Changes in thromboelastography to predict ecchymosis after knee arthroplasty: a promising guide for the use of anticoagulants: a case control study


 BackgroundEcchymosis is one of the worrisome complications after total knee arthroplasty (TKA) and interferes with functional rehabilitation. Current clinical guidelines do not provide individualized approaches for patients with ecchymoses. In this study, we used thromboelastography (TEG) to determine the coagulation state after TKA and to then explore markers that predict the occurrence of ecchymosis events after TKA.MethodsIn our cohort, patients were divided into ecchymosis (n=55) and nonecchymosis (n=137) groups according to whether ecchymosis events occurred after TKA. Rivaroxaban 10 mg/d was taken orally for thromboprophylaxis after surgery. All patients completed TEG testing. Correlation analysis was used to determine the risk factors for ecchymosis after TKA, and receiver operating characteristic (ROC) curves for variables with significant correlation were plotted.ResultsIn all, 55 of the 192 patients (28.65%) developed ecchymosis surrounding the surgical site. Multivariate analysis showed that hidden blood loss (OR = 1.003 and p =.005) and changes in the coagulation index (ΔCI) values (OR = 0.351 and p =.001) were risk factors for ecchymosis after TKA. Using the Youden index, 0.1805 was determined as the optimal threshold value of ΔCI for predicting the occurrence of ecchymosis, with a sensitivity of 74.55% and specificity of 72.99%.ConclusionsΔCI is a promising marker as an alarm for the occurrence of ecchymosis after TKA.Trial registrationChinese Clinical TrialRegistry, ChiCTR, NO. ChiCTR1800017245. Registered 19 July 2018 - Prospective registered, http://www.chictr.org.cn


Background
Total knee arthroplasty (TKA) is considered the most effective treatment for end-stage knee osteoarthritis (OA) [1].
Due to the high prevalence of OA, TKA is a fairly common surgery. In the United States, the number of TKAs is projected to increase by 673% by 2030 [2]. Venous thromboembolism (VTE) is a worrisome complication after TKA [3]. Perioperative anticoagulant prophylaxis has been shown to reduce the incidence of postoperative VTErelated mortality and complications [4]. Evidence-based guidelines recommend that patients undergoing TKA receive oral rivaroxaban for anticoagulant prophylaxis for 14 days[5, 6]. However, postoperative bleeding complications associated with anticoagulation are not uncommon, especially with the widespread use of factor Xa inhibitors, in which the incidence of ecchymosis around the wound is as high as 13% [7].
The formation of postoperative ecchymosis around the wound is related to the use of anticoagulants [3].
Ecchymosis around the surgical site can prolong the recovery time after TKA and may even lead to reoperation due to periprosthetic infection [8,9]. At present, there are still no clear guidelines for the balance between postoperative anticoagulation and bleeding [10]. The use of anticoagulants should prevent VTE and avoid the occurrence of bleeding events. The monitoring of coagulation function has guiding value for the use of anticoagulants [11].
Routine coagulation tests provide limited information about the quality of coagulation status [12,13]. Therefore, we need an alarm to predict the occurrence of ecchymosis events around the wound after TKA. Thrombelastography (TEG) provides a comprehensive evaluation of blood viscoelastic properties and has potential value in predicting postoperative bleeding and thrombotic events [12,14]. Moreover, previous studies have shown that the change of coagulation index (ΔCI) value was a risk factor for patients with ecchymosis after TKA and was expected to guide personalized anticoagulant therapy [6]. Thus, in this study, we sought to (1) explore whether the change in ΔCI can predict ecchymosis after TKA and (2) to calculate the threshold for predicting patients with ecchymosis based on ΔCI.

Methods
From October 2018 to October 2020, we prospectively enrolled patients who were scheduled to undergo primary unilateral total knee arthroplasty (TKA) for knee OA. We excluded patients who (1) underwent bilateral TKA; (2) did not undergo TEG testing; (3) had a history of cardiovascular surgery, VTE or prior anticoagulant therapy; (4) were concomitant with coagulation disorders; (5) were treated with anticoagulation agents other than rivaroxaban; or (6) had incomplete medical records. According to the occurrence of ecchymosis after TKA, the patients were divided into ecchymosis and nonecchymosis groups[6]. The study was registered in the Chinese Clinical Trial Registry (ChiCTR1800017245).
A tourniquet was used intraoperatively, and the tourniquet was loosened before the incision was closed. The anesthesiologist recorded the blood loss during the operation, mainly involving attracting blood from bottles and gauze. No drainage was used. All patients received standard physical therapy and rivaroxaban anticoagulant therapy for 14 days. Rivaroxaban (10 mg) was administered once daily starting 12 hours after surgery, monitoring the occurrence of ecchymosis closely and discontinuing rivaroxaban once ecchymosis was observed. Venous blood was collected 1 day before surgery to obtain baseline hematocrit (HCT) and TEG values. The HCT and TEG values were monitored daily postoperatively until the patient was discharged. For discharged patients, investigators followed them up daily to see if there were any ecchymosis events. Once there were ecchymosis events, HCT and TEG tests were completed within 24 hours, and anticoagulation therapy was stopped. TEG tests were performed by a TEG® Hemostasis Analyzer (Hemonetics Corporation, Braintree, MA, USA).
All indicators of TEG (R-time, α-angle, maximum amplitude and K-time) were recorded, and the CI was calculated using the formula CI = 0.1227(R) + 0.0092(K) + 0.1655(MA) − 0.0241(α) − 5.0220. We analyzed the ΔCI values for all patients. For patients without ecchymosis after TKA, we analyzed the maximum reduction in CI relative to preoperative values; for patients with ecchymosis, we analyzed the ΔCI values between the day of ecchymosis occurrence and preoperatively. The gross [15] equation was used to calculate the volume of human erythrocytes and total blood loss. Then, hidden blood loss was the residual value of the total blood loss during the removal of intraoperative blood loss.

Statistical Analysis
Statistical analysis was carried out with SPSS 24.0. Comparisons were made between the patients with and without ecchymosis. To clarify the values of factors related to ecchymosis in predicting the occurrence of ecchymosis events, the receiver operating characteristic (ROC) curve was established by MedCalc; the area under the ROC curve (AUC) was also calculated. The optimal cutoff values of each index for predicting the occurrence of ecchymosis events and the corresponding speci city and sensitivity were determined by Youden's J statistic. P < 0.05 was considered statistically signi cant.

Results
A total of 192 patients who received a unilateral primary TKA were eligible for the study. In all, 55 of 192 patients (28.65%) developed ecchymosis surrounding the surgical site. There were no statistically signi cant differences between the two groups in terms of age (p=.125), sex (p=.480), or BMI (P=.085) ( Table I). During the follow-up, only 3 patients developed ecchymosis around the wound, which improved after anticoagulant treatment was stopped, and blood samples of these patients were obtained on the day that ecchymosis was observed.   To measure the value of CI and hidden blood loss in predicting the occurrence of ecchymosis, we plotted the ROC curves of the two variables (Fig. 1). ∆CI discriminated between ecchymosis and nonecchymosis with an AUC of 0.794 (95% CI: 0.730, 0.849). However, hidden blood loss did not exhibit a superior AUC of 0.681 (95% CI: 0.610, 0.746) (Fig. 1).  ∆CI, change of coagulation index. CI, con dence interval; PPV, positive predictive value; NPV, negative predictive value. TKA, total knee arthroplasty.

Discussion
To our knowledge, this study was the rst attempt to use ΔCI to predict ecchymosis events after TKA and demonstrated its reliability. In our cohort, the ΔCI in the ecchymosis group was signi cantly higher than that in the nonecchymosis group (p = .001) and was an independent risk factor (OR = 0.351, p =.001) for ecchymosis after TKA. The optimal cutoff value of ΔCI (0.1805) re ected maximal sensitivity (74.55%) and speci city (72.99%) to predict ecchymosis events after TKA. This study also found a signi cant correlation (OR = 1.003, p =.005) between hidden blood loss and ecchymosis events after TKA. Unfortunately, hidden blood loss showed unsatisfactory sensitivity in predicting the occurrence of ecchymosis events. Hidden blood loss is generally de ned as blood deposited in the joint space and blood seeping into the tissue [16,17]. However, when the volume of hidden blood loss was not su cient, the blood in the tissue may have been absorbed by the body before it penetrated the mucosa of the skin, preventing the formation of ecchymosis.
Damage to the vascular wall at the surgical site provides a possibility for the formation of ecchymosis, but it may not be the only cause [18]. The effect of anticoagulants on the formation of ecchymosis cannot be ignored. In the absence of thromboprophylaxis treatment, the incidence of VTE after TKA can be as high as 40-84% [19]. However, anticoagulant chemoprophylaxis puts patients at risk for bleeding after TKA. Bleeding complications (including ecchymosis) following total joint arthroplasty are not acceptable, as they can lead to more important complications, such as infection, wound healing problems, dysfunction and loosening of the joints, with a high likelihood of affecting the surgical outcome[20]. Rivaroxaban, a direct oral factor Xa inhibitor, is commonly prescribed for the prevention of VTE after TKA due to its effectiveness, high safety and convenience of use [3,21].
However, evidence has shown that rivaroxaban use also increases the risk of postoperative bleeding[7]. This poses a challenge to the balance between anticoagulation and bleeding prevention. Therefore, accurate monitoring of coagulation status and early prediction of ecchymosis events are the key links of individual anticoagulation.
TEG is commonly used to evaluate the viscoelastic properties of a patient's whole blood during surgery [22]. As early as 2009, Jeffry et al. [14] showed the role of TEG in identifying hypercoagulability and predicting thromboembolic events in surgical patients. Previous studies have attempted to apply ∆R to adjust the use of anticoagulants, but the data showed that ∆R does not have this capability [11]. The possible explanation is that the clotting process involved in the formation of a thrombus or brinolysis is complex, and it is unreliable to represent the whole process by a single point of the clotting process. Therefore, in this study, we analyzed and compared the preoperative and postoperative changes in the comprehensive evaluation index CI. Through the analysis of two cohorts with or without ecchymosis after TKA, we found a high correlation between ∆CI and ecchymosis, which was consistent with previous research results[6]. In addition, we also con rmed that, when the CI was lower than 0.1805, it was a warning that the body was in a hypocoagulable state, and the probability of ecchymosis events was as high as 73.44%.
The patients were prospectively recruited, with each patient followed for at least 2 weeks. The 2-week follow-up covered the entire course of the patient's anticoagulant use, and it was assured that the patient's hemodynamics had stabilized by the end of the follow-up[8].
Some limitations need to be noted in this study. First, this study was performed in a single center, and more regional studies are needed to support our conclusions. We are, in a follow-up study, examining this issue. Second, the patients who underwent TKA were elderly and had different types of underlying diseases. Age and underlying diseases may also be risk factors for the occurrence of ecchymosis [23,24]. In this study, there was no subgroup analyses on the types of diseases and ages of patients, so the results of the study may be biased. Last, most of the patients were hospitalized for 5 days after surgery. Except for the patients with ecchymosis, blood samples were collected on the day that ecchymosis appeared, and the patients without ecchymosis only had TEG testing during hospitalization. Therefore, it was di cult to ensure that the maximum CI in the nonecchymosis group might not be the maximum.

Conclusion
This study, to our knowledge, is the rst to demonstrate that ∆CI in TEG parameters can be used as a predictor of ecchymosis events after TKA, and the optimal cutoff value was 0.1805. On the basis of our ndings, we believe that ∆CI is a promising indicator to guide the use of anticoagulants after TKA. When the ∆CI was lower than 0.1805, it was reasonable to consider stopping the use of rivaroxaban to avoid the occurrence of ecchymosis events. Of course, more research is needed to verify and con rm the reliability of this prediction.

Consent for publication
Patients signed informed consent regarding publishing their data.
Availability of data and materials The datasets supporting the results of this article are included within the article and its additional les.

Competing interests
All authors declare no con icts of interest.

Funding
The National Natural Science Foundation of China (82072443).

Authors' contributions
Ning Hu and Xuan Gong contributed to the experimental ideas and design of this study. Yuelong Chen drafted the manuscript. Leilei Qin collected and analyzed the data. Jiawei Wang performed the statistical analysis. Ning Hu, Xuan Gong examined and revised the contents of the manuscript. Figure 1