Anesthesia & Analgesia

Anticoagulation during cardiopulmonary bypass (CPB) is usually performed with unfractionated heparins (UFHs). However, despite profound anticoagulation, CPB still causes severe alterations in the coagulation system. Platelets and procoagulants are diluted by the priming solution of the CPB lines, consumed increasingly via contact activation on artificial surfaces, and showed contact activation activates the inflammatory response and fibrinolysis, all contributing to postoperative bleeding and transfusion requirements (^1,2).

Large-dose heparinization requires frequent, point-of-care monitoring of the anticoagulant effect. Global anticoagulation tests, such as the commonly used activated clotting time (ACT), reflect the more general coagulation status of the patient than the actual heparin concentration (^3–5). During longer perfusions, a continuing effect on the coagulation system can lead to a prolongation of the ACT, despite decreasing levels of UFHs (^1,2). There is evidence that the maintenance of larger heparin levels inhibits the system of contact activation and contributes to a reduced postoperative blood loss (⁶). Moreover, the reversal agent protamine itself reveals an anticoagulant effect. Excessive protamine dosing, due to overestimation of actual heparin level after conclusion of CPB, contributes to further damage of the coagulation system (⁷). Therefore, an anticoagulation protocol for CPB, based on measurement of the concentration of UFHs, effectively contributes to reduced blood loss and reduced transfusion requirements (^8–10) and appears to be desirable. However, the commercially available systems, such as the automated protamine titration assay (Hepcon HMS; Medtronic, Inc., Parker, CO), are controversial with reference to their reliability for precise measurement of heparin concentration (^11,12). Values of the Hepcon HMS are only valid if the heparin concentration is measured in the central channels of the cartridge. Coagulation detection in the left or right margin chamber requires repetition of the test with a smaller or a larger range cartridge to avoid over- or underestimation of the heparin level. This prolongs the time for the measurement and also contributes to the increased costs of the system.

Assessment of anti-Xa activity is the laboratory reference method for the measurement of heparins. The available assays require time-consuming preanalytical procedures and are not appropriate for point-of-care monitoring. We adapted the Heptest® (Accuclot™ Heptest®, Sigma Diagnostics, Deisenhofen, Germany) for measurement of plasmatic anti-Xa activity for performance in whole blood using the ACT II (Medtronic) device. We assessed the reliability of this test in vitro and in vivo during CPB.

Methods

Principle of Clot Detection in the ACT II Device

The ACT II device is designed for the measurement of the kaolin-ACT in prepared double channel cartridges. Kaolin is placed in the reagent chamber at the bottom of the cartridge. This chamber is made air-tight by a plastic flag, which is inserted above the reagent chamber. For measurement of ACT, the blood sample (150–200 μL) is placed over the flag-closed chamber and coagulation measured after an automated incubation period, which includes warming the sample to 37°C. Detection of coagulation is achieved by an optical/mechanical method. Flags are moved up and down in the cartridge to provide mixing of the stimulator and the blood. The velocity of the fall of the flags is detected by an optical sensor. Reduction of the velocity of the fall to a defined value is recognized as coagulation. The coagulation time for each channel, the mean value, and the difference between the two measurements are visualized in a display.

Performance of Whole Blood Heptest® on the ACT II Device

Recalcification of the citrated blood was achieved by transferring 75 μL of the Recalmix solution of the Heptest® kit into the reaction chamber of the ACT II cartridges and inserting the flag. Then, 75 μL of the citrated whole blood sample was placed in the cartridge. Thereafter, 75 μL of the Factor Xa reagent was added, and the test was started after the automated incubation period.

Preliminary Investigations and Determination of the Optimal Reagent

Concentrations of UFHs, which are necessary to achieve sufficient anticoagulation during CPB, normally vary from patient to patient. Therefore, the initial goal was to design an assay that would reveal linearity over a wide range of UFHs. In preliminary investigations using the same method as described later, the exclusive use of the large-range Heptest® (Heptest® HI) demonstrated linearity up to 8 IU/mL UFH. However, spread of the coagulation times between 0 and 8 IU/mL UFH was reasonably low (130 s) and correlation to the chromogenic assay was poor (r = 0.56).

In the case of UFH concentrations below 4 IU/mL, this was attributed to so much excess available Factor Xa that the differences in UFH levels cannot be discriminated.

The optimal reagent for the assay was determined. The Heptest® HI reagent was diluted with predetermined volumes of the small-range test kit (Heptest®). A ratio of 1:3 of the Heptest® HI and Heptest® LOW revealed linearity to concentrations of up to 6 IU/mL UFH and spread the measuring range to a 300-s gap between concentrations of 0.5–8 IU/mL UFH. This was regarded as optimal for precise differentiation of the concentrations of UFH and fast acquisition of test results.

In Vitro Investigations

With written informed consent, citrated whole blood samples (10 ml) were obtained from 10 healthy volunteers (7 male and 3 female students, mean age 25 yr). Samples from 10 volunteers were used to establish a standard calibration curve for UFHs (Liquemin, Hoffman-La Roche, Grenzach-Wyhlen, Germany).

The relationship between ACT II anti-Xa-UFH assay measurement and each heparin concentration was assessed under the following conditions:

• Variation in hematocrit (20%, 30%, and 60%), obtained by centrifugation and adjustment of the plasma fraction.
• Dilution of platelets (100, 50, and 20 × 103/μL), obtained by centrifugation of platelet-rich plasma and addition of platelet-poor plasma.
• Dilution of procoagulants (50%, 30%, and 20% of the initial value) obtained via substitution of platelet-poor plasma with corresponding volumes of a 5% solution of albumin. All measurements were performed in duplicate by the use of a two-channel cartridge.

The within-assay variance was measured using two different concentrations (2 and 6 IU/mL) of UFHs. Each measurement was performed 10 times.

Accordingly, the influence of storage time at room temperature (18°–24°C) was evaluated by repetition of the measurements (mean value of measurement in duplicates) after 30, 60, 120, 240, and 480 min.

In Vivo Investigations

After written informed consent and approval by the local ethics committee, samples were collected from 15 patients who underwent cardiovascular surgery that involved CPB.

According to our departmental protocol, preoperative anticoagulation with salicylic acid was stopped 10 days before surgery and replaced with either a single subcutaneous bolus injection (7500 IU) of UFH (n = 8 patients) or continuous infusion of UFH (n = 5 patients), according to values of the activated partial thromboplastin time of 40–60 s. Two patients had been treated with warfarin to an international normalized ratio value of 1.8.

Anesthesia was performed according to the departmental standard using a total IV technique (propofol, sufentanil, and pancuronium bromide). CPB was performed with standard equipment, such as roller pumps and membrane oxygenators. All components were nonheparin-coated for the CPB system.

The platelet count, fibrinogen level, and hematocrit were performed in parallel to each measurement.

Samples were collected before and after the initial UFH bolus. Thereafter, samples were collected at intervals of 30 min during perfusion and after protamine administration. The anticoagulation protocol was performed according to the departmental standard: an ACT value of 480 s was the predetermined target, and the necessary individual heparin concentration to achieve this ACT value was calculated by means of the heparin titration cartridge of the Hepcon HMS device. Five minutes after injection of the initial individual bolus of UFH, both the ACT and heparin level (as provided by the Hepcon HMS device) were measured. During a further course of surgery, heparin concentration was measured at intervals of 30 min and adjusted to the individual target value according to the calculations of the Hepcon HMS device. Anticoagulation was exclusively guided by maintenance of heparin concentration without further measurements of ACT. By the end of CPB, the protamine dose was based on the residual heparin concentration. In all patients, aprotinin (Antagosan; Hoechst, Frankfurt, Germany) was administered with a bolus of 2 × 106 kallikrein inhibiting units (KIU) for the patient, 2 × 106 KIU into the priming solution of the CPB, and a constant infusion of 0.5 × 106 KIU/h during extracorporeal circulation.

Chromogenic Anti-Xa Assay

The chromogenic assay was performed on an STA analyzer (Roche Diagnostics, Mannheim, Germany) that used the STA LMWH test kit (Roche Diagnostics). Fifty microliters of the sample were diluted with 50 μL of diluent buffer and 100 μL of Factor Xa, and 100 μL of substrate was added. For kinetic measurement of the remaining Factor Xa, the change in optical density was assessed at 405 nm.

Statistical Analysis

Correlation between the results of the plasmatic chromogenic test and the ACT II anti-Xa assay was analyzed using Pearson’s correlation coefficient. In vitro data were analyzed using analysis of variance and the Scheffé test. A P < 0.01 was considered significant.

Results

In Vitro Investigations

The calibration curve was established over the range 0–8 IU/mL (Fig. 1). The ACT II anti-Xa assay revealed a linearity up to concentrations of 6 IU/mL UFH (Fig. 1). The variation coefficient (VC) between the results obtained from each individual at a defined concentration was calculated with a mean of 10.5 ± 2.83 for all measured concentrations and a mean of 9.08 ± 0.77 within the linear range of 0–6 IU/mL UFH (Fig. 1).

Serial imprecision was calculated by the VC obtained by 10 time measurements of a sample using two different concentrations of UFH. VC for a concentration of 2 IU/mL UFH was 3.4 and for a concentration of 6 IU/mL UFH it was 2.0 (Table 1).

Storage of the prepared cartridges for 30–480 min did not impair the results as calculated by a VC of 2.3 for UFH concentrations of 6 IU/mL and a VC of 1.1 for concentrations of 6 IU/mL (Table 2).

The assay was not significantly influenced by the variations in platelet counts, plasmatic coagulation factors, and variation of the hematocrit to 20% and 30%. A hematocrit value of 60% led to a prolongation of the coagulation time, particularly with the larger UFH concentrations (Fig. 2A–C).

In Vivo Investigations

Surgery included coronary artery bypass grafting (n = 4), mitral and aortic valve replacement (n = 2), combined valve replacement and coronary artery bypass grafting (n = 5), repair of an aneurysm of the ascending thoracic aorta (n = 2), heart transplantation (n = 1), and repair of an aneurysm of the thoracic-abdominal aorta (n = 2).

The duration of CPB ranged from 85 to 321 min, with a mean of 137 ± 43 min. Surgery was performed in normothermia (n = 4), moderate hypothermia (32°–30°C, n = 9), and deep hypothermia (16°C, n = 2).

Platelet counts ranged from 243 to 24 ± 35 × 106/μL with a mean of 129 ± 23 × 106/μL the fibrinogen levels ranged from 54 to 250 mg/dL with a mean of 154 ± 31 mg/dL and the hematocrit values from 37% to 22% with a mean of 27% ± 3.6%.

Heparin concentrations obtained by the chromogenic method varied from 0.0 to 8.9 IU/mL with a mean of 3.2 ± 2.70 IU/mL. Values of the ACT anti-Xa UFH assay ranged from 22 to 425 s with a mean of 161 ± 104 s. The correlation coefficient of the ACT II anti-Xa UFH assay to the chromogenic method was 0.9 (Fig. 3).

Discussion

Measurement of the anti-Xa activity is the standard laboratory reference assay for monitoring UFH. Both laboratory anti-Xa assays, the Heptest® and the chromogenic test are performed in plasma and, therefore, are unsuitable for use at the site of patient care. However, online monitoring of anticoagulation is necessary during cardiovascular surgery that uses CPB. We developed an anti-Xa-based whole blood assay that can be easily performed with blank cartridges of the ACT II device, even by untrained personal. This ACT II anti-Xa-UFH assay performed in whole blood provided quick and reliable monitoring of the heparin concentration during CPB.

In the in vitro part of the investigation, the test revealed excellent reproducibility and linearity up to concentrations of 6 IU/mL UFH. The interindividual variation of the coagulation times for defined heparin concentrations exceeding 5% suggests that a preoperative individual calibration is mandatory. However, further investigations with a larger number of patients are necessary for the final evaluation of this subject.

Test results were nearly uninfluenced by the dilution to a hematocrit of 20% of platelets and procoagulants as often observed during CPB. However, an increase of the hematocrit to 60% led to a prolongation of the coagulation time. This effect was pronounced in the larger concentrations of UFH and can be explained by the larger concentration of UFH in the correspondingly smaller plasma volume.

In the in vivo part of the investigation, the test provided a close correlation (r = 0.90) to the values of the chromogenic reference method. This is noteworthy because the CPB procedures included extreme conditions of perfusion, such as profound deep hypothermia (16°C) and long-term CPB (>3 h) with depletion of platelets (<20,000/μL) and fibrinogen (<50 mg/dL), and confirmed the results obtained in the in vitro set-up. In order to avoid further imprecision by additional measurement of the hematocrit value, which would have been necessary, the values of the plasmatic chromogenic assay were not corrected to the hematocrit values of whole blood. Therefore, the values of the plasmatic chromogenic assay, in contrast to the constant volume of the whole blood, are influenced by the hematocrit as it determines the volume in which UFHs are dissolved. Apart from the individual response to heparins as influenced, for example, by cell binding or binding to platelet factor 4 in the sample, these variations of the hematocrit (37%–22%) explain the scatter along the correlation curve in Fig. 3. Consequently, a correlation coefficient of 0.9 can be considered as satisfactory.

The ACT II device appears to be preferable to other single channel options for the adaptation of other more specific coagulation tests because: first, the tests are performed in double cartridges so that duplicate measurements can be performed in parallel. Second, the tests can be easily performed because of the air-tight reaction chamber and stability of the reagent allows for prepreparation of the cartridges that effectively reduces the operating steps in the operation room. Moreover, the automated incubation period contributes to the further ease of handling. Third, the small-range cartridge only needs a volume of approximately 150–200 μL in comparison with the 2 mL needed for most other devices. Because the reagents are major contributors to the cost of the assay, the reduction of the volumes necessary for the test effectively reduced the costs.

The ACT II anti-Xa-UFH assay is in the series of tests transformed to the ACT II device for monitoring inhibitors of plasmatic coagulation, such as low molecular weight heparins, heparinoids, and direct thrombin inhibitors like recombinant-hirudin (^13,14). Therefore, current clinically important anticoagulants can be monitored with one device at the site of patient care.

However, although the test is easy and quick to perform, the need for preanalytical construction of an individual calibration curve and for precise pipetting steps in the operating room limit the value of the assay in its current form. Therefore, before routine use in clinical practice is practical, further automation of the assay is needed. In addition, further clinical investigations with a larger number of patients are necessary for the final evaluation of the assay.

We thank Miss Tonie Derwent for editorial assistance on the manuscript and Miss Annette Gaussmann and Mr. Helge Hasselbach for the figures. We are indebted to Miss Cornelia Harke and Mr. Gerhardt Kettelgerdes for the performance of laboratory tests.

References