Measuring direct oral anticoagulants—when, how

Physicians first performed a thromboelastogram (TEG) on this patient and said the results looked normal. Though the point of the case was not to discuss TEG parameters, Dr. Volod noted that the R parameter will affect the coagulation factor, “so anything that affects coagulation factors will potentially affect the R parameter,” she said. In case of factor/s deficiency or anticoagulant effect, R will be prolonged. In this case R was shortened, indicating an increased rate of thrombin generation. The patient also received thrombin concentrate, which also could potentially affect that parameter. The physicians asked, “Is it safe to take the patient to surgery, and is there a DOAC effect or not?”

Dr. Volod added a workup to measure PT, aPTT, and UFH level; she didn’t include an anti-Xa assay because Cedars-Sinai didn’t have apixaban or rivaroxaban assays at that time. (They were brought in-house because of this case and similar cases.) aPTT was normal; PT was above the reference range. UFH “showed a level [1.90 IU/mL] that potentially would be supratherapeutic, but we cannot apply this to the level for Xa inhibitors,” Dr. Volod said. “We can just say, ‘Based on these parameters, there is still the effect of rivaroxaban present.’”

Regarding Xa inhibitors, Dr. Volod raised three points. First, “we have to remember the viscoelastic assay main activator is kaolin, which is a contact surface activator that activates assays for aPTT and activated clotting time. And that is the reason the R parameter of TEG or the Intem parameter of ROTEM will be closer to the aPTT rather than PT and that abnormal clotting time may not exclude Xa inhibitor presence.” In this case the R was shortened, “and that can be misleading information,” she said.

Second, PT sensitivity is drug and reagent dependent. “In our institution, we use the Stago reagent and instruments, and based on personal experience and published literature, I know the sensitivity of the reagent we use in our laboratory.” Pathologists need to be aware of the sensitivity of their reagents to the different DOAC levels, she said.

Third, the unfractionated heparin and low-molecular-weight heparin anti-Xa assay can be used to detect or exclude drug presence. “By anti-Xa method, we can detect the presence of the drug but not the level of the drug,” she said. “For this you need an Xa assay calibrated with apixaban and rivaroxaban” (Billoir P, et al. Ann Pharmacother. 2019;53[4]:​341–347).

The following two cases highlight the potential interference of DOACs in other hemostasis and thrombosis laboratory tests.

In one case, a 54-year-old woman had been bedridden while treated for pneumonia. During her recovery, Dr. Moser said, she developed a PE, which prompted testing for lupus anticoagulant, among other things.

A DRVVT-based and an aPTT-based lupus anticoagulant test, both with screening, mixing, and confirmatory components as per current ISTH recommendations, were used. The PNP test used is a platelet neutralization procedure—an aPTT-based lupus anticoagulant confirmatory assay.

The patient’s PT result was prolonged (23 sec, RI 12.0–15.3 sec), “and that’s a little curious for a few reasons,” Dr. Moser said. One is that reagents tend to have a very high phospholipid concentration, “so we don’t typically expect lupus anticoagulants to prolong the PT to any significant degree,” she said. “Anytime I see that my radar is up, thinking about what else could be present in this patient plasma sample that’s causing that PT to be prolonged.” Lupus anticoagulant wouldn’t typically do that, she said, “so I’m wondering if something else is going on, and specifically if there’s an anticoagulant present.”

The DRVVT screen was prolonged (57 sec, RI 33–44 sec), “and there’s an inhibitory pattern that the mixing study doesn’t completely correct into our reference interval. But when we added back an increased phospholipid concentration in the DRVVT reaction, we weren’t able to demonstrate phospholipid dependence of this apparent inhibitor. So our confirmatory test came up as negative.”

The aPTT screen was also prolonged (61 sec, RI 32–48 sec). “Anytime we see prolonged aPTT screens we always wonder: Could this be heparin? A lot of us who work in thrombosis-hemostasis laboratories are conditioned to be suspicious for heparin, whether it’s heparin that’s therapeutic and expected to be present or heparin from a hep-lock IV line, or if there’s a line draw that wasn’t adequately flushed.” Thrombin time was normal, suggesting no presence of heparin or a direct thrombin inhibitor.

The aPTT-based mixing study “nearly but not completely” corrected (49 sec, RI 32–48 sec).

“In our initial aPTT confirmatory test, the platelet neutralization procedure also was negative,” she said. Another aPTT-based confirmatory test—hexagonal phospholipid neutralization—came up positive.

“Overall, this prolonged PT is kind of suspicious, making us think there could be some sort of drug present,” Dr. Moser said. “And it turned out there was.”

At the time the patient was tested, she was receiving rivaroxaban for her PE, and Dr. Moser suspected it was that causing the prolonged PT in their instrument-reagent system. “It also calls into question the lupus anticoagulant results” as potential false-positives.

The laboratory tested a follow-up specimen from the patient about 12 weeks later, for which her physician opted to pause the patient’s anticoagulant therapy. “Lo and behold, her PT was completely normal, as were her DRVVT and aPTT screens,” Dr. Moser said. “So that throws these initial results into question.”

While it’s possible the patient had a transient lupus anticoagulant detected initially and gone within 12 weeks, “the more likely scenario, given the clinical correlation, knowing that rivaroxaban was present initially, knowing that our prothrombin time was prolonged initially, is that the apparent lupus anticoagulant result we were picking up was likely an effect of the direct Xa inhibitor.”

Dr. Moser shared the next case for “contrast.” A 37-year-old man with an unprovoked PE underwent laboratory evaluation for various causes of thrombophilia, including lupus anticoagulant. This patient’s PT was normal, so there’s no clue of drug interference. His DRVVT screen was prolonged (91 sec, RI 33–44 sec). The mixing study remained prolonged (80 sec, RI 33–44 sec), “so it looked like an inhibitor effect,” she said. The DRVVT confirmatory test was also positive, and the aPTT screen was prolonged (119 sec, RI 32–48 sec).

The lab checked for heparin or direct thrombin inhibitor with thrombin time, which was normal, “so it doesn’t look like either heparin or direct thrombin inhibitor is present,” she said.

The aPTT mixing study didn’t correct (85 sec, RI 32–48 sec), “and our aPTT confirmatory assay, the platelet neutralization procedure, also looked positive,” Dr. Moser said.

Discussion with the ordering physician revealed that the patient had been on apixaban at the time of testing. “As Dr. Volod showed us, PTs are relatively insensitive to the presence of apixaban, and that was the case for our reagent in my laboratory.”

Of the lupus anticoagulant results, she asked, “Can we trust them in the presence of a direct Xa inhibitor?”

“It’s interesting that both the DRVVT and the aPTT-based testing came up as positive.” To be certain there’s no drug interference causing false results, it would be optimal, if possible, she said, to test while this patient is on low-molecular-weight heparin, if he transitions for a short time, or in the absence of anticoagulant if it would be safe to do so. There has also been work looking at drawing lupus anticoagulant samples at expected drug trough, Dr. Moser said. “So when the lowest level of, in this case, apixaban is present, to try to minimize interference. That could be another strategy.”

Dr. Moser requested but did not receive a follow-up sample from this patient to investigate whether lupus anticoagulant was present, so no follow-up data are available. But she contrasted this case with the prior case involving rivaroxaban in which the PT was prolonged. “As Dr. Volod showed, there’s significant variability, so understanding the performance expected from your PT and aPTT reagents is key to being able to interpret results in your laboratory.”

Table 2 sums up DOAC laboratory test interference, and Dr. Moser noted a few themes. “Rarely, false-negative results with lupus anticoagulants have been reported, mostly in the context of apixaban. That might be another point in favor of a potential real lupus anticoagulant” in the case of the 37-year-old male patient. In addition, PT- or aPTT-based factor assays can be falsely underestimated. And DOACs can interfere, too, in measurements of the natural anticoagulants antithrombin, protein C, and protein S. “In protein C and protein S activity, we’re expecting false overestimations with clot-based assays, so we run the risk of misclassifying a truly deficient patient as having a normal protein C or protein S level.”

Chromogenic protein C activity assays are not subject to this type of DOAC interference, she said, and free protein S antigen (or total S antigen) would show no effect from DOACs.

Antithrombin activity assays deserve special mention, she said, because the underlying assay design—whether factor Xa or IIa based—determines which of the DOACs could potentially interfere.

“Let’s take the case of direct Xa inhibitors,” Dr. Moser said. “If the antithrombin activity is factor Xa based, there’s a potential to overestimate antithrombin activity in the presence of a direct Xa inhibitor. But if the antithrombin activity is factor IIa based, we wouldn’t expect the direct Xa inhibitor to have a significant interfering effect. And that’s reversed for direct thrombin inhibitors.”

A few examples of hemostasis and thrombosis tests that are unaffected by DOACs are ELISA or latex immunoassay-based tests, such as D-dimer and von Willebrand antigen; DNA-based tests, such as those measuring prothrombin G20210A or factor V Leiden mutation; and fibrinogen activity using the Clauss method (Mani H. Int J Lab Hematol. 2014;36[3]:261–268; Gosselin RC, et al. Thromb Haemost. 2018;118[3]:437–450).

With the years of experience with heparin neutralization in the clinical laboratory, using compounds like heparinase or polybrene, one could ask, why not just neutralize the DOAC also?

The answer lies in an area of active investigation “not ready for prime time,” Dr. Moser said, “but there’s a growing body of literature in which laboratories are sharing their initial experiences. And this is something to keep an eye on in the coming years” (Siriez R, et al. Int J Lab Hematol. 2021;​43[1]:7–20; De Kesel PM, et al. J Thromb Haemost. 2020;18[8]:​2003–2017; Platton S, et al. Int J Lab Hematol. 2019;41[2]:227–233; Jilma-Stohlawetz P, et al. Int J Lab Hematol. 2021;​43​[2]:​318–323).

Two DOAC-neutralizing strategies—DOAC-Stop (Haematex Research, Australia) and DOAC-Remove (5-Diagnostics AG, Switzerland) are tablet-based compounds that incorporate activated carbon. DOAC Filter (Diagnostica Stago, France) is a filtration system. None of these is FDA approved, Dr. Moser said, and each assay has manual steps. “This would require careful consideration of how it would fit into a laboratory workflow.” And then there’s the added cost for the additional reagent and the time, and the possibility of needing additional patient plasma.

As for whether the strategies work, the preliminary, limited data “is promising,” Dr. Moser said. “Those neutralizing compounds appear to decrease interference in lupus anticoagulant testing, as well as some selected factor assays, activated protein C resistance, and antithrombin activity, based on data that have been published to date.”

Some groups report that DOAC-Stop can cause decreased factor activities in treated plasmas, she said, “so there is a question whether we are inadvertently absorbing some factors that we don’t mean to in addition to the direct oral anticoagulants.” The data are mixed, she said, but it’s a question worth considering: “Are we inadvertently taking something else out?”

“It’s an area wide open for investigation,” she said.

Amy Carpenter Aquino is CAP TODAY senior editor.