and a d-dimer under 1,000 ng/ml or those with moderate C-PTP and a d-dimer less than 500 ng/ml.

Authors’ Conclusions: “A combination of a low C-PTP and a D-dimer level of less than 1000 ng per milliliter identified a group of patients at low risk for pulmonary embolism during follow-up”

Quality Checklist for Observational Study:

  1. Did the study address a clearly focused issue? Yes
  2. Did the authors use an appropriate method to answer their question? Yes
  3. Was the cohort recruited in an acceptable way? Unsure
  4. Was the exposure accurately measured to minimize bias? Yes
  5. Was the outcome accurately measured to minimize bias? Yes
  6. Have the authors identified all-important confounding factors? Yes
  7. Was the follow up of subjects complete enough? Yes
  8. How precise are the results? Precise with narrow confidence intervals
  9. Do you believe the results? Yes
  10. Can the results be applied to the local population? Yes
  11. Do the results of this study fit with other available evidence? Yes

Key Results: They enrolled a total of 2,017 patients in the study (1,752 low risk C-PTP, 218 moderate risk C-PTP and 47 high risk C-PTP). The mean age was 52 years and two-thirds were female.

None of the low or moderate risk C-PTP patients who had “negative” adjusted d-dimers had a VTE diagnosed at 3 months.

  • Primary Outcome: VTE in low or moderate C-PTP patients at 3 months was 0% (95% CI; 0% to 0.29%).

There was one patient in the low-risk group had a + d-dimer test of 1,200ng/ml and a negative CT PE, was found to have a PE at follow up.

  • Secondary Outcomes:
    • Percentage of patients with VTE in predefined subgroups after initial testing and anticoagulant therapy: Low 1.15%, Moderate 0% and High 0%.
    • Bleeding Events: Seven major and 23 minor bleeding episodes
    • Deaths: None of the 34 deaths were attributed to PE
    • Avoiding Diagnostic Imaging: Reduced by 17.6% (51.9% with YEARS and 34.3% with PEGeD).

  1. Patients: Most of the patients included in the study were considered low risk (87%). There were only 11% (218/2017) moderate risk patients and 2% (47,2017) high risk patients. Then when you look at those who were stratified as low risk, only 5% of those patients had a PE. If you look back at some of the original studies on diagnosing PE (PIOPED 1990) the low risk group had about more than double that rule-in rate around 12%. This suggests that these were not low risk but very low risk patients. This can increase the diagnostic strategy given the low prevalence of disease. Even the moderate risk patients had a rule in rate of 20% compared to PIOPED that was 33%.
  2. Selection Bias: It is important to always look at who was included and excluded from the study. The authors lay out a clear exclusion criterion. However, some of the exclusion criteria were subjective (ex: life expectancy < 3 months). This could lead to some section bias in excluding patients. With regards to inclusion criteria, it is unclear how these patients were specifically recruited. They did not explicitly state in the manuscript that it was consecutive patients presenting to the clinic or emergency department. The methods section says the study population included those patients with signs or symptoms suggestive of pulmonary embolism. This is pragmatic, uses clinical judgment/gestalt but is also subjective. That lack of clear objective inclusion criteria also could introduce some selection bias.
  3. Comparison: No direct comparison between this study and current practice (age-adjusted and YEARS algorithm), the secondary analysis was done in retrospect.
  4. Confirms YEARS: The study provides more evidence for what we have already suspected with the YEARS study. That is where we learned that we can apply a <1,000ng/ml d-dimer limit for patients with low pretest probability. Furthermore, the YEARS algorithm was validated in pregnant patients which this study excluded. The results from the moderate pretest probability group where they tried to rule out PE based on a value of 500ng/ml are inconclusive given the low number of moderate risk participants.
  5. External Validity: They reported a potential 18% reduction in imaging using a PEGeD strategy comparted to YEARS. Less imaging results in less incidentalomas, over diagnosing, over treating, less harm and less cost. While it is very good to reduce imaging, we are wondering how persuasive this will be on some clinicians? This study was conducted in Canada. What would be the impact in other countries like the USA? They face much different medical legal pressure and patient expectations. It would be great to see an impact analysis done in different practice environments.

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We generally agree with the authors’ conclusions.

SGEM Bottom Line: It is reasonable to use the PEGeD strategy to safely rule out PE in low risk patients.

Case Resolution: With our patient who is low risk C-PTP and a d-dimer under 1,000 ng/ml, it is reasonable to rule out PE and forgo any further chest imaging in our Canadian practice environment.

Clinical Application: Adjusting d-dimer threshold to less than 1,000 ng/ml for low risk clinical pre-test probability patients can reduce imaging and appears safe.

Dr. Nour Khatib

What Do I Tell My Patient? We were concerned you might have had a blood clot in your lung. This can be very serious and even deadly problem. We ran some blood tests and they were normal. It’s very unlikely you have a blood clot. Now we need to look for other causes of your chest pain.

Keener Kontest: Last weeks’ winner was Glen Keating from Ohio. He knew BONES is a mnemonic BONES related to the potential complications of BVM ventilation and it stands for Beards, Obstructions, Neck stiffness or Neck mass, Expecting (pregnancy) and Stridor/Snoring. The MOANS mnemonic is similar and stands for Mask seal, Obesity, Age >55, No teeth and Sleep apnea and Stiff Lungs.

Listen to the podcast to hear this weeks’ question. Send your answer to with “keener” in the subject line. The first correct answer will receive a cool skeptical prize.

Other FOAMed:

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.