The Effects of Aprotinin on Thromboelastography with Three Different Activators

Approval for the study was obtained from the King’s College Hospital Ethics Committee (London, United Kingdom).

Blood samples were drawn from 45 healthy volunteers and 36 adult patients undergoing elective cardiac surgery at King’s College Hospital. Volunteers were doctors and nurses who agreed to donate blood specimens for this study. Patients gave verbal consent for thromboelastography tests to be performed on their blood at the same time as blood was drawn for other routine blood tests. Patients with known coagulopathies, those taking anticoagulants, and those with liver or renal dysfunction were excluded from the study.

Blood was drawn into tubes containing sodium citrate (0.105 m; ratio, 9:1). After thromboelastography tests had been performed, the blood was centrifuged at 2,000 g  to obtain platelet-poor plasma. Samples were stored in aliquots at −70°C until the time of assay.

Blood was drawn from 26 patients who received aprotinin as part of their routine management. Blood samples were drawn from a free-flowing, 14-gauge central catheter immediately before and 5 min after aprotinin administration. This was before surgery and heparin administration. Thromboelastography tests were performed using fresh blood.

Blood was drawn from 10 patients into a tube containing sodium citrate 1 h after protamine administration after CPB. These patients had not received aprotinin.

The international normalized ratio (INR), APTR, and fibrinogen values were obtained using a nephelometric method on an ACL 300R^®machine (Instrumentation Laboratory, Warrington, Cheshire, United Kingdom).

Celite and kaolin were obtained in powder form. A solution of 1 g/100 ml (1%) was mixed using normal saline to ensure that there was sufficient quantity for the whole study. The solutions were stored in a refrigerator at 2–8°C in 1.5-ml plastic containers with an aliquot of 37 μl kaolin or celite. At the time of testing, the plastic containers were warmed to room temperature, and 1 ml blood was added, giving final celite or kaolin concentrations of 0.357 mg/ml blood. This is the celite concentration recommended by the manufacturer. There are no previous reports of kaolin used as an activator for thromboelastography; initial dose–response studies (data not shown) suggested that this concentration produced a significant shortening of the R time compared with the “native” trace but retained sensitivity.

Recombinant TF has been used as an activator for thromboelastography in a previous study. 13Recombinant TF (Innovin^®; Dade Behring, Atterbury, Milton Keynes, United Kingdom) was diluted to a concentration of 1:120. Thirty-seven microliters of this solution was decanted into 1.5-ml plastic containers, which were stored at −20°C and warmed to room temperature before the thromboelastography test. One milliliter of blood was added to the containers to yield an equivalent final TF concentration of 0.3 μl/ml blood.

Two dual-channel thromboelastography machines (TEG^®; Haemoscope, Niles, IL) were used in parallel, connected to a notebook computer. The machines underwent regular quality control and calibration. All thromboelastography measurements were performed with the machines prewarmed to 37°C. The parameters recorded included R time, α angle, maximum amplitude (MA), and lysis index at 30 min (percentage reduction of MA at 30 min after the MA). R represents the time to initial clot formation, α angle relates to rate of clot extension, MA relates to clot strength, and the lysis indices reflect the extent of fibrinolysis present (fig. 1).

Citrated blood samples were analyzed 60 min after venepuncture. One milliliter sodium-citrated blood was added to a container holding 37 μl of activator (celite, kaolin, or TF). This was mixed, and then 340 μl was transferred into the thromboelastography cuvette to which 20 μl calcium chloride, 0.1 m, had been added.

Fresh blood was placed into a sterile plastic container. After 4 min, 1 ml blood was added to a container holding 37 μl of activator (celite, kaolin, or recombinant TF). This was mixed, and then 360 μl was transferred into the thromboelastography cuvette.

Thromboelastography tests were performed as described for healthy volunteers between 2 and 5 h after the blood was drawn.

Samples from 19 volunteers were run in duplicate with kaolin in two thromboelastography channels and celite in the other channels. Samples from 18 volunteers were run in duplicate with celite and TF as activators in two channels each. Aprotinin was added to blood samples to obtain a concentration of 200 KIU/ml. This blood was run in the third and fourth channels in each experiment. Thus, with each activator, there was a paired test with and without aprotinin.

Blood was drawn before and after aprotinin administration. Paired celite- and kaolin-activated thromboelastography tests were performed on samples from the first 11 patients, and celite- and TF-activated thromboelastography tests were performed on blood from the subsequent 15 patients.

Thromboelastography tests were performed using blood from eight volunteers with celite and TF as activators. Three channels were used for each test with aprotinin blood concentrations of 0, 100, and 200 KIU/ml, respectively, in each channel.

Blood from 10 patients after surgery was tested with the same methodology described in the previous paragraph.

The Wilcoxon signed rank test was used for paired comparisons. Results are expressed as median difference (95% confidence interval [CI]) and P  value for difference. Pearson and Spearman correlations were used where appropriate to evaluate association among parameters. Results are expressed as correlation coefficient (95% CI) and P  value for correlation). The Friedman test was used for analysis of variance. For agreement between tests, Altman and Bland bias plots and Passing-Bablok method comparisons were used. Bias with 95% CIs are reported. Data distribution was evaluated with the Shapiro-Wilk normality test. Statistical analysis was performed using Analyse-It^®Statistical Software (Analyse-It, Leeds, United Kingdom).

The thromboelastography R values were shorter when both kaolin and TF were used as activators compared with celite activation (P < 0.001). There was a strong correlation between kaolin- and celite-activated R values (r = 0.58 [CI = 0.27–0.78], P = 0.001) but not between celite- and TF-activated R values. Bias of 1.8 min (CI = 1.4–4.1) was detected between R values with celite and kaolin as activators. There was a small but consistent increase in the α angle with kaolin compared with celite (bias = 2°[CI = 1–3]) with a strong correlation between these values (r = 0.84 [CI = 0.68–0.92], P < 0.001). There was no significant difference in the celite and TF α angles, but there was also no correlation between these values. Similarly, there was a small increase in the MA with kaolin compared with celite (bias = 5 mm [CI = 2.3–7.6], P < 0.001) with a strong correlation between these values (r = 0.89 [CI = 0.79–0.95]). There was no significant difference or bias between the TF and celite MAs, and, unlike the R values and α angles, these values were strongly correlated (r = 0.79 [CI = 0.61–0.89], P < 0.001).

Aprotinin prolonged the R time of the thromboelastography trace (P < 0.001) when celite and kaolin were used as activators but not when TF was the activator (fig. 2and table 1) with both volunteers’ and patients’ blood. There remained a strong correlation between the celite and kaolin R values after aprotinin (r = 0.67 [CI = 0.4–0.83], P < 0.001). The celite and TF R values did not correlate after aprotinin. There was also a decrease in the celite α angles (median decrease =−2.5°[CI =−1.8 to −3.3], P < 0.001) and MAs (median decrease =−2 mm [CI =−1.3 to −3], P < 0.001) after aprotinin. The kaolin α angles did not decrease significantly, but the kaolin MAs were decreased (median decrease =−1.3 mm [CI = 0 to −2.5], P = 0.03) after aprotinin. TF α angles and MAs, like TF R times, were not significantly affected by aprotinin. There was a consistent prolongation of the APTR after aprotinin (median increase = 0.65 [CI = 0.6–0.7], P < 0.001). Fibrinogen concentrations and the INR were unaffected by aprotinin.

The aprotinin dose–response curves on blood tested from volunteers and from patients after CPB showed that at blood aprotinin concentrations of both 100 and 200 KIU/ml, aprotinin prolonged the R times of celite-activated thromboelastography (P < 0.05) but not TF-activated thromboelastography (fig. 3and table 1).

The celite-activated thromboelastography R values correlated with the APTR results (fig. 4) at baseline (r = 0.41 [CI = 0.12–0.64], P = 0.008) and after aprotinin (r = 0.53 [CI = 0.29–0.73], P < 0.001), but there was no correlation with the INR. The INR was not affected by aprotinin and correlated with the TF-activated thromboelastography R values (fig. 4) both before (r = 0.52 [CI = 0.2–0.74], P = 0.003) and after aprotinin administration (r =0.44 [CI = 0.1–0.68], P = 0.01). There was no correlation between the TF-activated thromboelastography R values and the APTR. There was a correlation between fibrinogen concentrations and celite α angles (r = 0.39 [CI = 0.09–0.62], P = 0.01) as well as celite MA (r = 0.8 [CI = 0.66–0.89], P < 0.001). Fibrinogen also correlated with TF MA (r = 0.78 [CI = 0.59–0.89], P < 0.001). The platelet count correlated with the celite MA (r = 0.67 [CI = 0.36–0.85], P < 0.001) and the TF MA (r = 0.75 [CI = 0.31–0.93], P < 0.001).

If thromboelastography is to be used with an algorithm to guide the use of blood components during surgery, then it is important to know whether pharmaceutical agents alter its trace. The results of this study show that when celite and kaolin are used as activators for thromboelastography in the presence of aprotinin, there is a prolongation of R time, a decrease in α angle, and a decrease in MA. Therefore, celite and kaolin may not be ideal activators for use in a thromboelastography-guided algorithm, in which prolongation of the R time is a trigger for fresh frozen plasma transfusion and decrease in MA is a trigger for platelet transfusion, 11in aprotinin-treated patients. TF is a cell surface glycoprotein that serves as the cellular receptor for factor VII, and the TF-FVII complex is a trigger of coagulation in vivo . 17,18Although TF has been described as an activator for thromboelastography, 11,13the effects of aprotinin have not previously been evaluated in this context. This study suggests that TF is a reproducible activator for thromboelastography, and the trace produced is not affected by aprotinin. This might favor its use in situations in which aprotinin is administered, so that an underlying coagulopathy is not obscured by aprotinin.

Aprotinin is well-recognized to cause prolongation of the ACT and APTR. These assays are specifically designed to evaluate the intrinsic pathway through activation of factor XII, which activates prekallikrein to kallikrein, the latter operating a positive feedback loop in factor XII activation. Aprotinin, as an antikallikrein agent, slows generation of activated factor XII and results in prolonged clotting times in assays of the intrinsic pathway. The prolongation of the celite and kaolin (both factor XII activators) thromboelastography R times by aprotinin is consistent with our understanding, as is the correlation between the APTR and the celite-activated R times. Similarly, the failure of aprotinin to affect TF-activated traces is predictable because aprotinin has no effect on either TF or the subsequently activated extrinsic pathway. As expected, the TF-activated thromboelastography R times correlated with the INR.

Actions of celite and TF at different sites in the coagulation pathways produced differences in the thromboelastography traces when the same blood samples were tested. This could have important clinical applications. For example, TF-activated thromboelastography is more reliable in detecting a warfarin effect than is celite-activated thromboelastography. 19The relation between thromboelastography parameters and established laboratory tests of coagulation requires in-depth investigation and validation for thromboelastography to gain widespread clinical acceptance. 20,21 

In conclusion, this study showed that aprotinin resulted in several changes in the thromboelastography trace when celite and kaolin were used as activators but not when TF was used. This mirrors the effects of aprotinin on laboratory tests of coagulation and is consistent with the mode of action of aprotinin. The clinical implication is that prolongation of the R time, decrease in the α angle, and decrease in the MA of both celite- and kaolin-activated thromboelastography should be interpreted with caution in the presence of aprotinin.