Kearon C, de Wit K, Parpia S, et al. Diagnosis of Pulmonary Embolism with d-Dimer Adjusted to Clinical Probability. N Engl J Med. 2019 Nov 28;381(22):2125-2134. doi: 10.1056/NEJMoa1909159. (Original study)

BACKGROUND: Retrospective analyses suggest that pulmonary embolism is ruled out by a d-dimer level of less than 1000 ng per milliliter in patients with a low clinical pretest probability (C-PTP) and by a d-dimer level of less than 500 ng per milliliter in patients with a moderate C-PTP.

METHODS: We performed a prospective study in which pulmonary embolism was considered to be ruled out without further testing in outpatients with a low C-PTP and a d-dimer level of less than 1000 ng per milliliter or with a moderate C-PTP and a d-dimer level of less than 500 ng per milliliter. All other patients underwent chest imaging (usually computed tomographic pulmonary angiography). If pulmonary embolism was not diagnosed, patients did not receive anticoagulant therapy. All patients were followed for 3 months to detect venous thromboembolism.

RESULTS: A total of 2017 patients were enrolled and evaluated, of whom 7.4% had pulmonary embolism on initial diagnostic testing. Of the 1325 patients who had a low C-PTP (1285 patients) or moderate C-PTP (40 patients) and a negative d-dimer test (i.e., <1000 or <500 ng per milliliter, respectively), none had venous thromboembolism during follow-up (95% confidence interval [CI], 0.00 to 0.29%). These included 315 patients who had a low C-PTP and a d-dimer level of 500 to 999 ng per milliliter (95% CI, 0.00 to 1.20%). Of all 1863 patients who did not receive a diagnosis of pulmonary embolism initially and did not receive anticoagulant therapy, 1 patient (0.05%; 95% CI, 0.01 to 0.30) had venous thromboembolism. Our diagnostic strategy resulted in the use of chest imaging in 34.3% of patients, whereas a strategy in which pulmonary embolism is considered to be ruled out with a low C-PTP and a d-dimer level of less than 500 ng per milliliter would result in the use of chest imaging in 51.9% (difference, -17.6 percentage points; 95% CI, -19.2 to -15.9).

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. (Funded by the Canadian Institutes of Health Research and others; PEGeD ClinicalTrials.gov number, NCT02483442.).

Discipline Area Score
Emergency Medicine 7 / 7
Hemostasis and Thrombosis 7 / 7
Hospital Doctor/Hospitalists 6 / 7
Internal Medicine 6 / 7
Comments from MORE raters

Emergency Medicine rater

A must-read for emergency medicine. PE is the poster child for emergency medicine over-testing and consequential downstream over-diagnosis and over-treatment (see https://onlinelibrary.wiley.com/doi/full/10.1111/acem.12820). Whereas useful diagnostic tests generally adjust pre-test probabilities for disease, this study uses an innovative approach to adjust "abnormal" for D-dimer based on pre-test probability for VTE. One challenge to widespread uptake of these results to align post-D-dimer imaging decisions with clinical practice is that physicians too often neglect evidence-based risk stratification in advanced imaging (see http://pmid.us/27689922). How would the authors overcome this knowledge translation barrier?

Emergency Medicine rater

In my opinion, this study is practice-changing. When combined with the previously published YEARS studies, this validates a strategy of doubling the D-Dimer threshold to rule out PE in low-risk patients. The low-risk definition of Wells score =4 allows for most patients to fall into this category (87% in this study) and potentially benefit from the increased likelihood of having PE ruled out without requiring chest imaging. About 1 in 6 patients were able to forego chest imaging who would have required it without the D-Dimer adjustment, and there were no identified misses. I am using this strategy now and encouraging my partners to do the same.

Hemostasis and Thrombosis rater

Important for all those who diagnose pulmonary embolism from the prestigious Canadian thrombosis research group. These data expand the class of patients for whom diagnostic imaging is not necessary - those with low clinical suspicion and a d-dimer below 1,000 ng/mL, and for moderate-risk patients with d-dimer below 500. Previously, it was considered safe to withhold imaging for low-risk patients with d-dimer below 500, but it was uncertain what to do with moderate risk patients. These new data will change clinical practice BUT with some provisos: 1. The clinician should perform a valid clinical risk score (e.g., Wells) - clinicians don't always do this but rely on gestalt; 2. The conclusions apply to outpatients - Do they mean in an office setting or emergency room? The authors do not specify and the latter could be sicker; 3. These data may not apply to inpatients.

Hemostasis and Thrombosis rater

An important advance using adjustment of d-dimer values to further curtail unnecessary imaging.

Hospital Doctor/Hospitalists rater

As a hospitalist, the D-dimer test when <500 ng/ml, is useful to avoid further imaging in a low-risk Wells score. Unfortunately, this combination is not particularly common; whether a slightly higher cutoff (<1000) or lower cutoff with moderate probability would also convey good outcomes was unknown. This study answers that question. Patients with low prob and d-dimer <1000 had excellent 90-day outcomes. 6/1200 were lost to follow-up, which could have affected the final answer slightly. There were too few moderate prob (about 40) to answer that question effectively. However the strategy reduced the need for CTPA by 17%, a very important finding that can stem the tide of over-imaging if applied by all...and it should be.

Internal Medicine rater

Useful info on decreasing imaging by expanding the definitions of low-prob clinical and negative D-dimer. There were only 40 participants in the moderate-prob group with a negative D-dimer, so the confidence interval here is quite wide and perhaps is not as reassuring. I would also be interested to know how many patients got off-protocol imaging due to clinician preference and what percentage of those images were positive (I didn`t see it in the Results).
Comments from EvidenceAlerts subscribers

No subscriber has commented on this article yet.