Comparison of three point-of-care testing devices to detect hemostatic changes in adult elective cardiac surgery: a prospective observational study

Excessive bleeding in cardiac surgery occurs in as much as 20% of patients. It is associated with increased morbidity and mortality, mainly caused by need for reexploration and transfusion of blood products [1‐6]. In 15-50% of the patients undergoing reexploration, a surgical cause of bleeding cannot be found [7, 8] and it is assumed to be of microvascular origin. The pathogenesis of microvascular bleeding in cardiac surgery is multifactorial and may be related to changes in the hemostatic system associated with increased age, preoperative medication with platelet and/or coagulation inhibitors, and transient platelet dysfunction associated with cardiopulmonary bypass (CPB) [6‐12]. The use of CPB induces a systemic inflammatory response with activation of both the coagulation and fibrinolytic systems, followed by coagulopathy caused by factor consumption and transiently reduced platelet count and function [13, 14].

To date, there is no “gold standard” test for the assessment of microvascular bleeding in cardiac surgery. Perioperative monitoring of the hemostatic process by a combination of conventional coagulation tests may be helpful in diagnosing possible causes of bleeding [15, 16]. However, the value of these tests has been questioned, mainly because they are time-consuming, performed in plasma and do not add information about platelet function or the contribution of the cellular components [17, 18].

In the last decade, viscoelastic point-of-care (POC) coagulation instruments such as the thromboelastograph (TEG), the thromboelastometer (RoTEM) and the Sonoclot have been increasingly used in clinical practice for perioperative monitoring of excessive bleeding in cardiac surgery [18‐20]. These POC instruments provide information about the quality and dynamics of the clot, are relatively simple to operate and can be used close to the patient. The results are usually ready for interpretation within 15–30 minutes. Guidelines for interpretation and use of POC instruments in the management of perioperative coagulopathy and transfusion algorithms have been developed [21]. These instruments may be helpful as a guide for transfusion therapy and to decide whether to reexplore a patient or not [17‐19, 22].

Although with some differences, these instruments are based on similar principles and provide output in the form of a similar graphical tracing. However, it is still not clear whether TEG, RoTEM and Sonoclot provide identical information. The performance of different POC instruments in cardiac surgery has been evaluated in previous studies [22‐24] but, to our knowledge, TEG, RoTEM and Sonoclot instruments have not previously been directly compared in the same group of cardiac surgery patients.

The primary objective of this study was to evaluate the comparability of the results obtained from the TEG, RoTEM and Sonoclot instruments by assessing their ability to detect changes in hemostasis caused by CPB. In addition, we also investigated whether variables from the TEG, RoTEM and Sonoclot analyses were correlated to results from a selection of conventional coagulation tests.

Patient demographics, medical data and data on surgical procedures are presented in Table 2.

All the analyzed conventional coagulation tests differed significantly with time (Table 3). There were significant changes with time for the TEG variables R (reaction time), K (clot kinetics), alpha (clot strengthening), MA (clot strength) and G (clot elasticity), mainly between the preoperative results and the results recorded 1 hour after surgery (Table 4). MA and G were also significantly different from baseline 24 hours postoperatively. The RoTEM variables INTEM CT (clotting time), INTEM CFT (clot kinetics) and FIBTEM MCF (clot strength) changed significantly with time. For the Sonoclot, only Son ACT values (time until fibrin formation) differed significantly with time.

Correlations between the conventional coagulation tests and POC variables are given in Table 5. TEG MA (clot strength) and RoTEM MCF (clot strength) correlated with fibrinogen at baseline. Even though both TEG alpha (clot strengthening) and TEG G (clot elasticity) were correlated to fibrinogen 1 and 24 hrs postoperatively and TEG K (clot kinetics) was correlated to fibrinogen 24 hrs postoperatively, TEG MA (clot strength) was the only TEG variable that correlated with fibrinogen at all time-points. The RoTEM MCF (clot strength) variables and the Sonoclot Clot Rate (rate of fibrin formation) also correlated to fibrinogen at all time-points.

Platelet counts correlated with TEG MA (clot strength) 24 hrs postoperatively, and two different RoTEM MCF variables (clot strength) correlated with platelet counts either 1 hr or 24 hrs postoperatively. The only correlations with aPTT were with the Sonoclot variables Son ACT (time until fibrin formation) and Clot Rate (rate of fibrin formation) at 1 and 24 hrs postoperatively. None of the POC variables correlated with INR.

Four patients needed a surgical reexploration, due to persistent postoperative bleeding (total bleeding 800–2690 mL). These patients were transfused with red cell concentrates (6–10 units), platelet concentrates (0–3 units) and FFP (3–6 units). Two of these patients received an additional dose of tranexamic acid postoperatively. The decision of giving tranexamic acid postoperatively was empirical, as the results of the POC instruments were not available for the attending physician. None of the bleeding patients had abnormal baseline values for the TEG, RoTEM or for the coagulation tests. Preoperative Sonoclot Son Act (time until fibrin formation) results were abnormal in three of these patients. However, pathological Son Act results were found in the majority of the patients preoperatively. Only one patient showed pronouncedly reduced RoTEM FIBTEM MCF (maximal clot firmness), plasma fibrinogen and TEG alpha (clot strengthening) values in the early postoperative period.

As expected, all the conventional coagulation tests showed significant changes following cardiac surgery. This was mirrored in changes in all the TEG variables, some of the RoTEM variables, and only one of the Sonoclot variables. The changes pointed towards a more hypocoagulable state, consistent with dilutional coagulopathy and consumption of platelets and coagulation factors occurring during CPB [7, 8]. In most patients, the changes reversed to baseline values within 24 hours after surgery. In contrast to other studies [9, 12], we did not observe increased fibrinolysis, which may be explained by the routine use of a prophylactic antifibrinolytic agent.

Several previous studies [22, 25‐28] do not support the routine use of POC instruments prior to surgery to predict the need for transfusion, due to the high number of false positive and false negative results obtained. However, rapid assessment of the cause of bleeding if it ensues in the postoperative period is essential to provide adequate treatment to the patient. Alterations in coagulation usually occur during CPB, mainly due to hemodilution and platelet activation caused by the contact with plastic surfaces. For a POC instrument to be useful, it should be sensitive enough to indicate such alterations in hemostasis in a timely fashion, and the data from 1 hr postoperatively may be the most relevant when comparing the instruments. In that regard, a variant of the TEG analysis, Rapid TEG, where tissue factor is added to kaolin for activation of the extrinsic pathway has shown promising results. Several studies have demonstrated that Rapid TEG is useful as the results are available in about 20 minutes, correlate well with conventional coagulation tests, and may be used to predict transfusion of red cells, plasma and platelets [29, 30].

From our study, TEG seemed to be the more sensitive of the POC instruments and the Sonoclot was the least sensitive. None of our patients had markedly abnormal hemostasis and we cannot exclude that the instruments would have given more comparable results in patients with severe hemostatic disturbances.

Previous comparative studies to assess the interchangeability between TEG and RoTEM have been reported [24, 31]. Even if TEG and RoTEM are very similar instruments, there are some differences between them. The RoTEM instrument analyses each parameter in a separate channel while in the TEG, all the parameters are analyzed simultaneously. In the TEG, the cup oscillates around the metal pin, whereas the pin moves around the cup in the RoTEM, conferring more stability against vibrations. The TEG uses kaolin and the RoTEM uses tissue factor (EXTEM) or the contact activator ellagic acid (INTEM) as activator of the coagulation process. The different plastic composition of the reaction cups has been pointed out as a possible additional factor for the lack of interchangeability between the two instruments [17]. All these differences may explain the divergent results obtained from the two instruments in our study and reported by other authors.

From our study, it seems that all three POC instruments may be equally useful to quantify changes in fibrinogen. The correlations found between various POC variables and fibrinogen concentrations are consistent with previous studies [20, 32‐34]. The importance of fibrinogen in promoting sufficient coagulation and hemostasis and as a contributor to the clot strength has lately been underscored [17]. Reduced fibrinogen levels during CPB identify patients at risk for post-operative bleeding and high transfusion requirements [13].

No correlations were found between INR and the TEG variable R (reaction time), the RoTEM CT (clotting time) or the Sonoclot Son ACT (time until fibrin formation), even if these three parameters all measure the time to the first fibrin formation. The reason for the discrepancy for TEG R might be the different activators used; kaolin for TEG and tissue factor for INR. For the other fibrin time measures, the reasons for the discrepancies are unknown. We found that the Sonoclot variables Son ACT and Clot rate (rate of fibrin formation) correlated significantly with aPTT. This is as expected, as aPTT is usually prolonged in hypocoagulable states, where a low Clot Rate can be found.

Preoperative platelet counts were within normal ranges in all the patients. None of the POC tests showed correlations with platelet counts at both time points in the postoperative period. The explanation may be that the POC instruments rather assess platelet function than platelet counts even if they may also indicate when platelet counts are too low to provide adequate hemostasis. From this point of view, the POC instruments may be more clinically useful than standard platelet counts.

Even if the TEG seemed to be more sensitive to detect hemostatic changes than the two other POC instruments, its use still carries potential problems. The references for the manufacturer’s proposed normal values for the TEG values are very limited and not well validated [34, 35]. The manufacturer’s technical manual does not contain information about factors that may influence the results, such as gender, age or underlying medical conditions. In one study in 118 healthy blood donors [35] the authors found that 18.6% of the donors had at least one abnormal TEG parameter according to the manufacturer’s reference values, with the greatest differences involving the R (reaction time), alpha (clot strengthening) and MA (clot strength) parameters, even if none of the volunteers had abnormal routine coagulation assays. The authors recommended that each institution should determine its own reference values before adopting TEG. Concerns regarding blood sample stability and storage conditions have also been raised [17].

Laboratories should establish standardized procedures to minimize variability. To solve the mentioned problems the International TEG-RoTEM Working Group has been created to try to standardize TEG and RoTEM and to establish protocols to ensure accurate results [36]. Until better standardization is possible and the knowledge about factors influencing the results is increased, the routine coagulation tests will have to remain the gold standard for assessment of many aspects of hemostasis.

The main limitation of this study is the small number of patients. In addition, there was some missing data for the RoTEM analysis, caused by short-time supplier problems (n = 1-4). However, our results are consistent with previous findings from studies with separate evaluation of the TEG, RoTEM and Sonoclot instruments in cardiac surgery. Our study was not designed to study clinical outcomes in relation to the POC tests, which would require a much larger patient population.

In conclusion, TEG and RoTEM can be used to detect postoperative hemostatic changes following cardiac surgery, whereas the Sonoclot seems to be less suitable, at least in patients without grave hemostatic changes. Variables from TEG, RoTEM and Sonoclot may be useful to monitor fibrinogen levels. Further studies are necessary to establish adequate reference values for different patient groups and to standardize these assays to achieve reliable results.