Inflammatory Biomarkers as Prognostic Factors of Acute Deep Vein Thrombosis Following the Total Knee Arthroplasty

Background and objectives: Deep vein thrombosis (DVT) is one of the most serious post-operative complications in the case of total knee arthroplasty (TKA). This study aims to verify the predictive role of inflammatory biomarkers [monocyte-to-lymphocyte ratio (MLR), neutrophil-to-lymphocyte ratio (NLR), platelets-to-lymphocyte ratio (PLR), systemic inflammatory index (SII), systemic inflammation response index (SIRI), and aggregate index of systemic inflammation (AISI)] in acute DVT following TKA. Materials and methods: The present study was designed as an observational, analytical, retrospective cohort study and included all patients over 18 years of age with surgical indications for TKA, admitted to the Department of Orthopedics, Regina Maria Health Network, Targu Mures, Romania, and the Department of Orthopedics, Humanitas MedLife Hospital, Cluj-Napoca, Romania between January 2017 and July 2022. The primary endpoint was the risk of acute DVT following the TKA, and the secondary endpoint was the length of hospital stay, and the outcomes were stratified for the baseline’s optimal MLR, NLR, PLR, SII, SIRI, and AISI cut-off value. Results: DVT patients were associated with higher age (p = 0.01), higher incidence of cardiac disease [arterial hypertension (p = 0.02), atrial fibrillation (p = 0.01)], malignancy (p = 0.005), as well as risk factors [smoking (p = 0.03) and obesity (p = 0.02)]. Multivariate analysis showed a high baseline value for all hematological ratios: MLR (OR: 11.06; p < 0.001), NLR (OR: 10.15; p < 0.001), PLR (OR: 12.31; p < 0.001), SII (OR: 18.87; p < 0.001), SIRI (OR: 10.86; p < 0.001), and AISI (OR: 14.05; p < 0.001) was an independent predictor of DVT after TKA for all recruited patients. Moreover, age above 70 (OR: 2.96; p = 0.007), AH (OR: 2.93; p = 0.02), AF (OR: 2.71; p = 0.01), malignancy (OR: 3.98; p = 0.002), obesity (OR: 2.34; p = 0.04), and tobacco (OR: 2.30; p = 0.04) were all independent predictors of DVT risk. Conclusions: Higher pre-operative hematological ratios MLR, NLR, PLR, SII, SIRI, and AISI values determined before operations strongly predict acute DVT following TKA. Moreover, age over 70, malignancy, cardiovascular disease, and risk factors such as obesity and tobacco were predictive risk factors for acute DVT.


Introduction
Total knee arthroplasty (TKA) is one of the most common orthopedic surgeries performed worldwide, addressing severe cases of knee osteoarthritis, and improving patients' quality of life and their capability to move [1][2][3]. Recently, advancement in operating techniques, as well as the vast variety of prostheses available, have resulted in a significant decrease in post-surgical complications, an increase in success rate, and therefore an improvement in patient quality of life [4][5][6].
Deep vein thrombosis (DVT) is one of the most serious post-operative complications in the case of TKA, leading to pulmonary embolism (PE) and mortality [7][8][9][10]. Furthermore, TKA is associated with a greater incidence of DVT occurrence compared to total hip arthroplasty (THA) [11].
Thrombosis occurs when the key components of Virchow's triad, including blood circulation stagnation, endothelial injury, and hypercoagulability, are present [12]. Numerous studies have demonstrated the association of inflammatory status with hypercoagulability status [13,14]. Moreover, various inflammatory biomarkers with predictive significance in the diagnosis of post-TKA DVT have been studied in the last ten years [15][16][17]. Interleukin-6 (IL-6) and C-reactive protein (CRP) are two of the most studied inflammatory biomarkers in the prediction of complications following TKA [18][19][20][21], but unfortunately, the results are inconsistent and unsatisfactory, with high sensitivity but low specificity [22,23].
The risk of DVT following TKA is unpredictable, making therapy for these patients problematic. In modern medicine, the use of diagnostic tools in risk group stratification is crucial since it allows us to establish a treatment plan and prevent thromboembolic events. Furthermore, there is a lack of understanding in the existing literature about the diagnostic methods for DVT risk following joint arthroplasty.
The purpose of this study is to confirm the prognostic significance of hematological ratios (MLR, NLR, PLR, SII, SIRI, and AISI) in the acute DVT following TKA.

Study Design
The present study was designed as an observational, analytical, retrospective cohort case-control study and included all patients over 18 years of age with surgical indications for TKA, admitted to the Department of Orthopedics, Regina Maria Health Network, Targu Mures, Romania, and the Department of Orthopedics, Humanitas MedLife Hospital, Cluj-Napoca, Romania between January 2017 and July 2022. Patients having a history of DVT, patients who required post-operative blood transfusions, and patients who benefitted from bilateral TKA were all excluded.
Patients included in the study were initially divided into two groups depending on the presence of DVT at four weeks named "no-DVT" and "DVT". The ideal cut-off value for MLR and AISI was used to calculate the DVT risk and length of hospital stay.

Systemic Inflammatory Markers
The systemic inflammation index was determined from the first blood test result. The ratio was calculated using the equations: -MLR = total number of monocytes/total number of lymphocytes -NLR = total number of neutrophils/total number of lymphocytes -PLR = total number of platelets/total number of lymphocytes -SII = (total number of neutrophils × total number of platelets)/total number of lymphocytes -SIRI = (total number of monocytes × total number of platelets)/total number of lymphocytes -AISI = (total number of neutrophils × total number of monocytes × total number of platelets)/total number of lymphocytes

Knee Osteoarthritis Severity
Regarding the severity of knee osteoarthritis, we used one of the most frequently applied classifications based on pre-operative radiography, the Kellgren-Lawrence classification, which ranges from grade 0 to 4, depending on the progression and severity of osteoarthritis [42]. Furthermore, patients with an injury below Kellgren-Lawrance grade 3 were excluded from the research.

Surgical Technique
All patients had the same surgical method, which was undertaken by the same specialists. The Zimmer Biomet NexGen LPS implant was used for the procedure, and tranexamic acid was injected shortly after the joint capsule was closed. All patients were given intravenous antibiotics and low molecular weight heparin within the first 72 h of surgery. Anti-embolism stockings were used on all patients to avoid DVT. Furthermore, for postoperative days 3 to 14, an anticoagulant (rivaroxaban, Xarelto ® , 10 mg, Bayer AG, Leverkusen, Germany) was administered. Lastly, all patients began physical therapy on the first day post-surgery.

Study Outcomes
The primary endpoint was the risk of acute DVT, and the secondary endpoint was the length of hospital stay. Outcomes were stratified for the baseline's optimal MLR, NLR, PLR, SII, SIRI, and AISI cut-off value.

Follow-up Strategy
Post-operatively, all patients were evaluated with a doppler ultrasonography, to identify DVT before being discharged and again at four weeks. During the hospitalization, and following the TKA, none of the 273 participants in the study had acute DVT.

Statistical Analysis
SPSS for Mac OS version 28.0.1.0 was used for statistical analysis (SPSS, Inc., Chicago, IL, USA). Chi-square tests were used to assess the associations of hematological ratios with category factors, while t-Student or Mann-Whitney tests were used to assess differences in continuous variables. To analyze the predictive power and to establish the cut-off values of hematological ratios, the receiver operating characteristic (ROC) curve analysis was utilized. The ROC curve analysis was used to determine the appropriate MLR, NLR, PLR, SII, SIRI, and AISI cut-off values based on the Youden index (Youden Index = Sensitivity + Specificity − 1, ranging from 0 to 1). To identify independent predictors of DVT risk, a multivariate logistic regression analysis using variables with p < 0.1 was undertaken.

Results
During the study period, 339 patients were hospitalized, with 28 requiring postoperative blood transfusions, 22 having a history of DVT, and 16 having bilateral TKA. Throughout the research procedure, 273 patients were enrolled in the study. At four weeks, 28 patients (10.25%) had a DVT diagnosis confirmed by ultrasonography, although there was no indication of pulmonary embolism. Of the 28 patients with DVT, 22 had DVT below the knee, 3 had mild edema at the level of the afflicted limb, and the remaining 19 had no symptoms. The six patients who presented with DVT above the knee had significant edema and functional impotence at the afflicted leg level and required anticoagulant medication.
In terms of demographics, confirmed DVT patients were older (p = 0.0006). Additionally, cardiovascular pathologies [AH (p = 0.02), AF (p = 0.01)], malignancy (p = 0.005), as well as risk factors [smoking (p = 0.03) and obesity (p = 0.02)] were more prevalent in the DVT group. Moreover, in the case of laboratory analyses, there was a low level of hemoglobin (p = 0.02), hematocrit (p = 0.03), the total number of lymphocytes (p < 0.0001), and albumin (p = 0.002), as well as an increased level of the total number of neutrophils (p = 0.0001), monocytes (p = 0.001), platelets (p = 0.0007), BUN (p = 0.02), creatinine (p < 0.0001), potassium (p = 0.03), and hematological reports (for all p < 0.0001), as seen in Table 1. The ROC curves of MLR, NLR, PLR, SII, SIRI, and AISI were created to determine whether the baseline of these markers was predictive of acute DVT following the TKA (Figure 1). The optimal cut-off value obtained from Youden's index, areas under the curve (AUC), and the predictive accuracy of the markers are listed in Table 2.
The ROC curves of MLR, NLR, PLR, SII, SIRI, and AISI were created to determine whether the baseline of these markers was predictive of acute DVT following the TKA (Figure 1). The optimal cut-off value obtained from Youden's index, areas under the curve (AUC), and the predictive accuracy of the markers are listed in Table 2.  The DVT risk and length of hospital stay were further analyzed after dividing the patients into paired groups, according to the optimal cut-off value of inflammatory  The DVT risk and length of hospital stay were further analyzed after dividing the patients into paired groups, according to the optimal cut-off value of inflammatory biomarkers. Regarding the hospitalization time, there was a longer inpatient stay in the high-NLR (p = 0.01), high-SIRI (p = 0.02), and high-AISI group (p = 0.0004). Moreover, there was a higher incidence of DVT risk for all the hematological ratios, as seen in Table 3. A multivariate analysis was used to determine the association between the hematological ratios, the underlying risk factors, and DVT. A high baseline value of all hematological ratios was a strong independent predictor of DVT (for all p < 0.0001). Moreover, as indicated in Table 4  The Kaplan-Meier plot for the DVT risk in the first 4 weeks post-operation based on the optimal cut-off value of the hematologic ratios is shown in Figure 2. AISI = aggregate index of systemic inflammation; AF = atrial fibrillation; AH = arterial hypertension; DVT = deep vein thrombosis; MLR = monocyte to lymphocyte ratio; NLR = neutrophil to lymphocyte ratio; PLR = platelets to lymphocyte ratio; SII = systemic inflammatory index; SIRI = systemic inflammation response index.
The Kaplan-Meier plot for the DVT risk in the first 4 weeks post-operation based on the optimal cut-off value of the hematologic ratios is shown in Figure 2. cut off value (p <0.001; log-rank p), (B) according to NLR optimal cut-off value (p < 0.001; log-rank p), (C) according to PLR optimal cut-off value (p < 0.001; log-rank p), (D) according to SII optimal cut-off value (p < 0.001; log-rank p), (E) according to SIRI optimal cut-off value (p < 0.001; log-rank p), and (F) according to AISI optimal cut-off value (p < 0.001; log-rank p).

Discussion
The primary outcome of this research is that pre-operative hematological inflammatory markers have a strong predictive role in the risk of DVT incidence following TKA. We further confirm that cardiovascular disease (AH and AF), older age, malignancy, as well as smoking and obesity all predict DVT risk, as shown in Table 4. To the best of our knowledge, this is the first study to evaluate the predictive role of all hematological biomarkers and risk of DVT following TKA.
The incidence of acute DVT in patients undergoing TKA has been identified to be associated with elderly patients, a history of malignancy, and cardiovascular pathology [43][44][45]. TKA presents all three factors of the Virchow triad. Venous stasis occurs both intra-operatively, due to surgical immobilization and the use of a tourniquet, and shortly post-operatively, due to poor mobilization [46,47]. Endothelial damage is also unavoidable in major surgery such as TKA [47,48]. Moreover, the operational act causes local inflammation as well as a systemic inflammatory response that promotes a hypercoagulable condition [49][50][51].
Another risk factor for DVT following TKA is the use of tranexamic acid, which, although reducing the need for a post-operative blood transfusion by up to 39%, has been shown to have a predictive role in the occurrence of DVT in articles by Myers et al. [52], Henry et al. [53], and Ng et al. [54].

Discussion
The primary outcome of this research is that pre-operative hematological inflammatory markers have a strong predictive role in the risk of DVT incidence following TKA. We further confirm that cardiovascular disease (AH and AF), older age, malignancy, as well as smoking and obesity all predict DVT risk, as shown in Table 4. To the best of our knowledge, this is the first study to evaluate the predictive role of all hematological biomarkers and risk of DVT following TKA.
The incidence of acute DVT in patients undergoing TKA has been identified to be associated with elderly patients, a history of malignancy, and cardiovascular pathology [43][44][45]. TKA presents all three factors of the Virchow triad. Venous stasis occurs both intraoperatively, due to surgical immobilization and the use of a tourniquet, and shortly postoperatively, due to poor mobilization [46,47]. Endothelial damage is also unavoidable in major surgery such as TKA [47,48]. Moreover, the operational act causes local inflammation as well as a systemic inflammatory response that promotes a hypercoagulable condition [49][50][51].
Another risk factor for DVT following TKA is the use of tranexamic acid, which, although reducing the need for a post-operative blood transfusion by up to 39%, has been shown to have a predictive role in the occurrence of DVT in articles by Myers et al. [52], Henry et al. [53], and Ng et al. [54].
The predictive role of inflammatory biomarkers was analyzed and also demonstrated in the case of periprosthetic joint infection [39][40][41][55][56][57]. In the systemic review published by Festa et al. [39], it was demonstrated that high values of MLR, NLR, and PLR are associated with the detection of hip and knee periprosthetic infection.
In the work published by Yao [59].
Hematological ratios (MLR, NLR, PLR, SII, SIRI, and AISI) are measures of acute myeloid-driven innate immune responses reported to chronic, lymphocyte-driven, immunological memory reflected by lymphocyte numbers. An increased hematological ratio may reflect an immunological imbalance between a potential ongoing clinical or sub-clinical acute inflammation and an impaired immune defense. The importance of these hematological indicators in predicting coagulopathy risk and thromboembolic risk was widely researched [24,26,29].
Although this study included TKA patients from two private hospitals over a five-year period and had significant results in terms of the high level of sensitivity and specificity of the investigated inflammatory biomarkers [MLR (78.6% Sensitivity and 76.7% Specificity), NLR (78.6% Sensitivity and 73.5% Specificity), PLR (75% Sensitivity and 80.4% Specificity), SII (82.1% Sensitivity and 80.4% Specificity), SIRI (75% Sensitivity and 78.4% Specificity), and AISI (71.4% Sensitivity and 84.9% Specificity)] in the prediction of acute DVT, it has certain limitations. Firstly, we must consider the retrospective design of the study. Additionally, tranexamic acid was used for intra-articular injections in all included patients. Because this research was done using data from two private medical centers, we may expect the patients to have better health status.

Conclusions
Our data revealed that higher pre-operative hematological ratios MLR, NLR, PLR, SII, SIRI, and AISI values highly predict acute DVT following TKA. Moreover, during the studied period, age above 70, malignancy, cardiovascular disease, and risk factors such as obesity and tobacco were predictive risk factors for acute DVT. Given their accessibility and low cost, these ratios can be used for pre-operative risk group stratification, for better patient management regarding the DVT enhancement of thromboprophylaxis and for the development of predictive patterns.