Reliability of the Heparin Management Test for Monitoring High Levels of Unfractionated Heparin: In Vitro Findings in Volunteersversus In Vivo Findings during Cardiopulmonary Bypass

Heparin concentrations were measured using the HMT (TAS Analyzer; Cardiovascular Diagnostics, Raleigh, NC; distributed as Rapidpoint Coag by Bayer Diagnostics, Mishawaka, IN). In this technique, a sample of approximately 30 μl citrated whole blood or plasma is spread along a large reaction chamber of the HMT test card. The surface of this chamber is coated with celite and small iron particles. The movement of the iron particles up and down in a magnetic field allows the transmission of light to a photodetector. The formation of a clot impairs this particle movement, which reduces light transmission. Coagulation is defined as the point at which light transmission is reduced to a predetermined percentage of the initial value. The measured time for clot formation is related to a heparin level by use of a previously generated standard curve.

With written informed consent, citrated whole blood samples (10 ml) were obtained from 10 healthy volunteers (classified as American Society of Anesthesiologists physical status I; 7 men, 3 women; mean age, 25 yr). The samples from five of the volunteers were used to construct a standard curve for the HMT using unfractionated heparin (UFH) to achieve blood concentrations over the range of 0–6 IU/ml (fig. 1). The relation between the HMT 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 (fig. 2A); dilution of platelets (200, 100, and 30 × 10³/μl), obtained by centrifugation of platelet-rich plasma and addition of platelet-poor plasma (fig. 2B); and dilution of procoagulants (50, 30, and 10% of the initial value), obtained via  substitution of platelet-poor plasma with corresponding volumes of a 5% solution of albumin (fig. 2C). In these studies, HMT was assessed two times at each of the seven heparin concentrations to include the reproducibility of measurements within an instrument. Thus, the standard curve was composed of 70 measurements, and the remaining studies were composed of 630 measurements (2 measurements per sample × 5 blood samples × 7 heparin concentrations) at three hematocrit values (n = 210;fig. 2A), three platelet concentrations (n = 210;fig. 2B), or three levels of procoagulants (n = 210;fig. 2C).

The reproducibility of the HMT measurements between instruments and their correlation to five predetermined heparin concentrations (0, 1, 2, 4, and 6 IU/ml) were assessed in the remaining five volunteers using two HMT analyzers in parallel. These studies were composed of 250 measurements (5 measurements per sample × 2 instruments × 5 blood samples × 5 heparin concentrations;table 1).

The influence of storage time on the measurements of the HMT (and their relation to four predetermined heparin concentrations) was evaluated by performing three measurements on the same sample with two HMT analyzers in parallel after 10, 30, 60, 90, and 120 min of storage (n = 120;table 2).

With written informed consent and approval of the local ethics committees, 60 patients undergoing cardiac surgery with CPB at either of two university hospitals participated. Forty of the patients were scheduled for elective coronary artery bypass grafting of one or two vessels (group I;table 3). The remaining 20 patients were scheduled for complex cardiac surgery (group II), including coronary artery bypass grafting in patients with impaired ventricular function (< 20%), combined coronary artery bypass grafting and valve replacement or repair, pulmonary thrombendarterectomy, thoracoabdominal aortic aneurysm repair, and heart and lung transplantation (tables 3 and 4). In this group, 10 of the patients were on therapy with coumadin (group II^cou); the remaining 10 patients were not (group II^ncou).

In the 40 patients of group I and the 10 patients of group II^ncou, preoperative antiaggregative therapy with salicylic acid was stopped 10 days before surgery and replaced with either a single subcutaneous bolus injection (7,500 IU) of UFH (n = 34 patients) or continuous infusion of UFH (n = 16 patients), according to values of the activated partial thromboplastin time of 40 to 60 s. In the 10 patients of group II^cou, preoperative anticoagulation consisted of coumadin, according to INR values of 1.8 to 2.5 (INR = international normalized ratio), which could not be replaced before surgery.

Premedication consisted of the oral administration of either midazolam, 0.05–0.1 mg/kg (group I), or clorazepate, 20 mg (group II). Induction of anesthesia was accomplished according to the standards of the two hospitals. Anesthesia was maintained using a total intravenous technique: sufentanil (0.05 μg · kg⁻¹· h⁻¹), propofol (100 μg · kg⁻¹· h⁻¹), and pancuronium bromide (supplementary if needed).

Surgery was performed under moderate (36°C rectal temperature, group I) or deeper hypothermia (≥ 28–30°C, 14 patients in group II) or deep hypothermic cardiac arrest (18°C, 6 patients in group II) (tables 3 and 4).

In 20 of the patients in group I (14 males, 6 females; mean age, 63 yr) and in the 20 patients in group II (13 males, 7 females; mean age, 68 yr), antifibrinolytic therapy was given using aprotinin. The therapy consisted of a 2 × 10⁶kallikrein inhibiting units (KIU) bolus given to the patient and a bolus of 2 × 10⁶KIU added to the priming solution, followed by a continuous infusion of 500,000 KIU/h.

In group I, systemic anticoagulation was achieved using UFH (350 IU/kg intravenous). Five minutes later, the ACT was measured (in seconds) using the Hemochron 401 device (International Technidyne Corporation, Edison, NJ). If the ACT exceeded 400 s, the extracorporeal circulation was initiated. Otherwise, an additional bolus of 5,000 IU of UFH was given. ACT values were obtained at intervals of 30 min; if the ACT value was less than 400 s, additional intravenous bolus injections of UFH (5,000–10,000 IU) were given. At the end of CPB, heparin was reversed with an initial fixed dose of protamine (0.8 mg per milligram of total heparin). In group II, the anticoagulation protocol was based on the automated heparin dose–response assay using the HepconHMS II device (Medtronic). According to departmental standards, an ACT value of 480 s was the predetermined target, and the necessary individual heparin concentration was calculated by means of the heparin titration cartridge of the heparin dose–response assay. Five minutes after injection of the initial individual bolus of UFH, both the ACT and the heparin level (as provided by the heparin dose–response assay) were controlled. During the course of the surgery, the heparin concentration was measured at intervals of 30 min and adjusted to the individual target according to the calculations of the heparin dose–response assay. By the end of CPB, the protamine dose was based on the residual heparin concentration. In both groups, infusion of fresh-frozen plasma and erythrocytes was based on clinical demands and departmental standards.

In all 60 patients, whole blood samples were obtained before heparinization (baseline), 5 min after the initial bolus of UFH, during CPB (30-min intervals), and 5 min after the administration of protamine. The samples were analyzed immediately using both the HMT (whole blood) and the chromogeneic standard laboratory measurement of the heparin anti-Xa activity (citrated plasma), which was obtained using our current anti-Xa analyzer (ACL 3000, Instrumentation Laboratory, Kirchheim, Germany). All measurements were performed in duplicate using two HMT analyzers in parallel.

The reproducibility of the HMT measurements (between and within instruments) and their correlation with heparin concentration were analyzed using multifactorial analysis of variance (StatView 4.02 for Macintosh; Abacus Systems, Berkeley, CA) and post hoc  analysis according to the Scheffé test; the correlation was obtained by polynomic (entire range of heparin concentrations) and linear regression analysis (slope of the relation for heparin concentrations over 0–3 and 4–6 IU/ml). The differences between the HMT measurements caused by selective variations in erythrocytes, platelets, and procoagulants were analyzed using the Wilcoxon test and multifactorial analysis of variance. For all studies, P < 0.05 was defined as statistically significant.

The correlation between the measured HMT values and heparin concentration over the entire range of heparin concentrations was r²= 0.98, indicating a strong relation between the two variables (figs. 1 and 2and tables 1 and 2). The slope of the curve relating HMT to heparin concentration was greater for heparin concentrations over the range 0–3 IU/ml (i.e. , linearity) than it was for heparin concentrations over the range 4–6 IU/ml (fig. 1and table 1). There was no significant difference between the measured values for HMT obtained in the two analyzers; the difference between the values obtained within a given analyzer did not exceed 4% of the mean value (figs. 1 and 2and tables 1 and 2), and the between-analyzer P  values varied between 0.58 and more than 0.99 (tables 1 and 2). The HMT values for hematocrits of 20 and 30% were similar to each other greater, but they were less than those for an hematocrit of 60% at heparin concentrations less than 3 IU/ml (fig. 2A). Selective dilution of platelets did not alter the HMT values (fig. 2B). Selective dilution of plasmatic procoagulants to 50 and 30% of baseline did not alter the HMT values, whereas dilution to 10% of baseline nearly doubled the HMT values across the entire range of heparin concentrations (fig. 2C). The HMT values were not affected by storage time (table 2).

In the 40 patients in group I, CPB did not exceed 90 min. Systemic anticoagulation required heparin doses from 25,000 to 35,000 IU, resulting in heparin concentrations during CPB of 3.2–4.7 IU/ml, assessed by the reference anti-Xa procedure (table 3). The total heparin requirement ranged from 26,200 to 164,000 IU. The corresponding HMT values averaged 146–155 s (baseline) and 326–422 s during CPB, and they correlated well with the anti-Xa values in the absence and presence of aprotinin (r²= 0.96 and 0.97, respectively;table 3). At the termination of CPB, coagulation time ranged from 168 to 208 s, despite total reversal of heparin (table 3). The difference between HMT values obtained from the two HMT analyzers never exceeded 4% of the mean value; the between-analyzer P  values were between 0.55 and more than 0.99.

In the 10 patients in group II^ncou, CPB duration was between 190 and 270 min (227.5 ± 27.5 min). Systemic anticoagulation required heparin doses from 27,000 to 55,000 IU, resulting in heparin concentrations during CPB of 4.1–10.2 IU/ml, assessed by the heparin dose–response assay (table 3). The total heparin requirement ranged from 65,000 to 145,000 IU. The corresponding HMT values averaged 143–179 s (baseline) and 451–790 s during CPB, and they correlated well with the anti-Xa reference (r²= 0.89;table 3). At the termination of CPB, coagulation time ranged from 79 to 203 s, despite total reversal of heparin (table 3). The difference in HMT values obtained from the two HMT analyzers never exceeded 4%; the between-analyzer values varied between 0.54 and more than 0.98.

In the 10 patients in group II^cou, CPB duration was between 190 and 320 min (242 ± 34.4 min). The amount of heparin necessary to achieve systemic anticoagulation ranged from 25,000 to 55,000 IU. The total heparin requirement during CPB ranged between 42,000 and 125,000 IU, producing heparin levels during CPB of 3.7–10.5 IU/ml, assessed by heparin dose–response assay (table 4). The corresponding HMT values averaged 144–167 s (baseline) and 385–768 s during CPB. The difference between the duplicate measurements within a HMT analyzer never exceeded 4% of the mean, but the difference between HMT values obtained from the two analyzers ranged from 0 to 47%. The correlation of the two HMT devices with the anti-Xa control within each patient covered a relatively wide range (r²= 0.57–0.94), although the mean correlation of each of the two analyzers (HMT vs.  anti—Xa) in the total of 10 patients was acceptable (r²= 0.79 and 0.87, respectively) (table 4). After reversal of heparin, the reproducibility was comparable to that found in vitro  and in groups I and II^ncou, despite a wider range in coagulation time of 75–210 s (table 4).

The HMT is an extension of the TAS Analyzer line of point-of-care anticoagulation testing, that is, activated partial thromboplastin time 22and ecarin clotting time. 23Although the HMT originally was designed for the requirements of the cardiac catheter laboratory, preliminary studies have reported successful use of the HMT during CPB in adults. 3Recent studies, addressing the suitability during standard CPB, 7,10,11have reported similar 11or better 7,10performance of the HMT in vitro  and in vivo  compared with the ACT. The current study extended this previous work by assessing under well-controlled conditions in vitro  the influence of variables common to CPB, e.g. , dilution of erythrocytes and platelets, on the HMT; and by evaluating the reliability of the HMT to assess the heparin activity during different conditions of CPB, such as standard and complex (prolonged) surgeries (CPB ≤ 90 min vs.  CPB ≤ 180 min), use or nonuse of aprotinin, different anticoagulation protocols during perfusion (ACT vs.  automated protamine titration and heparin dose–response assay), and preoperative use or nonuse of coumadin.

Our in vitro  studies indicated that the HMT values were reliable and reproducible (within and between instruments) and were independent of storage time (up to 120 min). The relation between the HMT values and heparin concentration was strong across the entire range of heparin concentrations (i.e. , a definite relation between HMT and predetermined heparin levels), but linearity was restricted to heparin concentrations less than 4 IU/ml (fig. 1and table 1). Although a concentration of 4 IU/ml UFH rarely is achieved during cardiologic intervention, heparin levels of more than 4 IU/ml often occur during CPB and would be easier to calculate if there was linearity over the entire range of heparin.

Reductions of hematocrit to 30 and 20% had no effect on HMT results, but an increase in hematocrit to 60% was associated with a marked prolongation of the coagulation time (fig. 2A). This effect may be relevant in patients with cyanotic heart disease, who often present with significantly increased hematocrit values. The HMT result was independent of decreases in platelet concentration to as low as 30 × 10³/μl (fig. 2B). It was also independent of decreases in procoagulant concentration to 50 and 30% of baseline, although decreases to 10% of baseline nearly doubled the HMT values (fig. 2C). Clinically, this effect could lead to underheparinization during and overprotaminization following CPB, and to clotting within the bypass circuits following the administration of coagulation factors via  transfusion of fresh-frozen plasma, even without the administration of protamine.

During standard CPB with a fixed regimen of anticoagulation controlled by monitoring of the ACT (group I), the HMT technique also showed good reproducibility and reliability (table 3). In keeping with previous data, 3,7,10,11there was a strong correlation between the HMT values and the chromogeneic heparin anti-Xa activity (laboratory reference) during the entire period of extracorporeal circulation. In contrast to findings obtained with the celite-based ACT, 3,6,17,18,22this correlation was not affected by aprotinin.

The HMT method also revealed good reproducibility and reliability (table 3) even under the conditions of complex cardiac surgery and global alterations of the coagulation system. This included an extended CPB of 180 min or more in patients without preoperative therapy with coumadin, hypothermia (≥ 28–30°C), marked hemodilution (tolerated hematocrit ≈ 20%), and maintenance of the heparin levels according to the automated heparin dose–response assay compared with patients undergoing standard CPB (group II^ncouvs.  group I).

The impaired interinstrument reproducibility of the HMT values in the presence of coumadin (group II^couvs.  group II^ncouand group I) is another noteworthy finding (tables 3 and 4). Figure 3demonstrates the poor reproducibility in the duplicate HMT values and anti-Xa measurements plotted separately for 3 of the 10 patients in group II^cou. This finding may be explained by the alterations in the coagulation system caused by coumadin, which were aggravated by extended perfusion times (up to 320 min), the hypothermia (18–30°C), low hematocrits (≈ 20%), and maintenance of heparin levels according to the automated heparin dose-response assay. The baseline values for the HMT were comparable to group I and group II^ncou, and the reproducibility of the HMT measurements improved after total reversal of heparin by the end of CPB (anti-Xa = 0). After total reversal of heparin, the difference between the two instruments decreased to 4%, although the HMT values after heparin reversal showed significant variation (between 75 and 210 s;fig. 3, table 4). For example, in patient 2 (fig. 3), the overestimation of blood heparin level could prompt the use of excessive protamine to restore clotting (despite total reversal of heparin). Because protamine has direct anticoagulant influence on platelets and proteins, including fibrinogen, 1,16,18this could further aggravate post-CPB disorders of the hemostatic system. 9,19The findings in group II^couare consistent with the results obtained in vitro , demonstrating that a depletion of the procoagulants to less than 10% causes a marked prolongation even of the baseline HMT value.

The current study suggests the need for improvement in the reliability of the HMT. This may be achieved by the use of a more potent stimulator of coagulation in manufacturing of the HMT card (i.e. , kaolin in place of celite) or coating of the card with lyophilized plasma.

The hardware may require additional features for routine clinical use. Although the HMT has satisfactory electronic quality control and several notable assets, including simplicity, compact size, and silence of operation, the absence of a digital display of clotting time is a disadvantage. As advocated for the ACT, 12the capability for duplicate HMT measurements, i.e. , two channels, is most desirable.

In conclusion, our findings suggest that the HMT (originally developed for monitoring in the catheterization laboratory) is well-suited for use in standard and even complex surgeries with prolonged CPB of 180 min or more, including deep hypothermia and use of aprotinin. The current results, however, also suggest that the HMT test card is not suitable as the sole criterion for therapeutic decisions during complex surgery with extended CPB in patients preoperatively treated with coumadin. In these patients, further studies must provide evidence that anticoagulation according to a specific range for the HMT ensures safety compared with the anticoagulation protocol based on the kaolin-based ACT (400–480 s).

The authors thank Tonie Derwent for reviewing and editing the manuscript and Annette Gaussmann for performing statistical analysis and editing the figures (German Heart Center, Department of Anesthesiology, Berlin, Germany).