Rationale and request for industry guidance to standardize the threshold for an abnormal D-dimer concentration in the evaluation of patients with suspected pulmonary embolism
June 30, 2014

Jeffrey A. Kline, MD
Vice Chair of Research
Department of Emergency Medicine
Professor, Department of Cellular and Integrative Physiology
Indiana University School of Medicine
720 Eskanazi Avenue
Indianapolis, IN 46202
317 880 3869 office
317 670 0541 cell

Zachary Kahler MD
Research Faculty
Department of Emergency Medicine
Professor, Department of Cellular and Integrative Physiology
Indiana University School of Medicine
720 Eskanazi Avenue
Indianapolis, IN 46202

This report concerns the variable concentrations used by different manufacturers of D-dimer assays. The primary issue is what threshold defines an abnormal D-dimer value in the diagnostic consideration of patients with possible deep venous thrombosis and pulmonary embolism. Pulmonary embolism occurs in about 1 in 1000 persons per year and can cause sudden death.1
The concern arises from the multiple cutoff values that physicians are exposed to by their hospital laboratory, depending on the manufacturer of the D-Dimer assay. The root cause of the problem stems from varying capture antibody epitopes used in different devices. This problem is particularly acute in the emergency care setting. For example, at the Methodist Hospital in Indianapolis, IN, the point of care laboratory in the emergency department uses the HemosIL D-dimer with an abnormal cutoff of 230 ng/mL, but the central laboratory, which is used for inpatient care, uses a D-dimer with a 400 ng/mL cutoff. As is the case with many metropolitan cities, some physicians at Methodist Hospital also work at nearby Eskenazi Hospital, which uses a different assay, and has a cutoff of 500 ng/mL. Physicians are confused, and this confusion contributes to underuse of the D-dimer test.2 In the absence of a D-dimer, the only practical diagnostic test to exclude pulmonary embolism is with medical imaging, primarily with contrast-enhanced computerized tomography of the chest—an expensive test that imparts 10 mSv radiation and leads to kidney injury in 15% of patients.3 Underuse of D-dimer increases unnecessary medical imaging in patients with symptoms of pulmonary embolism, leading to high medical radiation exposure in patients with no significant cardiopulmonary problems.4
Recent research has found that the D-dimer threshold for abnormal can be varied by age, using the formula age×10 ng/mL, assuming a standard cutoff of 500 ng/mL.5 Adjusting the D-dimer threshold for patient-specific conditions will increase its utility, obviating more medical radiation and risk of unnecessary imaging. It can be expected that future research will allow threshold adjustments to reduce medical imaging for pulmonary embolism in patients with cancer and pregnancy.6;7 The variable D-dimer cutoffs of different assays reduces physician willingness to use the D-dimer at all, let alone adjust its value for patient specific conditions. The basis for this assertion comes from audience input at national meetings (e.g., the American Academy of Emergency Medicine, NY, NY, February 2014), recent social media postings (https://twitter.com/EMSwami/status/443118390178484224), and a survey.
In May 2104, we created a survey that we sent primarily to emergency care physicians in the US to inquire about the significance of the varying D-dimer threshold. The survey consisted of 8 questions and was administered using the REDCap archiving system to physicians from a wide geographic and demographic representation in the United States. The questions were designed to assess the magnitude of the problem and the desire for a solution. Clinicians had the option of submitting their electronic authorization to sign a petition to the FDA to take steps to normalize the D-dimer threshold in lieu of a live signature.
Approximately 1,000 emergency physicians responded to the survey, which represents 3.1% of all board-certified emergency physicians in the US. The survey asked 8 questions and at the end, asked if the respondent would authorize his or her name to be added to a petition to FDA about this matter.

The following shows the question asked, the format for the answer, and the main results of each question.

Question 1: Various D-dimer assays have differing cutoffs to define an abnormal value. Do you believe this is a problem?
Format: Visual Analog Slider, 0-100%

Mean 70.05% (SD 23.322%)
Median 73% (IQR 58% to 87%)

Question 2: The INR was created to standardize the prothrombin time. How strongly do you believe that a similar standardization should be performed with the D-dimer?

Format: Visual Analog Slider, 0-100%

Mean 79.24% (SD 20.614%)
Median 85% (IQR 85% to 97%)

Question 3: Which of the following best describes how often you use a D-dimer in your practice?

Format: Multiple Choice

Daily (315, 32.4%)
Weekly (523, 53.9%)
Monthly (123, 12.7%)
Yearly (10, 1.0%)

Question 4: Which D-dimer assay does your hospital use?

Format: Multiple Choice

I don't know (861, 88.7%)
VIDAS (bio- Merieux SA, Marcy-Etoile, France) (36, 3.7%)
Liatest (Diagnostica Stago, Asnières sur Seine, France) (21, 2.2%)
Hemosil (Dade Behring, Marburg, Germany) (27, 2.8%)
Advanced D-dimer (Dade Behring) (5, 0.5%)
Biopool Minutex (diaPharma, West Chester, OH) (2, 0.2%)
MDA (Organon Teknika Corporation, Durham, NC)(2, 0.2%)
Other (17, 1.8%)

Question 5: Do you ever adjust the "abnormal" D-dimer threshold in your clinical practice? (e.g. for patient age, trimester in pregnancy)

Format: Multiple Choice

I didn't know I could. (184, 19.0%)
I would, but I don't know how my hospital's d-dimer cutoff would scale to published values. (326, 33.6%)
I do not want to adjust this myself. (245, 25.3%)
Yes, I do this. (215, 22.2%)

Question 6: If a D-IMR* ratio was created, what numerical cutoff would you choose to define as "normal"?
*D-dimer International Managed Ratio

Format: Multiple Choice

1.0 (580, 59.8%)
500 (371, 38.2%)
other (19, 2.0%)
(Unique answers from “other:” 0.50, 50, 100, 300, 1000)

In summary, from Question 1, 81% of physicians surveyed believed the varying D-dimer cutoff was a problem (defined as a response more than 50 on the visual analogue scale), and 87% believed that the D-dimer should be standardized (Question 2). Based on the results of this survey, we note several concerns. Almost 90% of emergency physicians do not know what type of D-dimer assay their hospital uses, which is a major patient safety issue given the different cutoff values. Along the same lines, the most common response to the question “would you adjust a D-dimer cutoff” is “I would, but don’t know how my hospital’s assay compares to national guidelines.”
In addition, 600/973 (61.6%) of all survey participants chose to sign the petition. These signatures are attached in Appendix 1 after the references and represent 1.9% of all emergency physicians in the country.
Query of the FDA 510K database indicates the following D-dimer assays have an indication for use statement that includes to some extent, the diagnosis or diagnosis and exclusion of venous thromboembolism. We found that VIDAS, Stratus, Innovance, HemosIL, Stratus Dade, TinaQuant, Triage, Cardiac (Roche), Advanced, and the MDA assays had evidence of FDA clearance to market for VTE. The thresholds for normal range from 230 ng/mL with HemosIL, 450 ng/mL for Stratus, and 500 ng/mL for the remainder. The 510K website shows clearance of dozens of other D-dimer assays to measure fibrin split products and no indication for use specific to VTE. The thresholds for abnormal for these assays vary widely, and many are not reported in the filing. Unfortunately, in many emergency departments, clinicians must make decisions about possible pulmonary embolism using D-dimer assays without FDA clearance to market for VTE exclusion.
Our request is the FDA consider issuing draft guidance to industry that that D-dimer devices provide an arithmetic correction to the mass unit of D-dimer concentration output by their device, normalized to a specific cutoff value, possibly 1.0 (as suggested by 59.8% of respondents), analogously to the international normalized ratio for the prothrombin time (INR). We recognize that such a step would require extra cost and time for each manufacturer perform standard curve testing using a validated external standard for D-dimer concentration and then rewrite software controlling device output. However, we believe the extra work would be justified as standardization would reduce confusion and permit wider use of threshold adjustment to the benefit of patients.

Reference List

​(1) ​Goldhaber SZ. Venous thromboembolism: epidemiology and magnitude of the problem. Best Pract Res Clin Haematol 2012; 25(3):235-242.

​(2) ​Venkatesh AK, Kline JA, Courtney DM, Camargo CA, Plewa MC, Nordenholz KE et al. Evaluation of pulmonary embolism in the emergency department and consistency with a national quality measure: quantifying the opportunity for improvement. Arch Intern Med 2012; 172(13):1028-1032.

​(3) ​Mitchell AM, Jones AE, Tumlin JA, Kline JA. Prospective study of the incidence of contrast-induced nephropathy among patients evaluated for pulmonary embolism by contrast-enhanced computed tomography. Acad Emerg Med 2012; 19(6):618-625.

​(4) ​Kline JA, Shapiro NI, Jones AE, Hernandez-Nino J, Hogg MM, Nelson RD et al. Outcomes and radiation exposure of emergency department patients with chest pain and shortness of breath and ultralow pretest probability. Ann Emerg Med 2014; 63:281-288.

​(5) ​Righini M, van EJ, den Exter PL, Roy PM, Verschuren F, Ghuysen A et al. Age-adjusted D-dimer cutoff levels to rule out pulmonary embolism: the ADJUST-PE study. JAMA 2014; 311(11):1117-1124.

​(6) ​Kline J.A., Hambleton GW, Hernandez J. D-dimer concentrations in normal pregnancy: New diagnostic thresholds are needed. Clinical Chemistry 2005; 51(5):825-829.

​(7) ​Righini M, Le GG, De LS, Roy PM, Meyer G, Aujesky D et al. Clinical usefulness of D-dimer testing in cancer patients with suspected pulmonary embolism. Thromb Haemost 2006; 95(4):715-719.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License