Chromogenic and Clot-Based Bivalirudin Assays for Monitoring Anticoagulation

Abstract

Background:

Direct thrombin inhibitors (DTIs) are usually monitored with the activated partial thromboplastin time (aPTT) or activated clotting time (ACT). Both are complex assays with multiple enzymatic steps, and performance may be influenced by physiologic and pathologic factors unrelated to the DTI. Simpler systems, such as clot-based dilute thrombin time (dTT) and chromogenic anti-factor IIa assays, have been developed for monitoring DTIs, but there is limited data on their performance in clinical settings.

Methods:

Medical records of patients who received bivalirudin between March 2020 and April 2022 at a single institution were reviewed for demographic data and adverse outcomes. Plasma samples drawn for aPTT testing were analyzed with chromogenic anti-IIa and dTT bivalirudin assays. Results were compared to bivalirudin dosing.

Results:

Results of aPTT assays from 32 patients were compared with the chromogenic (n = 136) and dTT (n = 120) bivalirudin assays. Correlations between the aPTT and the chromogenic and dTT assays were poor (Spearman coefficients 0.55 and 0.62, respectively). There was a stronger correlation when results of the chromogenic and dTT assays were compared to each other (Spearman coefficient 0.92). When assay results were compared to bivalirudin dose, there were stronger correlations with the chromogenic and dTT assays than with the aPTT (Spearman coefficients 0.51, 0.63 and 0.22, respectively).

Conclusions:

There was considerable variation between results of specific bivalirudin assays and the aPTT. While bivalirudin assay results correlated better with administered drug dose, suggesting improving reliability, more studies are needed to determine if there is correlation between testing and clinical outcomes.

INTRODUCTION

Direct thrombin inhibitors (DTIs) are small-molecule anticoagulants that block the active site of the coagulation protease thrombin. They are used to treat or prevent thrombosis. DTIs currently used in clinical practice include parenterally administered bivalirudin and argatroban. While heparin and its analogs are the most widely used anticoagulants in inpatient settings, DTIs are emerging as important drugs in a growing number of clinical scenarios involving pediatric and adult patients, including the management of ventricular assist devices (VADs) in heart failure, extracorporeal membranous oxygenation (ECMO) support, and treatment of heparin-induced thrombocytopenia (HIT). In the multicenter Hirulog and Early Reperfusion or Occlusion-2 (HERO-2) trial involving adults with acute coronary syndrome, bivalirudin was associated with fewer adverse bleeding events and fewer episodes of reinfarction when compared to heparin (1). DTIs are often frontline therapy in patients with HIT, an important cause of in-hospital thrombosis. In pediatric patients with congenital heart disease with VADs, multicenter cohort studies have demonstrated that bivalirudin is as efficacious as heparin, while reducing bleeding risk (2). There is also limited data showing short-term mortality benefits and less bleeding and thrombosis with bivalirudin compared to heparin in pediatric patients on ECMO (3).

DTIs are usually monitored with 2 blood coagulation assays developed in the 1950s: the activated partial thromboplastin time (aPTT) and the activated clotting time (ACT). Both tests measure time to clot formation following induction of contact activation (4). Typically, DTI therapy is adjusted to prolong the clotting time into a predetermined range that is 1.5 to 2.5 times higher than the mean for the normal reference range for the assay (5). The strategy is predicated on the patient’s plasma coagulation mechanism being intact before anticoagulant administration. However, the aPTT (performed on cell-free plasma), and the ACT (performed on whole blood) are complex assays involving multiple enzymatic steps, and their performance may be altered by common clinical conditions including thrombosis (6), inflammation (through increased factor VIII activity), acquisition of non-specific inhibitors such as lupus anticoagulant, sepsis (7), disseminated intravascular coagulation, underlying bleeding disorders, and other illnesses (8). Furthermore, if plasma samples are collected from catheters or lines that have had anticoagulants run through them, the aPTT may be altered by anticoagulant contamination, confounding clinical interpretation. In critically ill patients, lower volumes of plasma are often collected and this creates an excess of citrate in collection tubes leading to suboptimal calcium concentration which leads to aberrant clotting time measurements; extremes of hematocrit or extremes of patient plasma calcium concentration can also cause abnormalities in aPTT (9). Additional monitoring complications arise because changes in renal function and hepatic function not only affect DTI metabolism (bivalirudin is primarily cleared renally, while argatroban is primarily cleared hepatically), but also the performances of the aPTT and ACT (10). Thus, there are numerous scenarios in which aPTT and ACT results will not accurately reflect the concentration of an anticoagulant in plasma. This leads to significant interpatient variability in testing that makes establishing therapeutic ranges for these assays difficult.

The issues outlined above are well recognized when using the aPTT and ACT to monitor the parenteral anticoagulant heparin. Chromogenic assays measuring plasma heparin concentration based on inhibition of factor Xa provide more accurate measurements of drug concentration (11). Such assays use cleavage of a chromogenic substrate rather than a measurement of clot formation as an end point (12). Experience with chromogenic assays for monitoring heparin demonstrate that there is often a poor correlation between the aPTT and plasma heparin level (13). Clinical outcomes, including bleeding and thrombosis rates and need for anticoagulant dose adjustments, tend to be better with the chromogenic heparin assay (14). There is a concern that treatment algorithms based on monitoring DTIs with the aPTT will also be inaccurate, particularly in situations where standard clotting assays are affected by disease processes (15).

With increasing use of DTIs, more reliable assays for monitoring therapeutic anticoagulation are required for safe use of these medications. Previous research has shown that results with a clot-based diluted thrombin time (dTT) assay correlate well with doses of DTI (16). This assay is less affected by increased D-dimer, decreased vitamin K-dependent coagulation factors, or lupus anticoagulants (17). Chromogenic assays that measure inhibition of thrombin activity (anti-IIa assays) have shown similar promise (especially with monitoring of the oral DTI dabigatran), although there are limited data from which to establish therapeutic ranges (18, 19). Both of these assays have been available for use for over a decade, but there are still very significant gaps in understanding their utility for clinical use of increasing scope. The pharmacokinetics of DTIs are not yet well understood in the pediatric population, and clinical monitoring of DTIs across the age spectrum, especially in critically ill patients, has not yet been optimized. The objectives of this study were to compare a chromogenic anti-IIa assay and a dTT assay to the aPTT for monitoring DTI therapy with bivalirudin.

MATERIALS AND METHODS

Institutional review board approval was obtained prior to research being conducted. Patients of any age who received bivalirudin between March 2020 and April 2022 at Vanderbilt University Medical Center were identified via pharmacy records and patient medical record review. Retrospective data collected included age, indication for bivalirudin use, baseline laboratory values for hemoglobin, platelet count, prothrombin time, aPTT, serum creatinine, and total bilirubin. Medical records were reviewed and bivalirudin doses were recorded for time points at which plasma samples were collected for standard aPTT monitoring of drug; there was inconsistent documentation in the medical record to determine if blood was drawn by venipuncture or from indwelling catheters. Records were reviewed for duration of bivalirudin treatment, documentation of bleeding- and thrombosis-related adverse outcomes, and administration of red blood cells, platelets, frozen plasma, or cryoprecipitate.

Plasma samples drawn for aPTT monitoring were stored at −80°C. It must be noted that the institution in this study has requirements for volume of blood draws; the concern of underfilled tubes causing abnormal aPTT results would thus not be observed, as such samples would not have been analyzed, and a new sample would have been requested. Samples were analyzed with a chromogenic anti-IIa assay (BIOPHEN^™ DTI, Hyphen BioMed) and a dTT assay (HEMOCLOT^™ Thrombin Inhibitor, Hyphen BioMed). The dilution factor for the anti-IIa assay is 1:2 and for the dTT assay 1:10. All assays were run on a Diagnostica Stago STA-R Max analyzer (Stago). Visually icteric or hemolyzed samples were included. In the chromogenic assay, a fixed amount of thrombin is added to patient plasma, and the rate of thrombin inhibition is determined by measuring residual thrombin activity with a chromogenic tripeptide conjugated to p-nitroaniline. Liberation of the p-nitroaniline leads to a change in color that is measured at 450 nm with a spectrophotometer. The values are inversely correlated with drug concentration. With the dTT, patient plasma is diluted with pooled plasma that is then mixed with a thrombin–calcium reagent. A drug level is extrapolated from the time to clot formation. Bivalirudin calibrators (BIOPHEN^™, Hyphen BioMed) were used to develop standard curves for both assays. Results were compared to aPTT measurements.

Statistical Considerations

A linear regression model was used to study the relationship between the chromogenic anti-IIa assay, dTT, and aPTT. A Wald test was used to test the null hypothesis that the slope of the model is equal to zero. Spearman correlations between the 2 variables were used to measure the strength of the association. Patient outcome analyses were descriptive. Adverse events were summarized using counts and proportions for categorical variables, and medians with interquartile range for continuous variables.

RESULTS

Thirty-two patients were included in the study. Three were in the pediatric age range (<18 years old), with one patient less than 1 year old. The remainder were adults. Of the 29 patients in the adult cohort, 25 were on anticoagulation for venous (n = 23) or arterial (n = 2) thrombi. Of patients with a thrombus, 9 were on bivalirudin for HIT confirmed with a serotonin release assay and another 9 were suspected of having HIT but subsequently had negative serotonin release assays. Two patients were on bivalirudin for suspected HIT without evidence of thrombus, 11 were on ECMO support, and one had a VAD. In the pediatric group, one patient with a venous thrombus was on bivalirudin for suspected HIT while on ECMO, while the other two had VADs placed while being evaluated for heart transplantation. Twenty-seven patients (one pediatric) were on heparin immediately prior to starting bivalirudin, one was on warfarin, and another was on apixaban. Baseline laboratory data (Table 1) reflect the level of illness and anticoagulation history in the cohort.

Table 1.

Mean laboratory values (with ranges in parentheses) in pediatric and adult patients immediately prior to initiation of bivalirudin for anticoagulation.^a

	Patients under 18 years of age (n = 3)	Patients over 18 years of age (n = 29)
Mean platelet count, /μL	112 000 (70 000-142 000)	118 000 (25 000-389 000)
Mean hemoglobin, g/dL	11.4 (7.7-14.9)	8.9 (6.3-11.8)
Mean partial thromboplastin time, s	42.9 (32.3-104.6)	56.3 (24.8-193.4)
Mean baseline partial thromboplastin time prior to any anticoagulation, s	30.3 (25.5-33.2)	29.6 (24.3-39.8)
Mean prothrombin time, s	14.7 (13.2-15.7)	15.7 (13.2-34.8)
Mean creatinine, mg/dL	1.00 (0.46-1.85)	1.19 (0.47-2.66)
Mean estimated glomerular filtration rate, mL/min/1.73 m2	80 (39-111)	70 (19-130)
Mean total bilirubin, mg/dL	1.73 (1.6-2.1)	0.97 (0.3-4.0)

^aThis table shows the average values of listed laboratory values right before bivalirudin initiation in the patient cohort, except where otherwise specified (baseline partial thromboplastin time). Overall, in the adult patients, anemia, and kidney injury were more common. In the pediatric patients, hyperbilirubinemia was more common.

One hundred and thirty-six unique plasma samples were tested with the chromogenic bivalirudin assay and 120 with the dTT assay. All 136 samples had an aPTT tested simultaneously. Of these samples, 10 were visibly icteric and 4 had visible evidence of hemolysis; because of the retrospective nature of sample analysis, plasma-free hemoglobin was not determined for visibly hemolyzed samples because it was not ordered by clinicians. The correlation between aPTT values and plasma bivalirudin levels determined by the chromogenic anti-IIa and dTT assays was poor (Figs. 1A and B; Spearman coefficients 0.55 and 0.62, respectively). In contrast, there was a strong linear correlation between results from the chromogenic and dTT assays (Fig. 2; Spearman coefficient 0.92). When comparing assay results to the dose of bivalirudin administered at each time point, there were stronger correlations between bivalirudin dose and results with the chromogenic anti-IIa (Fig. 3A, Spearman coefficient 0.51) and clot-based dTT (Fig. 3B, Spearman coefficient 0.63) assays than between drug dose and the aPTT (Fig. 3C, Spearman coefficient 0.22).

Fig. 1.

aPTT values plotted against bivalirudin levels determined by (A) a chromogenic anti-IIa assay (Spearman coefficient 0.55) or (B) a dTT assay (Spearman coefficient 0.63). Dotted lines were generated by linear regression analysis. For both panels, note the substantial range of aPTT values in samples that lacked bivalirudin.

Fig. 2.

Plasma bivalirudin levels as determine by a chromogenic anti-IIa assay plotted against the level measured by the clot-based dTT assay (Spearman coefficient 0.92). Levels of the chromogenic anti-IIa assay (independent axis) compared to levels of the dTT clot-based assay (dependent axis).

Fig. 3.

Results for (A), bivalirudin levels measured with a chromogenic anti-II assay, (B), bivalirudin levels measured with a clot based dTT assay, and (C), aPTT plotted against bivalirudin dose. The Spearman coefficients for the three plots are 0.51, 0.63 and 0.22, respectively.

The ranges of values for the chromogenic and dTT assay within the assigned therapeutic aPTT range (60 to 90 s) were also examined. The range of bivalirudin concentrations as determined by the chromogenic anti-IIa assay was 0.09 to 1.86 μg/ml, and for the dTT assay 0.05 to 2.07 μg/mL. The median aPTT was 68 s, while the median chromogenic anti-IIa bivalirudin level was 0.53 mg/μL and the median dTT bivalirudin level was 0.72 μg/mL. Median bivalirudin levels within groups with aPTTs of 60 to 70, 70–80, or 80–90 s were 0.46, 0.53, and 0.59 μg/mL, respectively, for the chromogenic anti-IIa assay (Fig. 4A), and 0.59, 0.75, and 1.26 mg/μL respectively, for the dTT assay (Fig. 4B). Thus, while aPTT values correlate rather poorly with bivalirudin measurements (Fig. 1A and B), there is a rough overall correlation between the aPTT and the bivalirudin levels.

Fig. 4.

aPTT values between 60 and 90 s (the assigned therapeutic range) were divided into tertiles (60 to 70, 70 to 80, and 80 to 90 s), and the median aPTT for each group was determined (66, 73, and 86 s, respectively). Results are plotted against the median bivalirudin concentration for each group determined by (A), a chromogenic anti-IIa assay (Spearman coefficient 0.96) or (B), a clot-based dTT assay (Spearman coefficient 0.99).

Fifteen patients in the cohort had adverse outcomes while receiving bivalirudin. Nine (29%) had bleeding, 4 (13%) died, and 2 (6%) had a venous or arterial thrombus. We did not have samples available from the time of these events to measure drug concentration levels, but the aPTT values preceding bleeding events ranged from 44 to 91 s. All patients who had bleeding events were placed back on bivalirudin with lower aPTT goals. In most cases, outside of bleeding episodes, transfusions of red blood cells, platelets, plasma, and cryoprecipitate were given according to standard protocols to ensure clinical stability while on ECMO. There was a total of 198 red blood cell, 12 platelet, 18 plasma, and one cryoprecipitate transfusions.

DISCUSSION

We observed better correlations between results for the anti-IIa and dTT bivalirudin assays than between either assay and the aPTT across a range of patient ages and clinical conditions. There was better agreement between the bivalirudin assays and the amount of drug administered. For several samples, aPTT values were less than 40 s despite chart reviews indicating patients were receiving bivalirudin at those times. There were also several samples where the aPTT was prolonged when the drug was not being administered. Examples of marked discordance between aPTT and drug levels are well known when monitoring heparin therapy with the aPTT, and likely reflect the sensitivity of the aPTT to multiple factors that influence the rate of clot formation independent of the effect of a therapeutic anticoagulant. Because a plateau effect was not observed for the aPTT (results ranged from 34.1 to 160 s for patients with charted bivalirudin dosing), this discordance was quite clear. The aPTT assay used in this study has an upper limit of 200 s, and no samples reached this level in the study cohort. Furthermore, in our cohort, the highest recorded bivalirudin dose was 0.8 mg/kg/h and the aPTT at this instance was 75.8 s, while the chromogenic assay resulted in 1.81 mg/μL and the dTT clot-based assay resulted in 1.66 mg/μL. The dose noted when the aPTT was at its highest at 160 s was 0.14 mg/kg/h. This study population therefore does not provide insight into whether a dose-dependent plateau exists for aPTT. We noted a high degree of interpatient variability in correlation between aPTT levels and bivalirudin dose. For one patient, the Spearman coefficient between aPTT level and bivalirudin dose was 0.76 (11 data points), while for another, the coefficient was 0.11 (13 data points). This emphasizes that the aPTT does not provide the same degree of reliability for anticoagulant monitoring for each patient. However, as noted previously, there is quite a degree of variability between the chromogenic and dTT assays (when looking at the median values for each assay [0.59 μg/mL for chromogenic, 1.26 mg/μL for dTT] compared to aPTT ranges from 80 to 90 s, for example) in analysis of this cohort. With the current clinical data available, it cannot be clearly stated from this study if the newer assays are complementary to each other or if they are equally as useful. More clinical and laboratory data in future work is needed to discern the clinical utility of these assays.

While there were stronger correlations between bivalirudin levels and the amount of drug administered, we noted that the chromogenic (anti-IIa) and clot-based (dTT) assays gave somewhat different dose responses. The values for plasma drug levels tended to be higher with the dTT assay than with the anti-IIa assay, despite the use of common controls. The reasons for this are not clear but may reflect fundamental differences between chromogenic and clot-based assays. The situation is reminiscent of the differences in therapeutic plasma heparin ranges reported in the literature for factor Xa-based chromogenic assays and protamine sulfate titration assays. This was highlighted when analyzing results for individual patients. For a 47-year-old male with a pulmonary embolus requiring ECMO for severe COVID-19–related illness, the Spearman coefficient for comparisons of assay results and drug dosing (13 data points) was strongest for the dTT (0.6), weaker with the chromogenic anti-IIa assay (0.25), and weakest for aPTT (0.1). A similar pattern was noted in a 60-year-old male with myocardial infarction and HIT (0.51, 0.42, 0.13, respectively, 9 data points), but the difference between the dTT and anti-IIa assays was less drastic. Such variability between assays was noted in other patients and did not seem to be related to differences in organ function or specific clinical conditions. It should be noted that decisions regarding dose titration were made based on aPTT results. It will be interesting to see if future studies using bivalirudin assays to adjust dosing alters correlations.

While the results of this analysis seem to favor using the drug-specific assays over the aPTT for therapeutic monitoring of patients on bivalirudin, this small study could not address the issue of which test correlates best with clinical outcome. Again, this is reminiscent of the situation with heparin therapy. While specific heparin assays give more accurate estimates of the amount of drug in plasma, it could be argued that clotting-based aPTT or ACT assays more directly demonstrate the consequences of a given drug concentration on the clotting mechanism of an individual patient. While there is some indication of superior performance for the chromogenic heparin assays, this is not a settled issue. At this point, in the absence of data regarding clinical outcomes, we cannot make strong recommendations regarding therapeutic ranges for plasma bivalirudin concentrations, or if a specific assay is preferable to another.

In principle, high plasma levels of bilirubin or significant hemolysis can interfere with chromogenic assays, while elevated levels of fibrin degradation products may interfere with clot-based assays (20, 21). The dilution factor of the anti-IIa assay was only 1:2, perhaps making this assay susceptible to bilirubin-related interference. However, the chromogenic anti-IIa assay failed to identify bivalirudin in one patient with severe liver disease because of a substance in plasma that gave a strong background signal at 450 nM; bivalirudin was measurable with the clot-based dTT assay. In our study the highest bilirubin concentration was 4.3 mg/dL, and only a few samples had visual evidence of hemolysis. Additional work is required to establish the sensitivities of the chromogenic and clot-based bivalirudin assays to various adverse testing conditions.

Our study has several limitations. First, it compares the bivalirudin assays to a single type of aPTT assay using a single type of coagulation analyzer. As there are many possible combinations of aPTT reagents and coagulation analyzers, correlations with bivalirudin assays may be different at different institutions. In addition, we were not able to establish if laboratory evidence of hemolysis correlated with assay discordance. As mentioned previously, it was not possible to obtain plasma-free hemoglobin on visibly hemolyzed samples and so correlation between this and assay discordance could not be ascertained. As mentioned before, there was no documentation to indicate that heparin contamination contributed to prolonged aPTT results, and so it is not known if this is a confounding influence of any discordance. Furthermore, data for the study were collected retrospectively from medical records and clinical outcomes and/or adverse outcomes may have been missed due to variation in documentation. The clinical outcome analysis in this study is, as a result, largely descriptive, as we usually did not have plasma samples drawn at the time of adverse bleeding or thrombotic events. Finally, the patient cohort in this study only included 3 patients in the pediatric age range due to a low number of patients having an indication to receive bivalirudin. Renal and hepatic function, especially as they relate to medication metabolization, evolve as patients grow older, as does the interplay between hemostasis and thrombosis. More data for these tests in pediatric patients is needed to understand if there are age-specific effects on assay reliability.

CONCLUSIONS

Use of the parenteral direct thrombin inhibitor bivalirudin is expanding in clinical practice. Drug administration is primarily monitored by aPTT testing, but there is little data pertaining to the sensitivities of different aPTT reagents to bivalirudin, or the range of conditions that can affect aPTT performance in patients on this drug. The chromogenic anti-IIa and clot-based dTT assays for bivalirudin are designed to determine plasma drug concentration using fewer enzymatic reactions than in the aPTT. While there is a better correlation between the newer assays and the amount of administered bivalirudin, more work is required to determine which assay correlates best with clinical outcomes and to establish therapeutic ranges for different patient populations. As direct thrombin inhibitors replace heparin for a wider range of clinical indications, better monitoring assays will be required to improve efficacy and safety.

IMPACT STATEMENT.

The paper describes the experience at a single institution with acutely ill patients treated with the anticoagulant bivalirudin. This drug is becoming popular as a replacement for heparin for a variety of indications. There is currently no standardized method for monitoring the drug, and most institutions default to the activated partial thromboplastin time in the absence of specific assays. The data in our manuscript suggest that specific assays for bivalirudin are a better method for monitoring.

Nonstandard Abbreviations:

DTI

direct thrombin inhibitor

aPTT

activated partial thromboplastin time

ACT

activated clotting time

dTT

diluted thrombin time

VAD

ventricular assist device

ECMO

extracorporeal membranous oxygenation

HIT

heparin-induced thrombocytopenia