Development of a risk-oriented algorithm for the combined use of hemostatics and anticoagulants to prevent thrombosis and bleeding cases after total arthroplasty of knee or hip joints

Introduction: The goal of our study was to develop a risk-oriented algorithm for the combined use of hemostatics and anticoagulants in patients after total arthroplasty of the knee or hip joints to reduce the risk of thrombohemorrhagic complications. Materials and methods: We performed a retrospective study (n=253). In group (Gr.) 1, the time interval (TI) between the administration of hemostatic and anticoagulant prophylaxis was ≤17 hours (n=145; 57.31%), and in Gr. 2 – 18-24 hours (n=108; 42.68%). We analyzed the influence of different factors on the development of thrombosis and bleeding cases after the operation. Results and discussion: Thrombohemorrhagic complications were observed in 27 (10.67%) patients. Thrombosis in Gr. 1 was associated with the use of tranexamic acid, and were recorded 2.2 times more often than in Gr. 2 (p<0.05). The development of thrombosis in Gr. 1 was influenced by: class II obesity, type 2 diabetes mellitus, myocardial infarction, venous pathology, age of patients >75 years, for women – an initially low level of international normalized ratio, and activated partial thromboplastin time (APTT) (p<0.05). The development of bleeding in Gr. 1 was influenced by: age >75 years, among men and women – an increased preoperative level of APTT, for women – a decreased level of fibrinogen and platelets (p<0.05). Conclusion: To prevent thrombosis and bleeding after arthroplasty of large joints, the TI between the use of hemostatics and anticoagulants should be at least 18 hours, especially in patients with the above risk factors, in particular, when using tranexamic acid and low molecular weight heparins.


Introduction
Osteoarthritis (OA) is a chronic degenerative-inflammatory disease in which the articular cartilage, subchondral bone, synovial membrane, ligaments, capsule and periarticular muscles are affected (Kraus et al. 2015;Grässel and Muschter 2020 ). According to various sources, OA can occur in 6 to 30% of people (Mustafin and Khusnutdinova 2015;Helo et al. 2018;Ariani et al. 2019) and is more often observed in people over 40-50 years old (Palazzo et al. 2016). Total arthroplasty is the most effective treatment for terminal OA (Helo et al. 2018). All over the world, there is a tendency for an increase in the number of surgeries for endoprosthetics of large joints of the lower extremities (Oremus 2015;Leitner et al. 2018). However, with these operations, the development of various complications is possible. Thrombosis after surgery can develop in 4.3-60% of cases; to prevent this, anticoagulants are used (Bozhkova et al. 2018). The prescription of anticoagulants for the prevention of venous thromboembolic complications after surgery is regulated in the guidelines of different countries (Johanson et al. 2009;Hill et al. 2010;Falck-Ytter et al. 2012;Mironov et al. 2012;Wickham et al. 2012;Bokeriya et al. 2015). Besides, arthroplasty may be followed by bleeding of varying severity, up to 20-40% of the circulating blood volume (Nakopia et al. 2017). Hemostatic agents are traditionally used to prevent cases of bleeding after surgery (Nakopia et al. 2017).
It is important to note that the combined use of anticoagulants and hemostatics is a complex and debatable question in pharmacology due to the bidirectional effects of these drugs. It is necessary to take into account the pharmacokinetics of the drugs, their half-life, and the duration of action in order to determine the regimen of co-administration of the drugs. Currently, there is no consensus among scientists about the combined use of anticoagulants and hemostatics. Fraval et al. (2019) reports of efficacy of combined hemostatic and anticoagulant prophylaxis in patients after hip arthroplasty. In the group that had received tranexamic acid and enoxaparin sodium, the volume of intraoperative and postoperative blood loss was less (0.51 L and 1.13 L, respectively) than in the group that had received saline and enoxaparin sodium (0.698 L and 1.48 L, respectively) (p<0.001). However, that article does not contain information about postoperative complications, that is, it does not mention the safety of the drugs used (Fraval et al. 2019). Hourlier and Fennema (2018) studied the complications of combined use of hemostatic and anticoagulant drugs among patients with endoprosthetics. The number of thrombosis and bleeding cases was higher in patients treated with tranexamic acid compared with patients not treated with this hemostatic (p>0.05). However, some limitations in their work were recorded (variable prescription regimens of tranexamic acid, criteria for prescribing tranexamic acid to patients, insufficient size of the comparison group) and recommend that the results of this study should be interpreted with caution (Hourlier and Fennema 2018). Clavé et al. (2019) found that tranexamic acid together with rivaroxaban was more effective in reducing surgical blood loss after hip arthroplasty without an increase of thrombosis compared with patients who received only rivaroxaban (p<0.05) (Clavé et al. 2019). In the Russian clinical guidelines Prevention of Venous Thromboembolic Complications in Traumatology and Orthopedics (Mironov et al. 2012) and Russian Clinical Guidelines for the Diagnosis, Treatment and Prevention of Venous Thromboembolic Complications (VTEC) (Bockeria et al. 2015), there is no description of how to jointly prescribe hemostatics and anticoagulants to prevent thrombosis and bleeding after this type of surgery.
To predict the possibility of postoperative complications, it is important to take into account the individual characteristics of patients: preoperative coagulogram parameters and the presence of comorbid pathology. The presence of arterial hypertension (AH) increases the risk of cardiovascular complications after surgery by 35%; another risk factor here is a decreased renal function (Kristensen et al. 2014). An increase in body mass index (BMI) for every 5 kg/m 2 increases the risk of postoperative complications by 1.5 times (Yousef and Akhtyamov 2017).
Thus, the goal of our study is to develop a risk-oriented algorithm for the combined use of hemostatics and anticoagulants in patients after total arthroplasty of the knee or hip joints to reduce the risk of thrombohemorrhagic complications.

General characteristics of the study
We conducted a retrospective study on the basis of the traumatology and orthopedic department of the clinic of Rostov State Medical University in 2017-2019. The study was approved by the local independent Committee for Ethics of Rostov State Medical University (29 Nakhichevan Lane, Rostov-on-Don, Russia), Protocol № 18/17 of 23 October 2017. The patient inclusion criteria for the study: joint prophylaxis with hemostatics and anticoagulants, and knee or hip arthroplasty. The exclusion criteria: failure to conduct planned surgery due to the patient's condition, lack of joint anticoagulant and hemostatic prophylaxis, and no knee or hip arthroplasty. We divided our research into several stages. The first stage involved the selection of medical records of the patients based on the inclusion criteria. The second stage was a retrospective analysis of the selected cases (n=253): an analysis of the anticoagulants and hemostatics used. Stage 3 involved dividing patients into two groups according to the time interval (TI) between the last prescription of a hemostatic agent and the first prescription of an anticoagulant. Stage 4 -assessment of the developed thrombohemorrhagic complications in the early postoperative period. Stage 5 -an analysis of risk factors of the patients and their in-fluence on the development of thrombohemorrhagic complications and assessment of the influence of the initial parameters of the hemostasiogram on the development of thrombosis and bleeding after surgery.
Hemostatics and anticoagulants used in patients after surgery to prevent bleeding and thrombosis We analyzed the used hemostatics and anticoagulants. Dosages and frequency of administration of the used anticoagulant drugs: heparin -5000 IU/ml, (5 ml), 4 times a day; enoxaparin sodium -4000 anti-Xa ME/0.4 ml (40 mg), once a day, subcutaneously; nadroparin calcium -3800 IU, 0.4 ml, once a day, subcutaneously; dalteparin sodium -5000 ME (anti-Xa)/0.2 ml, once a day, subcutaneously; parnaparin sodium -4250 anti-Xa IU/0.4 ml, once a day, subcutaneously; rivaroxaban -10 mg/15 mg, once a day, orally; dabigatran -110 mg, 2 tabs once a day or 150 mg 1 capsule 1-2 times a day. The frequency of use of various anticoagulants in patients is presented in Table 1.
Consequently, a large number of patients began anticoagulant prophylaxis with low molecular weight heparins (LMWHs) or unfractionated heparins (UFHs). Five days later, some patients were transferred to new oral anticoagulants (NOACs).
About 95% of the patients used only one hemostatic, and about 60% used tranexamic acid. In case of using the two drugs, the first was administered during the operation, and the second -at the end of the first day after the operation. Considering the fact that most patients used tranexamic acid, it should be noted that the antifibrinolytic activity of tranexamic acid can continue in various tissues of the body for up to 17 hours (according to the Register of Medicines of Russia. Encyclopedia of Drugs).

Dividing patients into groups
Based on this, we divided our patients into two groups, depending on the TI between the last administration of hemostatics and the first administration of anticoagulants. The first group (Gr.1) consisted of 145 patients, of whom women -112 (77.24%), and men -33 (22.76%), with TI≤17 hours. The second group (Gr. 2) included 108 patients, of whom 78 (72.22%) were women and 30 (27.78%) were men, with TI being 18-24 hours. Duration of hospitalization in Gr. 1 for men was 11.87±4.14 days, for women -11.37±3.87 days. Duration of hospitalization in Gr. 2 for men was 11.63±2.71 days, and for women -11.55±3.06 days. In Gr. 1, the average age of men was 63.35±9.21 (with 63.35 being the mean and 9.21 being the standard deviation) years, and of women -64.32±10.22 years. In Gr. 2, the average age of men was 62±13.34 years, and of women -66.36±10.43 years. We divided the patients in Groups 1 and 2 into age categories (Table 3) in accordance with the classification of the World Health Organization: young age -18-43 years, middle age -44-59 years, older age -60-74 years, senile age -75-90 years, over 90 -long-livers (Ganin 2019).
Thus, in our study, most of the patients were of older age.

Statistical processing of the obtained data
The statistical data were processed on a personal computer using the MS Office software package (Excel 2010) and Sta- Note: data are represented in the form of n (%), where n -number of cases, % -percentage of cases. Note: data are represented in the form of n (%), where n -number of cases, % -percentage of cases. Note: data are represented in the form of n (%), where n -number of cases, % -percentage of cases. In Group 1 -time interval between the last administration of hemostatics and the first administration of anticoagulants≤17 hours. In Group 2 -the time interval 18-24 hours. tistica 10.0 (StatSoft, USA). All the indicators were checked for normal distribution by the Kolmogorov-Smirnov test. The Mann-Whitney U-test was used to compare the quantitative indicators between the groups and assess their impact on the development of thrombosis and bleeding cases. When comparing qualitative indicators, Pearson's χ² test was used, with Yates' correction. For qualitative indicators that demonstrated their influence on the development of thrombosis and bleeding cases, the relative risk (RR) and the limits of the 95% confidence interval (CI) were calculated. The influence of risk factors were analyzed separately for thrombosis and bleeding cases in Groups 1 and 2. Separately for men and women in Groups 1 and 2, the RR and CI of the preoperative hemostasiogram parameters were calculated. Differences were considered statistically significant at p<0.05.

Results and discussion
When analyzing the schemes for the combined prescription of hemostatics and anticoagulants, 29 combinations were identified (Table 4).
In total, among all the patients, we registered 27 (10.67%) cases of thrombosis and bleeding after surgery. In Gr. 1, we identified 22 (14.48%) cases of complications: 6 (27.27%) in men and 16 (72.73%) in women. Among the cases of complications in patients of Gr. 1, we found 11 (50%) cases of thrombosis: 2 (18.18%)in men and 9 (81.82%) -in women. We also registered cases of bleeding in 11 (50%) patients: 4 (36.36%) -in men and 7 (63.64%) -in women. In Gr.2 after surgery, we recorded 5 (4.63%) cases of complications, which is 4.5 times fewer (p=0.0098) than in Gr. 1. We found that in Gr. 2, there were only cases of thrombosis recorded (n=5; 100%), with no cases of bleeding. In Gr. 2, the complications were recorded in 3 (60%) women and 2 (40%) men. The incidence of thrombosis in Gr. 1 was 2.2 times higher than in Gr. 2 (p=0.023). The data indicate the importance of maintaining the TI between the use of hemostatic and anticoagulant prophylaxis. The distribution of cases of thrombosis and bleeding after surgery, depending on the tactics of pharmacoprophylaxis, is shown in Table 5.
Thus, we found that when using the combination "aminomethylbenzoic acid + parnaparin sodium + heparin", Table 4. Combinations of hemostatics and anticoagulants in groups 1 and 2 and the frequency of their use. Note: data are represented in the form of n (%), where n -number of cases, % -percentage of cases. In Group 1 -the time interval between the last administration of hemostatics and the first administration of anticoagulants ≤ 17 hours. In Group 2 -the time interval 18-24 hours. bleeding was observed in 100% of the patients, and when prescribing the combination "tranexamic acid + aprotinin + enoxaparin sodium + rivaroxaban", thrombosis in the early postoperative period was registered in 66.67% of the patients. Of all the complications (n=27), 18 (66.67%) developed with the use of tranexamic acid. Moreover, 16 (88.89%) of these 18 cases were recorded in Gr. 1, with the TI of ≤ 17 hours. All the thrombosis cases in Gr. 1 (n=11) were associated with tranexamic acid use (p=0.038). In addition, all the complications were linked to LMWHs as the onset of anticoagulant prophylaxis.

Structure of the comorbid status of patients and its influence on complication development
We assessed the comorbid status of patients in Groups 1 and 2 (Table 6). To assess renal function in the patients before surgery, we calculated the glomerular filtration rate (GFR), using the CKD-EPI formula (ml/min/1.73 m 2 ). The obtained data on the value of GFR were analyzed by the standard interpretation (Shutov 2014 Thus, the groups of patient were comparable in terms of gender, age, term of hospitalization, and baseline health (p>0.05).
We analyzed the possibility of the effect of the comorbid pathology shown in Table 6 on the development of thrombosis and bleeding cases in the early postoperative period. Among the patients in Gr. 1 with developed thrombosis, we revealed statistically significant differences in terms of: BMI -between the patients with class II obesity and overweight (p=0.037); age -between patients of young and middle age (p=0.011), young and old age (p=0.015), and old and senile age (p=0.029); a history of myocardial infarction (MI) (p=0.00002), type 2 diabetes mellitus Note: data are represented in the form of n (%), where n -number of cases, % -percentage of cases. In Group 1 -the time interval between the last administration of hemostatics and the first administration of anticoagulants ≤17 hours. In Group 2the time interval 18-24 hours. GFR -glomerular filtration rate.
(DM) (p=0.00001), venous pathology (p=0.045), and impaired renal function -between patients with a mild loss and mild to moderate loss of kidney function (p=0.046).
In the patients of Gr. 2 who developed thrombosis after surgery, we found significant differences between the patients with class II obesity and overweight (p=0.044), with class II and class I obesity (p=0.039) and between old and senile patients (p=0.037). We also found that in the patients of Gr. 1 and Gr. 2 with thrombosis after surgery, a statistically significant difference in TI was revealed (p=0.023). Among the patients of Gr. 1 with bleeding after endoprosthetics, statistically significant differences were between old and senile patients (p=0.015). In Gr. 2, no analysis of risk factors for cases of bleeding was carried out, as no cases were registered. When analyzing other risk factors among the patients of Groups 1 and 2 with thrombosis and bleeding developing after surgery, no statistically significant differences (p>0.05) were found. Next, we calculated the analysis of the risk ratio and difference for factors that statistically significantly influenced the development of thrombosis or bleeding cases after surgery. We found out that it was statistically significant in the patients of Gr. , and mild to moderate loss of kidney function -6.231 times (RR=6.231, CI=1.157-33.554). In the patients of Gr. 2 with cases of thrombosis after surgery, there was no variation in significant factors, so we did not assessment of the ratio and difference of risks. In the patients in Gr. 1, the age over 75 years was associated with a 12-time higher risk of bleeding (RR=12, CI=1.445-99.678).

Analysis of the baseline coagulogram parameters and their influence on the development of thrombosis and bleeding cases
Along with the assessment of a comorbid status, the baseline hemostasiogram parameters and their effect on the development of thrombosis and bleeding cases after surgery were analyzed. The study of the coagulogram involved the analysis of the following indicators: 1) activated partial thromboplastin time (APTT); 2) prothrombin time (PT); 3) concentration of fibrinogen (Fg); 4) international normalized ratio (INR); 5) number of platelets (PLT).
After that, we assessed of the ratios and difference in the risk of thrombosis and bleeding cases, depending on the initial (preoperative) level of the coagulogram, gender and the Group of the patients. In the men of Group 1 with thrombosis after surgery, no statistically significant   differences were found between the initial hemostasiological parameters and thrombotic complications. In the women of Gr. 1 with thrombosis after surgery, statistically significant differences were found in the following initial coagulogram indicators: 1) INR: between the patients with values below normal and normal (p=0.00032) and between the patients with values below and above normal (p=0.00001); 2) APTT: between the patients with values below normal and normal (p=0.0037); 3) Fg: between the patients with values below normal and normal (p=0.0062). In Group 2 patients, both men and women, with thrombosis in the early postoperative period, no statistically significant difference were found between the preoperative parameters of the coagulogram and the cases of thrombosis. In Group 1. in men with bleeding after surgery, there were differences in the baseline values of the following indicators: 1) APTT: below normal and normal (p=0.039), above normal and normal (p=0.012); 2) PT: above normal and normal (p=0.042). Among the women of Group 1 with hemorrhagic complications, there were statistically significant differences in the following baseline indicators: 1) APTT: above normal and normal (p=0.00022); 2) Fg: below normal and normal (p=0.00065) and below and above normal (p=0.00001); 3) PLT: below the reference values and normal (p=0.038). When evaluating other baseline hemostasiogram parameters, no statistically significant differences were revealed in the patients with cases of thrombosis and bleeding in Groups 1 and 2, both males and females (p>0.05). For the factors that had a statistically significant effect on the development of thrombosis or bleeding cases in the early postoperative period, with variations, we analyzed the ratio and difference of risks. When analyzing the level of influence of hemostasiogram indicators before surgery on the development of thrombosis in the women of Group 1,    we found that INR below normal was associated with a 13.33-time increased risk of thrombosis (RR=13.333, CI=4.49-39.591), and APTT below normal before surgery was associated with a 5.8-time increased risk of thrombosis (RR=5.8, CI=1.357-24.796). In the women of Group 1, a preoperative APTT level above normal was associated with a 28-time increased risk of bleeding (RR=28, CI=3.426-228.831), and low levels of Fg and PLT in the preoperative period were associated with an increased risk of hemorrhages by 23.25 times (RR=23.25, CI=3.117-173.423) and 10.2 times (RR=10.2, CI=1.805-57.619), respectively. We found that in the men of Group 1, a baseline APTT level above normal was associated with an 18-time increased risk of bleeding cases (RR=18, CI=2.679-120.922). Thus, the overall incidence of thrombosis and bleeding after surgery in the patients of our study was 6.32% and 4.35%, respectively. In other studies related to the analysis of the incidence of thrombosis and bleeding after knee or hip arthroplasty, the similar data are obtained (Ershov et al 2015;Bozhkova et al. 2018). A higher incidence of thrombosis can, firstly, be influenced by the presence of ischemic heart disease and hypertension, because these pathologies can "shift" the hemostatic system towards hypercoagulation (Krasnova et al. 2019). Secondly, in patients with OA, endothelial dysfunction and baseline hypercoagulability are observed, which also increases the risk of thrombosis (Suzuki et al. 2014). In our study, we found that the use of tranexamic acid with TI between hemostatics and anticoagulants of less than 18 hours was statistically significantly associated with a higher risk of thrombosis cases after arthroplasty. Kim et al. (2018) provided the similar data, but in that study the difference was not statistically significant (Kim et al. 2018), in contrast to our work. We demonstrated the importance of adherence to TI between hemostatics and anticoagulants of at least 18 hours. The frequency of all complications, thrombosis in particular, was higher in Group 1, with a TI under 18 hours. Different studies consider diseases most often associated with OA (according to the Centers for Disease Control and Prevention 2015; Alabut and Sikilinda 2016), and their impact on the risk of thrombosis and bleeding cases is assessed (Boutsiadis et al. 2017;Gronbeck et al. 2019;Sloan et al. 2019). It is important to pay special attention to those risk factors that can provoke the onset of complications in the postoperative period after total knee or hip arthroplasty.
However, it should be noted that our study has some limitations, as it is retrospective and its results should be interpreted with caution.

Conclusion
Based on the presented results of our retrospective study, we have developed a risk-oriented algorithm for the prevention of thrombosis and bleeding after arthroplasty of the joints of the lower extremities in patients with different risk factors.
In the presence of such risk factors as: for men and women -type 2 diabetes mellitus, a history of myocardial infarction, class II obesity, senile age, venous pathology, mild to moderate loss of kidney function before surgery; for men -an initially high level of activated partial thromboplastin time; for women -initially high/low level of activated partial thromboplastin time, reduced level of international normalized ratio/fibrinogen/platelets, it is recommended that anticoagulant prophylaxis should not be started with low molecular weight heparins, especially in combination with tranexamic acid. However, if there is no chance of choosing unfractionated heparins or new oral anticoagulants, then the time interval between low molecular weight heparins and tranexamic acid should be at least 18 hours.
In the absence of risk factors: there are no restrictions on the use of drugs (except for individual contraindications), but with joint prophylaxis with hemostatics and anticoagulants, it is recommended that the time interval between the last administration of hemostatics and the first administration of anticoagulants should be at least 18 hours.