Last updated 3/18/25
For years, we've operated under the teaching that D-dimer testing is a no-go in patients with a high pre-test probability for pulmonary embolism (PE). The rationale? The higher prevalence of PE in this group decreases the negative predictive value of the test, increasing the risk of missing the diagnosis. A recent publication in Academic Emergency Medicine is challenging this long-held belief. But before we overhaul our practice and start withholding CTPA in high-risk patients, let's take a closer look at the data.
The Paper
Bannelier H, Kapfer T, Roussel M, et al. Failure rate of Dādimer testing in patients with high clinical probability of pulmonary embolism: Ancillary analysis of three European studies. Acad Emerg Med. 2025;32(2):116-122. doi:10.1111/acem.15032
Summary
This study is a post hoc analysis of three European cohorts: PROPER, MODIGLIANI, and TRYSPEED. These studies included 651 patients with a high clinical probability of PE (assessed by Wells or revised Geneva scores) who underwent D-dimer testing. The authors sought to evaluate the safety of ruling out PE based on D-dimer testing in this patient population. Key findings included:
- 70 patients had D-dimer levels below the age-adjusted threshold, and 0 were diagnosed with PE at 3-month follow-up.
- This resulted in a 0.0% failure rate (95% CI 0.0%ā6.5%), with Bayesian analysis estimating a 76.2% probability that the true failure rate is <2% (the accepted safety threshold).
- CTPA had a 0.7% failure rate in the subset of patients, underscoring that even āgold standardā imaging isnāt perfect.
Author conclusion
āIn this study, ruling out PE in high-risk patients based on D-dimer below the age-adjusted threshold was safe, with no missed PE. However, the large CI of the primary endpoint precludes a definitive conclusion.ā
Interpretation
Iāve been a long-time reader of First10EM and a fan of Dr. Morgensternās RAMMBO approach to interpreting clinical research, so I applied it to this study.
Recruitment
Key Questions: Were participants representative of the target population? Is the study size appropriate? Were there important exclusions that could affect the results?
ā Strengths:
- The study analyzed 12,300 patients from three European studies (PROPER, MODIGLIANI, and TRYSPEED) but may not be applicable to a US population.
- The final analysis included 651 patients with a high clinical probability of PE, making the findings relevant to high-risk ED populations.
ā ļø Limitations:
- Selection bias: The study used post hoc analysis, meaning the original trials were not designed to assess D-dimer in high-risk PE patients.
- The majority of included patients (90%) came from the TRYSPEED study, a retrospective study lacking 3-month follow-up, potentially affecting reliability.
- Missing data and exclusions (e.g., 2,963 patients were excluded due to missing D-dimer values), which could introduce bias.
Allocation
Key Questions: Were the groups similar at the beginning of the trial? Was allocation to groups truly random?
ā Strengths:
- Wells and revised Geneva scores were used to stratify patients into high-risk groups, which provides an objective measure of risk assessment.
- Use of Bayesian analysis helps account for uncertainty in failure rates.
ā ļø Limitations:
- The study was not a randomized controlled trial (RCT); instead, it was an observational secondary analysis of previous studies.
- Patient inclusion was based on available data rather than a prospective randomized assignment, making the findings prone to selection bias.
Maintenance
Key Questions: Were groups treated equally during follow-up? Were the outcomes of interest measured for all (or at least most) of the patients in the trial? (In other words, were patients lost to follow up, which could affect the reliability of the results?)
ā Strengths:
- The PROPER and MODIGLIANI cohorts had 3-month follow-up, ensuring events were captured over time.
- Death records were reviewed for patients lost to follow-up.
ā ļø Limitations:
- TRYSPEED lacked 3-month follow-up, meaning PE events in this group could have been missed, affecting reliability.
- Inconsistent follow-up methods across studies (some used adjudication committees, while others relied on retrospective chart reviews).
Measurement - Blind & Objective
Key Questions: Were the outcomes measured in an objective, valid and reliable way? Were outcome assessors blinded to exposure status?
ā Strengths:
- Clear primary outcome: Failure rate of D-dimerābased strategy (defined as PE diagnosed within 3 months in patients with negative D-dimer).
- Standardized D-dimer cutoff values (age-adjusted and fixed 500 ng/mL thresholds).
ā ļø Limitations:
- No blinding was mentioned in the study design.
- Heterogeneous study designs: Some patients had follow-up with CTPA only, while others had a 3-month clinical follow-up.
- D-dimer variability: Different assays may have been used across the European centers, affecting test performance.
- The study did not explicitly measure harms associated with delayed diagnosis or additional imaging/tests triggered by false-positive D-dimer results.
Conclusion
This study certainly challenges standard practice, but it doesnāt provide enough evidence to justify a paradigm shift in how we approach high-risk PE patients. While the 0% failure rate of the D-dimer strategy is intriguing, the wide confidence intervals (0.0%ā6.5%), the post-hoc nature of the analysis, and inconsistent follow-up across the three included studies all limit its impact.
From an emergency medicine perspective, we have to ask: Would this change how we practice? For now, the answer is no. Remember ā this is secondary analysis with small numbers of true D-dimer negative patients, leading to uncertainty in its conclusions. The retrospective inclusion of patients who were never intended to be the focus of this question makes it difficult to trust that the findings would hold up in a real-world setting.
That said, the study does raise an important question: Have we been too rigid in our exclusion of D-dimer testing in high-risk patients? A properly powered, prospective RCT is needed to determine whether D-dimer can safely rule out PE in these patients. Until then, the risk of missing a life-threatening diagnosis outweighs the potential benefit of reducing CTPA use in this population.For now, stick with CTPA in high-risk PE patients and letās wait for better data before changing our approach.
Read more on FOAMed
Sources
- Bannelier H, Kapfer T, Roussel M, et al. Failure rate of Dādimer testing in patients with high clinical probability of pulmonary embolism: Ancillary analysis of three European studies. Acad Emerg Med. 2025;32(2):116-122. doi:10.1111/acem.15032
- Morgenstern J. The December 2024 Research Roundup. First10EM. Published December 2024. Accessed March 18, 2025. https://first10em.com/the-december-2024-research-roundup/
- Milne K. SGEM#469: You Take My Breath Away ā D-Dimer for Ruling Out PE in High-Risk Patients. The Skepticsā Guide to Emergency Medicine. Published March 2025. Accessed March 18, 2025. https://thesgem.com/2025/03/sgem469-you-take-my-breath-away-d-dimer-for-ruling-out-pe-in-high-risk-patients/
- Radecki R. D-Dimer in High-Risk PE: A Gamble Worth Taking? REBEL EM. Published March 2025. Accessed March 18, 2025. https://rebelem.com/d-dimer-in-high-risk-pe-a-gamble-worth-taking/
This post is for education only and not medical advice.