“We can’t use our d-dimer for that.”
“No, no. Our d-dimer cannot be age adjusted.”
“Umm, so what’s our cutoff supposed to be?”
Hello friends, and welcome to today’s episode of information overload. It is very long and took a week to put together, but it's an important one! Nerd out with me for a bit.
Today’s topic - can we age-adjust Maimo’s d-dimer? Scroll down to the bottom for summary points.
Shout out to Dr. Lalley for inspiring this one!
The D-dimer Debate.... Debacle?
Or alternatively, a pearl of the day to (maybe) clear up some confusion and take another look at the commonly asked question of -- can we age-adjust the d-dimer we use at Maimo?
The diagnosis of PE is a particular enemy of mine, and I often turn to a wonderful POTD Dr. Alison Leung, Dr. Anna Bona, and Dr. Errel Khordipour to remind myself of a reasonable stepwise approach. https://www.maimonidesem.org/blog/potd-pulmonary-embolism-decision-rules-beyond-the-basics
In her original post, Dr. Leung writes “At Maimo, you cannot age adjust your d-dimer!” I have been told similarly since starting at Maimo, but I never understood why. We don’t have the right d-dimer. Our lab test is different. Hand wavy-ness. Seemed good enough for me... But not anymore! To figure this out, we go back in time a bit.
The Start of Age-Adjusted D-dimers
Studies (https://pubmed.ncbi.nlm.nih.gov/11020391) show that d-dimer levels typically increase with age, yet our upper limit on our d-dimer assays used to be at a fixed cutoff 500 ng/ml. This theoretically meant that the clinical usefulness of the d-dimer in ruling out VTE in low risk patients decreased the older the patient.
In 2014, Dr. Righini and his team published the ADJUST-PE study (https://jamanetwork.com/journals/jama/fullarticle/1841967) which assessed whether or not an age-adjusted d-dimer cutoff (defined as age x 10) could increase the number of patients safely ruled out for PE / VTE. Or phrased differently, could more unnecessary scans be prevented with a higher d-dimer cutoff, without sacrificing the sensitivity of the test? Their conclusions: Yes!
Below, people much smarter than I am taking deep dives into the study.
Core EM Summary: https://coreem.net/journal-reviews/age-adjusted-d-dimer/
RebelEM Summary: https://rebelem.com/age-adjusted-d-dimer-testing/
Original Study that ADJUST-PE Validated: https://www.bmj.com/content/346/bmj.f2492.full
Sharp AL, Vinson DR, Alamshaw F, Handler J, Gould MK. An Age-Adjusted D-dimer Threshold for Emergency Department Patients With Suspected Pulmonary Embolus: Accuracy and Clinical Implications. Ann Emerg Med. 2016;67(2):249-57.
Righini M, Van es J, Den exter PL, et al. Age-adjusted D-dimer cutoff levels to rule out pulmonary embolism: the ADJUST-PE study. JAMA. 2014;311(11):1117-24.
Schouten HJ, Geersing GJ, Koek HL, et al. Diagnostic accuracy of conventional or age adjusted D-dimer cut-off values in older patients with suspected venous thromboembolism: systematic review and meta-analysis. BMJ. 2013;346:f2492.
Awesome sauce! So instead of using a fixed upper limit of 500 µg/L, I can use the formula of age x 10 to get the upper limit (for patients older than 50). So for an 80-year-old man, the upper limit for a VTE rule-out would be 80 x 10 = 800. Right? Right?!?!
..... Yeah, if only.
Turns out, D-dimers are not the easiest things to measure.
What the heck is a D-dimer?
Also, let’s refresh our memories as to what a D-dimer is. (I’m lying, this never existed in my brain before now).
To get to a d-dimer we have to make a clot and break it down. (I promise this will become relevant later) To make a clot we start with fibrinogen. Fibrinogen is made of two “D” domains connected by an “E” domain in the middle.
Now, to make a clot we add:
fibrinogen + thrombin + factor XIIIa = fibrin mesh clot
Then to break down the clot:
Fibrin mesh clot + plasmin = broken down fibrin mesh = fibrin degradation products, one of which is two “D” domains directly connected to each other, called a D-dimer!
Great! That seems simple enough.
How D-dimer Assays Work, and How They Differ
The problem in D-dimer variability comes from the differences in all the assays. There are:
Five major types of D-dimer assay technology
Two main reagent types: D-dimer Units (DDU) or Fibrinogen Equivalent Units (FEU)
Eight (ish) units for the results of the assay
DDU/mL, mg DDU/L, μg DDU/L, μg DDU/ mL
FEU/mL, FEU/L, FEU/mL, FEU/mL
Combining these differences, that’s forty-five plus ways of getting D-dimer results. Not great.
In an attempt to figure this out further, I dug deep to try to remember some biochemistry, but turns out there’s none left in my brain. So I checked out a textbook (so retro). Book pdf attached for reference.
John D. Olson, Gregory S. Makowski Chapter One - D-dimer: An Overview of Hemostasis and Fibrinolysis, Assays, and Clinical Applications, Advances in Clinical Chemistry, Volume 69, 2015, Pages 1-46, https://doi.org/10.1016/bs.acc.2014.12.001.
The different assay types all work slightly differently, but they all essentially work like this.
As un-subtly highlighted in red in the above diagram, the output value of these assays are based on the assay’s calibrator reagent, namely DDU or FEU.
Why do we care? Because the initial studies including ADJUST-PE were based on the FEU calibrator reagent, so the units are expressed in FEUs. What do we have at Maimo?
..Dang it.
D-Dimer Units (DDU) vs Fibrinogen Equivalent Units (FEU)
Okay fine. The rabbit hole has led us to DDU vs FEU. What’s the difference?
Essentially in the early days of D-dimer testing, these assay creators had trouble creating (and agreeing on) a unified agent by which to calibrate their d-dimer machines.
One group used a known concentration of purified d-dimer as the calibrator, called D-dimer Unit, or DDU
One group used a known concentration of purified fibrinogen + added factor VIII to create a controlled clot + and then added plasmin to degrade the clot = to create a solution containing D-dimer. Because the exact final concentration of D-dimer is not known, the machine was calibrated based on the amount of initial fibrinogen. The calibrator contains an amount of D-dimer relative to a known input concentration of fibrinogen. So the unit is called Fibrinogen Equivalent Unit, or FEU.
Fibrinogen weighs 340 kDa. D-dimer weights 195 kDa. 340 divided by 195 is ≈ 1.75.
Therefore, because of the difference in weight, D-dimer values reported in FEU are roughly 1.75 higher than values reported in DDU.
This ratio is typically approximated to 2. As in, DDU * 2 = FEU.
Have you ever used a FEU-based D-dimer cutoff and applied it to a DDU lab test? Or put differently, have you ever used D-dimer cutoffs that you learned from literature or FOAMed and applied it to our Maimo D-dimer test (which reports on DDUs)? If so, you have probably sent home “rule-out PEs or DVTs” patients based on a d-dimer test that was not actually negative.
... Fantastic.
Can I just safely convert between DDU and FEU?
Short answer? Most likely. I’ll give you the facts, you decide.
Turns out, the DDU and FEU confusion gets even worse. Efforts to develop a standard calibrator agent that would react the same or similarly with every assay failed. In fact, one of the reasons FEU was even created at all was because the efforts to create a standard calibrator solution that contained only highly purified d-dimer lead to wide variability in the different assays.
Currently, each manufacturer uses their own standards, agents, and assays.
Even beyond this, there is huge variability in how D-dimer results are reported. Up to 34% of labs even convert their units, including between FEU and DDU, before ever reporting the lab value (although this percentage is decreasing, 13% on the most recent self-reported survey by the College of American Pathologists). As D-dimer plays a role in assessing severity of COVID-19 illness, the issues in D-dimer confusion has recently been brought into focus. See articles below.
Lippi, Giuseppe & Tripodi, Armando & Simundic, Ana-Maria & Favaloro, Emmanuel. (2015). International Survey on D-Dimer Test Reporting: A Call for Standardization. Seminars in thrombosis and hemostasis. 41. 10.1055/s-0035-1549092.
Olson JD , Cunningham MT, Higgins RA, Eby CS, Brandt JT. D-dimer—simple test, tough problems. Arch Pathol Lab Med2014;137:1030–8.
Moser Karen A, D-dimer: Common Assay, Challenges Abound, Caution Advised The Journal of Applied Laboratory Medicine, Volume 3, Issue 5, 1 March 2019, Pages 756–759, https://doi.org/10.1373/jalm.2018.027847
Favaloro EJ, Thachil J. Reporting of D-dimer data in COVID-19: some confusion and potential for misinformation. Clin Chem Lab Med. 2020 Jul 28;58(8):1191-1199. doi: 10.1515/cclm-2020-0573. PMID: 32432563.
https://researchoutput.csu.edu.au/ws/portalfiles/portal/46574546/46574490_published_article.pdf
Get to the point! Can I Use Maimo’s D-dimer or Not?
Scroll down for the faster answer.
Shoutout to our colleagues in the hematology lab who helped me figure out this information.
Our Maimo labs uses HemosIL D-Dimer HS (part number 0020007700) reagent made by Instrumentation Laboratory, A Werfen Company. https://www.instrumentationlaboratory.com/us/en/hemosil-reagents. This reagent reports D-dimer values in DDU.
As mentioned, the original studies including ADJUST-PE used a number of different types of D-dimer assays from different companies, but nearly all of them reported in FEUs.
However, articles from the last few years have looked into whether or not DDU-based assays can also be age adjusted. Below is one such article, very well summarized by our very own Dr. Turchiano.
Original Article:
https://pubmed.ncbi.nlm.nih.gov/28079562/
Summary and Review by Dr. Turchiano
https://coreem.net/journal-reviews/age-adjusted-d-dimer-2/
While the study had its limitations, especially as a retrospective cohort study at a single hospital center, it found age-adjusting DDUs (with the converted formula of age x 5) that specificity improved from 68% to 78% without additional missed PEs.
ACEP 2018 Clinical Policy
https://www.acep.org/patient-care/clinical-policies/acute-venous-thromboembolic-disease/
https://www.acep.org/globalassets/new-pdfs/clinical-policies/clinical.policy.suspected.acute.venous.thromboembolic.disease.pdf
“In adult patients with low to intermediate pretest probability for acute pulmonary embolism (PE), does a negative age-adjusted D-dimer result identify a group of patients at very low risk for the diagnosis of PE for whom no additional diagnostic workup is required?”
Level B Recommendations
In patients older than 50 years deemed to be low or intermediate risk for acute PE, clinicians may use a negative age-adjusted D-dimer* result to exclude the diagnosis of PE. *For highly sensitive D-dimer assays using fibrin equivalent units (FEU) use a cutoff of age×10 μg/L; for highly sensitive D-dimer assays using D-dimer units (DDU), use a cutoff of age×5 μg/L.
RebelEM Dr. Anand Swaminathan Summary of These Guidelines
https://rebelem.com/acep-clinical-policy-on-acute-vte-2018/
Per ACEP guidelines with a Level B recommendation, we can age-adjust our DDU-based d-dimer with the formula of age x 5 μg/L.
Ultimately it is up to you whether or not you wish to age-adjust a DDU-based d-dimer. However, for me personally after reading extensively about D-dimers and learning that the variance in reliability does not seem to not hinge on whether or not the units are FEU vs DDU (but rather the manufacturers themselves), I feel comfortable age adjusting my DDU D-dimers.
What about adjusting for clinical probability, like in YEARS or PEGeD?
So in more recent studies, the d-dimer is not adjusted for age, but rather by clinical probability. Meaning that patients for whom the clinician has a lower suspicion for VTE (based on the algorithm), the upper limit d-dimer cutoff can be higher. Two such notable recent approaches are YEARS and PEGeD.
The YEARS algorithm clinically stratifies patients based on the number of YEARS criteria, namely clinical signs of deep vein thrombosis, hemoptysis, or clinical suspicion of PE. Those with fewer criteria have a higher d-dimer cutoff. (See below for further dive into YEARS). In the original study, while it makes no mention of DDU vs FEU, the threshold of a ‘negative’ d-dimer set at 500 ng/ml implies that FEU-based tests were used.
YEARS Algorithm 2017 - Original Paper
https://pubmed.ncbi.nlm.nih.gov/28549662/
However, Dr Anand Swaminathan of Rebel EM in reviewing the YEARS approach and its expansion to include pregnant patients, suggests that DDU based tests can be utilized as well with the d-dimer cutoffs converted by a factor of 2, accordingly. (500 FEU ng/ml ≈ 250 DDU ng/ml)
https://rebelem.com/the-years-study-simplified-diagnostic-approach-to-pe/
https://rebelem.com/pregnancy-adapted-years-algorithm-for-pe-ready-for-prime-time/
In the PEGeD study, a similar approach with small differences to YEARS, they actually included 14 patients with DDU tests (with cutoffs at 230 ng/ml, 460 ng/ml) mixed with the FEU tests (with cutoffs at 500 ng/ml, 1000 ng/ml).
PEGeD study - Nov 2019
https://www.nejm.org/doi/full/10.1056/NEJMoa1909159
Summaries Dec 2019
https://www.thebottomline.org.uk/summaries/icm/peg-ed/
https://rebelem.com/peged-study-is-it-safe-to-adjust-the-d-dimer-threshold-for-clinical-probability/
https://rebelem.com/pregnancy-adapted-years-algorithm-for-pe-ready-for-prime-time/
All this to say, there seems to be a growing consensus that DDU and FEU can be used in place of each other.
In Summary
Major difference that ED physicians should be cognizant of when reviewing d-dimer lab results is whether or not the unit is D-dimer Units (DDU) or Fibrinogen Equivalent Units (FEU)
FEU = 2 * DDU
Most literature and algorithms are published using d-dimer assays that report FEUs.
If at an institution that uses DDUs, adjust your limits by 1/2.
For example, if the typical “negative” d-dimer is < 500 FEU ng / ml, the equivalent is < 250 250 ng / ml.
While the amount of literature directly addressing the DDU vs FEU issue is limited, ACEP Clinical Policy as of 2018 recommends age-adjusting DDU-based D-dimers.