Association between triglyceride glucose index and deep vein thrombosis following total knee arthroplasty: a retrospective case-control study

Abstract

Background

Total knee arthroplasty (TKA), a frequently performed procedure for treating severe knee arthritis, is associated with notable postoperative risks, particularly deep vein thrombosis (DVT). The triglyceride-glucose (TyG) index, a marker of insulin resistance, has been linked to cardiovascular and metabolic disorders, yet its role in predicting DVT following TKA remains underexplored.

Objective

This study aimed to assess the relationship between the TyG index and DVT incidence in patients undergoing TKA.

Methods

A retrospective case-control analysis was performed on patients who underwent TKA at the First Affiliated Hospital of Soochow University from January 2020 to January 2022. The TyG index, calculated using fasting triglyceride and glucose levels, was used to categorize patients into tertiles. Multivariate logistic regression models, smooth curve fitting, and threshold effect analyses were applied to investigate the TyG-DVT relationship. The predictive capacity of TyG for DVT was further examined via receiver operating characteristic (ROC) curve analysis.

Results

Higher TyG index values were strongly linked to an elevated risk of DVT post-TKA. In the fully adjusted model, the odds ratio for the highest TyG tertile (T3) was 36.82 (95% CI: 3.24–418.16, P = 0.0036). A threshold effect emerged at a TyG index value of approximately 9.36, where DVT risk surged rapidly below this point and plateaued above it. The area under the ROC curve for TyG in predicting DVT was 0.7283, reflecting moderate predictive power.

Conclusion

The TyG index shows potential as a valuable marker for identifying DVT risk in TKA patients, offering a means for early detection of high-risk individuals and enabling timely interventions to mitigate thrombotic complications.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-025-06863-z.

Introduction

Total knee arthroplasty (TKA) is a widely performed surgical intervention aimed at alleviating pain and restoring function in patients with advanced knee osteoarthritis [1]. In 2012, approximately 700,000 TKA procedures were conducted in the United States, with projections estimating this figure to reach 3.48 million by 2030 [2]. Despite its substantial clinical benefits, TKA is linked to considerable postoperative complications, with deep vein thrombosis (DVT) being one of the most prevalent and concerning outcomes [3]. The economic burden of venous thromboembolism treatment is significant, with an average cost of $5,000 per patient within the first three months post-onset, potentially rising to $33,000 per patient within a year [4, 5]. If left unchecked, DVT may progress to life-threatening conditions such as pulmonary embolism [6]. As a result, identifying high-risk individuals and implementing timely interventions are critical to optimizing surgical outcomes.

The triglyceride-glucose (TyG) index, calculated from fasting triglyceride (TG) and glucose (Glu) levels, has emerged as a reliable surrogate marker for insulin resistance [7]. Insulin resistance contributes to a range of metabolic abnormalities, including hypercoagulability, endothelial dysfunction, and chronic inflammation—key factors in DVT development [8]. Recent studies have established a strong association between the TyG index and cardiovascular events, such as coronary artery disease and ischemic stroke, underscoring its potential role as a vascular health marker [9, 10]. Patients undergoing TKA are particularly susceptible to DVT due to prolonged immobility, surgically induced vascular damage, and local hemodynamic alterations. While the link between the TyG index and subclinical atherosclerosis is increasingly recognized, its potential to predict venous thrombotic events in patients undergoing orthopedic joint replacement remains largely unexplored.

Considering these risk factors, the TyG index may serve as a valuable tool for assessing DVT risk in this population. This study aims to investigate the relationship between the TyG index and DVT incidence following TKA, with a focus on its predictive efficacy for post-surgical thrombotic events.

Methods

Study design and patience

This retrospective case-control study enrolled patients who underwent conventional TKA via the medial parapatellar approach for osteoarthritis at the First Affiliated Hospital of Soochow University between January 2020 and January 2022. The Inclusion criteria were a diagnosis of Kellgren-Lawrence grade III or IV osteoarthritis and unilateral TKA following the failure of conservative treatments. Exclusion criteria were: (1) preoperative diagnoses of rheumatoid arthritis (RA), villonodular synovitis, gouty arthritis, or hemophilic arthropathy; (2) history of trauma, previous surgeries, or infections affecting the involved lower limb; (3) prior DVT or pulmonary embolism in the lower extremity on the operated side; and (4) incomplete diagnostic data for perioperative thrombosis. After screening, a total of 188 participants were enrolled in the study (Fig. 1). The Institutional Review Board of the First Affiliated Hospital of Soochow University approved the study protocol (No. 2021140), and informed consent was waived due to the retrospective extraction of medical records.

Fig. 1.

Flow diagram of patient selection

Data collection

Trained personnel measured the weight and height using standardized procedures, and BMI was calculated as weight in kilograms divided by height in meters squared. Participants were classified as current smokers if they reported regular smoking in the past six months, and as current drinkers if they consumed alcohol at least once weekly in the same period. Blood pressure was measured on the non-dominant arm with an automated device (Omron model HEM-752 FUZZY; Omron Corporation, Dalian, China) after a five-minute seated rest. The average of three readings was used for analysis. Hypertension was defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, or current antihypertensive medication use. Diabetes was defined per American Diabetes Association guidelines: fasting glucose ≥ 126 mg/dL (7.0 mmol/L), 2-hour glucose ≥ 200 mg/dL (11.1 mmol/L) during an oral glucose tolerance test, Hemoglobin A1c ≥ 6.5%, or a self-reported history of diabetes diagnosis. Coagulation function, biochemical parameters, and fasting blood glucose levels were assessed within 2–3 days postoperatively, coinciding with the day of the postoperative ultrasound examination of both lower limbs. Blood samples for biochemical parameters and glucose levels were collected following a minimum 10-hour fasting period.

Evaluation of TyG index

Given that both TG and Glu levels in the laboratory reports were measured in mmol/L, the formula used to calculate the TyG index is as follows: TyG index = Ln [TG (mmol/L) * 88.57 * Glu (mmol/L) * 18 / 2]. To further investigate the dose-response relationship of TyG index with postoperative DVT, TyG was divided into tertiles, with the lowest tertile (T1) serving as the reference group.

DVT diagnosis and treatment strategies

Ultrasound examinations were conducted by at least two qualified sonographers, all of whom completed standardized training to ensure uniformity in imaging techniques and interpretation. All participants underwent bilateral lower extremity venous ultrasonography on either the second or third postoperative day, coinciding with the day of postoperative biochemical examinations. The diagnostic criteria for DVT followed established protocols, as described in previous studies [11, 12], which included: (1) the absence of venous blood flow on Doppler imaging; (2) complete non-compressibility of the vein; and (3) the presence of echogenic thrombi in veins that are typically anechoic. This methodical approach aimed to ensure precise and unbiased DVT detection in the postoperative context. All patients received prophylactic low molecular weight heparin (LMWH) starting on the first postoperative day to prevent lower extremity venous thrombosis. During hospitalization, LMWH was administered routinely, followed by prophylactic treatment with novel oral anticoagulants such as rivaroxaban after discharge, with a duration of at least 10 to 14 days and up to a maximum of 35 days. For patients who developed distal thrombosis postoperatively, rivaroxaban was prescribed for 2 weeks after discharge, followed by re-evaluation. Patients with proximal thrombosis were assessed by the interventional department for appropriate management, which included continuation of anticoagulant therapy or placement of an inferior vena cava filter as indicated.

Statistical methods

EmpowerStats and R software were used for statistical analysis. The measurement data were represented by mean ± sd. When the data were normally distributed, the independent sample t-test was used for comparison between the two groups. The rank sum test was used when the data were not normally distributed. Count data were expressed as rates, and pairwise comparison was made using the Chi-square test or Fisher’s exact probability method. Three logistic regression models with different covariates were used to explore the potential association between TyG and DVT. Model 1 provided unadjusted estimates, whereas Model 2 adjusted for age and sex. Model 3, the primary model, included further adjustments for body mass index (BMI), thrombin time (TT), fibrinogen (FIB), antithrombin III (AT III), fibrin degradation products (FDP), activated partial thromboplastin time (APTT), D-dimer (DD), international normalized ratio (INR), prothrombin time (PT), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), total cholesterol (TC), alcohol consumption, hypertension, diabetes, and smoking status. In addition, ensemble smoothing fitting and threshold effect analysis determined the nonlinear relationship and potential saturation point of the metrological effect. Receiver operating characteristic (ROC) curves were used to compare the predictive ability of TyG index with other lipid and blood glucose markers for DVT after TKA. P < 0.05 was considered statistically significant.

Result

Baseline characteristics

This study included 188 subjects with an average age of 68.20 ± 7.55 years. Finally, 18 patients were diagnosed with DVT after surgery, accounting for 9.57%. The majority were female, accounting for 71.28% of the cohort. Supplementary Table 1 showed no statistically significant differences (P > 0.05) in baseline characteristics between the DVT and non-DVT groups following TKA, including sex, age, BMI, alcohol consumption, smoking status, hypertension, diabetes, and postoperative coagulation and lipid profiles. However, the DVT group demonstrated significantly higher glucose levels (P = 0.015) and TyG index (P = 0.005) compared to the non-DVT group. To explore this further, the analysis was stratified by TyG index tertiles, as shown in Table 1. Significant positive associations were observed between the TyG index and several variables, including BMI (P < 0.001), AT III (P = 0.031), APTT (P = 0.007), PT (P = 0.033), HDL-C (P < 0.001), TC (P = 0.012), TG (P < 0.001), and glucose levels (P < 0.001). Furthermore, DVT occurrence was significantly more frequent in participants within the highest TyG index tertile (T3) compared to those in the lowest tertile (T1) (P = 0.001).

Table 1.

Baseline characteristics of the participants following TKA according to the TyG index

Characteristics	Tertiles of TyG index				P value
	Overall (N = 188)	T1 (N = 62)	T2 (N = 61)	T3 (N = 65)
Age (years)	68.20 ± 7.55	68.45 ± 7.60	67.49 ± 8.32	68.62 ± 6.75	0.672
Sex					0.150
Male	54 (28.72%)	23 (37.10%)	13 (21.31%)	18 (27.69%)
Female	134 (71.28%)	39 (62.90%)	48 (78.69%)	47 (72.31%)
Alcohol use					0.702
Yes	24 (12.77%)	9 (14.52%)	6 (9.84%)	9 (13.85%)
No	164 (87.23%)	53 (85.48%)	55 (90.16%)	56 (86.15%)
Hypertension					0.639
Yes	54 (28.72%)	20 (32.26%)	15 (24.59%)	19 (29.23%)
No	134 (71.28%)	42 (67.74%)	46 (75.41%)	46 (70.77%)
Diabetes					0.137
Yes	43 (22.87%)	10 (16.13%)	13 (21.31%)	20 (30.77%)
No	145 (77.13%)	52 (83.87%)	48 (78.69%)	45 (69.23%)
Smoking status					0.736
Yes	24 (12.77%)	7 (11.29%)	7 (11.48%)	10 (15.38%)
No	164 (87.23%)	55 (88.71%)	54 (88.52%)	55 (84.62%)
BMI (Kg/m2)	24.77 ± 1.92	24.77 ± 1.92	24.66 ± 1.53	25.88 ± 2.15	< 0.001
TT (sec)	18.88 ± 1.88	18.96 ± 3.03	18.81 ± 1.00	18.85 ± 0.81	0.908
FIB (g/L)	3.11 ± 0.71	3.02 ± 0.70	3.19 ± 0.74	3.11 ± 0.70	0.445
AT III (%)	89.95 ± 10.78	87.32 ± 12.24	90.05 ± 11.02	92.35 ± 8.39	0.031
FDP (mg/L)	2.67 ± 3.14	2.50 ± 2.40	3.03 ± 4.49	2.48 ± 2.05	0.541
APTT (sec)	26.61 ± 2.48	27.40 ± 3.29	26.36 ± 1.91	26.09 ± 1.82	0.007
DD (mg/L)	0.91 ± 1.43	0.83 ± 1.00	1.08 ± 2.16	0.83 ± 0.78	0.537
INR	1.00 ± 0.05	1.01 ± 0.06	1.00 ± 0.05	0.99 ± 0.06	0.139
PT (sec)	11.51 ± 0.62	11.67 ± 0.66	11.46 ± 0.51	11.39 ± 0.65	0.033
LDL-C (mmol/L)	2.96 ± 0.82	2.94 ± 0.82	3.09 ± 0.85	2.85 ± 0.78	0.274
HDL-C (mmol/L)	1.12 ± 0.24	1.16 ± 0.24	1.18 ± 0.25	1.02 ± 0.19	< 0.001
TC (mmol/L)	4.95 ± 0.90	4.68 ± 0.80	5.02 ± 0.98	5.13 ± 0.87	0.012
TG (mmol/L)	1.83 ± 1.38	0.97 ± 0.26	1.50 ± 0.28	2.96 ± 1.82	< 0.001
Glu (mmol/L)	5.62 ± 1.22	5.14 ± 0.80	5.57 ± 1.10	6.11 ± 1.47	< 0.001
DVT					0.001
Yes	18 (9.57%)	1 (1.61%)	4 (6.56%)	13 (20.00%)
No	170 (90.43%)	63 (100.00%)	55 (91.67%)	52 (80.00%)

Table 2 presented the relationship between TyG index and DVT risk as a continuous variable and as a three-group variable. The results consistently demonstrated a strong association between higher TyG levels and increased DVT susceptibility across all three models. In the unadjusted Model 1, the odds ratio (OR) for TyG was 2.98 (95% confidence interval [CI]: 1.34–6.60, P = 0.0072). In Model 2, adjusted for sex and age, the OR remained significant at 3.03 (95% CI: 1.36–6.77, P = 0.0069). In the fully adjusted Model 3, which accounted for confounding factors such as BMI, coagulation parameters, lipid profiles, alcohol consumption, hypertension, diabetes, and smoking status, the OR increased to 8.82 (95% CI: 2.06–37.72, P = 0.0033). The analysis of TyG tertiles indicated a significant trend of increasing DVT risk, with the highest tertile (T3) showing a markedly elevated risk, particularly in the fully adjusted model (OR = 36.82, 95% CI: 3.24–418.16, P = 0.0036). This analysis confirmed a robust and independent association between elevated TyG levels and DVT risk.

Table 2.

Association between TyG and DVT risk

	OR (95%CI), P value
	Model 1	Model 2	Model 3
TyG index	2.98 (1.34, 6.60) 0.0072	3.03 (1.36, 6.77) 0.0069	8.82 (2.06, 37.72) 0.0033
T1	1.0	1.0	1.0
T2	4.28 (0.46, 39.43) 0.1993	4.27 (0.46, 39.69) 0.2021	6.03 (0.54, 67.75) 0.1452
T3	15.25 (1.93,120.48) 0.0098	15.56 (1.96, 123.54) 0.0094	36.82 (3.24, 418.16) 0.0036
P for trend	12.44 (2.69, 57.60) 0.0013	12.82 (2.74, 59.92) 0.0012	31.48 (4.26, 232.53) 0.0007

OR Odds ratio, 95% Cl 95% Confidence interval

Model 1: Unadjusted, with no covariates included

Model 2: Adjusted for age and sex

Model 3: Adjusted for age, sex, BMI, TT, FIB, AT III, FDP, APTT, DD, INR, PT, LDL-C, HDL-C, TC, alcohol consumption, hypertension, diabetes, and smoking status

Figure 2; Table 3 presented the findings of smooth curve fitting and threshold effect analysis. Across all three models, an inflection point was observed at a TyG index value of approximately 9.36. Below this threshold, DVT risk escalated sharply with rising TyG levels (all P < 0.05). However, once TyG levels surpassed the inflection point, the correlation between TyG and DVT risk diminished, suggesting a plateau effect. These results highlighted a nonlinear relationship between the TyG index and DVT, with a pronounced increase in risk below the identified threshold.

Fig. 2.

Smooth curve fitting analysis of the relationship between TyG and DVT. The solid red line represents the smooth curve fit illustrating the relationship between variables, with blue bands indicating the 95% confidence intervals derived from the fit. Panel (A) shows the relationship without adjusting for covariates. In panel (B), adjustments were made for age and sex. Panel (C) incorporates further adjustments, accounting for age, sex, BMI, TT, FIB, AT III, FDP, APTT, DD, INR, PT, LDL-C, HDL-C, TC, alcohol consumption, hypertension, diabetes, and smoking status

Table 3.

Threshold effect analysis of the TyG index on DVT risk using a two-segment piecewise linear regression model

DVT	Adjusted OR (95% CI), P value β (95% CI) P value β (95% CI) P valueβ (95% CI) P value
	Model 1	Model 2	Model 3
TyG index
The standard linear mode	2.98 (1.34, 6.60) 0.0072	3.03 (1.36, 6.77) 0.0069	8.82 (2.06, 37.72) 0.0033
Inflection point	9.36	9.36	9.36
TyG < Inflection point	17.41 (2.71, 111.88) 0.0026	17.83 (2.75, 115.57) 0.0025	50.19 (4.73, 532.92) 0.0012
TyG > Inflection point	0.24 (0.02, 3.33) 0.2905	0.25 (0.02, 3.42) 0.2975	0.28 (0.01, 8.55) 0.4646
Log-likelihood ratio	0.015	0.015	0.020

Model 1: Unadjusted, with no covariates included

Model 2: Adjusted for age and sex

Model 3: Adjusted for age, sex, BMI, TT, FIB, AT III, FDP, APTT, DD, INR, PT, LDL-C, HDL-C, TC, alcohol consumption, hypertension, diabetes, and smoking status

Predictive efficacy of TyG for DVT

The ROC analysis depicted in Fig. 3 reveals that the TyG index achieved the highest area under the curve (AUC) of 0.7283 (95% CI: 0.6231–0.8335) among lipid parameters for predicting DVT. The optimal TyG index threshold was determined to be 8.990, yielding a sensitivity of 0.7222 and a specificity of 0.7118. In contrast, TG exhibited a relatively high AUC of 0.7008 (95% CI: 0.6125–0.7891), with elevated sensitivity (0.9444) but reduced specificity (0.4824). Other lipid markers, such as LDL-C, HDL-C, and TC, demonstrated weaker predictive capacities, with AUCs of 0.5703, 0.5291, and 0.5877, respectively.

Fig. 3.

ROC curve analysis for predicting DVT

Discussion

In the present study, patients who developed DVT after TKA had significantly elevated TyG index compared with patients who did not develop DVT. As the TyG index increases, the risk of DVT increases. Logistic regression analysis and smooth fitting curve confirmed the positive nonlinear relationship between TyG index and DVT development. The ROC curve further confirmed the ability to predict DVT based on the TyG index. These results suggest that TyG index, a simple tool, may help in the early identification of high-risk groups in need of timely intervention.

The TyG index is widely used in clinical and epidemiological studies as a surrogate for assessing insulin resistance because of its convenient calculation and high cost-effectiveness. Through the analysis of triglyceride and blood glucose levels, the TyG index can help doctors preliminarily assess metabolic status without complex tests, especially in resource-limited settings. Over the past five years, in addition to diabetes and metabolic syndrome, some studies have shown that the TyG index is also associated with a variety of health problems and their prognosis, such as fatty liver [13], hypertension [14], stroke [15], depression [16], erectile dysfunction [17], osteoporosis [18], lung disease [19], and kidney disease [20], making it a versatile health assessment tool.

Since our initial publication linking an elevated TyG index to increased arthritis risk, an extensive literature has explored its potential connection to joint diseases [7]. Huang et al. [21], analyzing data from 25,514 NHANES participants (2015–2020), reported a 6.34-fold increase in osteoarthritis risk for each unit rise in the TyG index. Similarly, Li et al. [22], examining 11,768 participants from the 2007–2017 NHANES database, found a 40% higher likelihood of gout for each unit increase in the TyG index. Additionally, Liu et al. [23] confirmed the TyG index’s predictive value for arthritis onset in individuals over 45 years through a longitudinal study of 4,418 participants in the China Health and Retirement Longitudinal Study. Beyond its utility in predicting arthritis, the TyG index’s association with cardiovascular risk in patients with arthritis has been extensively investigated. Wang et al. [24] analyzed a cohort of 418 patients with RA, revealing that an elevated TyG index independently predicted a heightened cardiovascular disease risk. By integrating the TyG index with the rheumatoid factor, they constructed an ROC curve, achieving an AUC of 0.791, thereby proposing the TyG index as a robust predictor of cardiovascular risk in patients with RA. Similarly, Zang et al. [25] conducted a large-scale study with 1,613 patients with RA, tracking the incidence of major adverse cardiovascular events (MACEs) over a median follow-up of 5.32 years. Their findings indicated that a higher TyG index was independently associated with an increased MACEs risk in male participants, further underscoring its predictive value for cardiovascular events in RA men. Elazab et al. [26], in a case-control study involving 100 patients with RA and 50 matched healthy controls, reported a positive correlation between the TyG index and both RA disease activity and subclinical atherosclerosis. Additionally, Xie et al. [27] examined 165 patients with psoriatic arthritis who underwent serial carotid ultrasounds, identifying a positive association between the TyG index and atherosclerotic burden, independent of conventional cardiovascular risk factors and psoriasis-specific variables. These results suggest the TyG index could serve as a marker for atherosclerosis in patients with psoriatic arthritis. While numerous studies have examined the TyG index’s role in predicting cardiovascular risk in osteoarticular diseases, its link to thromboembolic events following TKA in patients with arthritis remains insufficiently explored.

With the growing interest in the TyG index in recent years, its association with venous thrombosis across various conditions has been investigated, yielding somewhat varied results. In an observational case-control study by Katipoglu et al. [28], which included 492 individuals aged 40 to 90 recruited from an ophthalmology outpatient clinic in Turkey, the TyG index was significantly higher in those with retinal vein occlusion (RVO) compared to controls. An elevated TyG index was associated with a roughly 50% increased risk of developing RVO, leading the researchers to suggest its potential as a reliable marker for identifying high-risk individuals and enabling timely preventive interventions. Similarly, Aslan [29] demonstrated that the TyG index independently predicts branch retinal vein occlusion (BRVO), with a ROC analysis yielding an AUC of 0.749. At a TyG index cutoff of 8.52, the sensitivity and specificity for predicting BRVO were 83% and 70%, respectively. Wang et al. [30] conducted a related study involving 4,748 non-diabetic individuals from the Asymptomatic Polyvascular Abnormalities Community study, exploring the relationship between the TyG index and carotid plaque stability in adults without diabetes. While no significant correlation was found between the TyG index and stable carotid plaques, a higher TyG index was significantly associated with an increased incidence of unstable plaques in this population. Since orthopedic surgery is a known risk factor for lower extremity venous thrombosis, investigating the link between the TyG index and postoperative outcomes in TKA for patients with osteoarthritis has become an area of interest. Our results also showed a positive correlation between the TyG index and postoperative DVT in elderly patients, demonstrating its potential value.

This study represented the first real-world investigation, according to available literature, exploring the relationship between the TyG index and post-TKA outcomes in patients with osteoarthritis. The findings indicate that this association is robust across populations and remains significant even after adjusting for multiple confounders in multivariable models. These results suggest that the TyG index could be a valuable marker for DVT risk following TKA, aiding in the early identification of patients at higher risk. While the underlying mechanisms are not fully understood, several hypotheses have been proposed. The TyG index, a surrogate marker for insulin resistance, integrates fasting triglyceride and glucose levels, reflecting metabolic disturbances that contribute to a pro-thrombotic state. Insulin resistance, indicated by a high TyG index, leads to hyperglycemia and dyslipidemia, both of which are critical factors in endothelial dysfunction. Hyperglycemia induces the formation of advanced glycation end-products, which promote oxidative stress and inflammatory responses, thereby damaging endothelial cells and impairing their ability to maintain vascular homeostasis [31]. Concurrently, elevated triglycerides contribute to the development of atherosclerosis by modifying lipoprotein particles, making them more atherogenic and prone to plaque formation, which can destabilize the vascular environment and facilitate thrombus formation [32, 33]. Furthermore, insulin resistance is associated with increased levels of inflammatory cytokines such as interleukin-6 and tumor necrosis factor -α, which activate the coagulation cascade by upregulating tissue factor expression and inhibiting natural anticoagulant pathways, thereby enhancing the propensity for clot formation [34, 35]. Additionally, a high TyG index is linked to elevated plasminogen activator inhibitor-1 levels, which impede fibrinolysis and promote the stability and persistence of thrombi [36, 37]. Endothelial dysfunction, characterized by reduced nitric oxide bioavailability due to insulin resistance and dyslipidemia, results in vasoconstriction, increased vascular permeability, and heightened platelet adhesion and aggregation, all of which are conducive to thrombosis, especially in the perioperative setting of TKA where surgical trauma further exacerbates endothelial injury [38, 39]. Moreover, insulin resistance can enhance platelet reactivity through oxidative stress, leading to the expression of pro-thrombotic markers such as P-selectin and the release of factors that facilitate platelet aggregation [40, 41]. Additionally, microvascular dysfunction resulting from insulin resistance impairs blood flow and oxygen delivery, creating a hypoxic environment that upregulates hypoxia-inducible factors and further stimulates pro-coagulant gene expression, thereby fostering thrombus formation [42, 43].

Study strengths and limitations

This study is the first to explore the association between the TyG index and thrombus formation in patients with arthritis following TKA, offering novel insights into the potential application of the TyG index in assessing perioperative thrombotic risk. As a readily measurable metabolic marker, the TyG index may serve as a practical and supplementary tool in clinical settings, complementing bilateral lower extremity venous assessments, particularly for preoperative evaluation and perioperative management. It holds promise in identifying and managing high-risk patients.

Nonetheless, several limitations should be acknowledged. Retrospective design inherently introduces vulnerabilities to biases, such as selection and information bias, which could influence the findings. Additionally, the single-center nature of the study may constrain the generalizability of the results. Variations in surgical techniques, patient demographics, and perioperative management across different institutions could affect the applicability of the observed link between the TyG index and thrombotic risk. Third, the timing of fasting blood glucose measurements (mostly on postoperative day 2, with a small minority on day 3) might have introduced variability in the TyG index calculation, which could affect the results. Moreover, although the study adjusted for multiple covariates in the multivariate analysis, residual confounding may persist, particularly regarding unmeasured or unaccounted-for factors, potentially affecting the observed association between the TyG index and thrombotic risk.

Conclusions

This study revealed a significant correlation between elevated TyG index levels and an increased risk of DVT in post-TKA patients. As an indirect marker of insulin resistance, the TyG index proved to be a reliable predictor of DVT, with the risk rising notably to a threshold value of approximately 9.36. These results indicate that the TyG index may serve as a simple and effective tool for the early identification of high-risk patients following TKA, potentially enhancing clinical outcomes through targeted preventive interventions.

Abbreviations

APTT

Activated partial thromboplastin time

AT III

Antithrombin III

AUC

Area under the curve

BMI

Body mass index

BRVO

Branch retinal vein occlusion

CI

Confidence interval

DVT

Deep vein thrombosis

DD

D-dimer

FDP

Fibrin degradation products

FIB

Fibrinogen

Glu

Glucose

HDL-C

High-density lipoprotein cholesterol

INR

International Normalized Ratio

LDL-C

Low-density lipoprotein cholesterol

LMWH

Low molecular weight heparin

MACEs

Major adverse cardiovascular events

NHANES

National Health and Nutrition Examination Survey

OR

Odds ratio

PT

Prothrombin time

ROC

Receiver operating characteristic

RVO

Retinal vein occlusion

RA

Rheumatoid arthritis

TC

Total cholesterol

TT

Thrombin time

TKA

Total knee arthroplasty

TG

Triglycerides

TyG

Triglyceride-glucose

Authors’ contributions

Qian Wu: Conceptualization, Methodology, Investigation, Writing – original draft. Wenquan Ding: Conceptualization, Methodology, Data curation, Investigation. Jiacheng Du: Conceptualization, Methodology, Data curation, Investigation.Lingfen Liu: Conceptualization, Methodology, Data curation, Validation.Rui Xie: Conceptualization, Methodology, Data curation, Validation. Zhanghuan Chen: Conceptualization, Methodology, Data curation. Dinghua Jiang: Conceptualization, Methodology, Data curation. Liyu Zhou: Conceptualization, Methodology, Data curation, Validation, Writing—review & editing, Funding acquisition.Lixin Huang: Conceptualization, Methodology, Data curation, Validation, Writing—review & editing, Funding acquisition. Wu Xu: Conceptualization, Methodology, Data curation, Validation, Writing—review & editing, Funding acquisition.

Data availability

Data are available upon request to the corresponding author.

Declarations

Ethics approval and consent to participate

Clinical trial number: not applicable. This study was a retrospective analysis of medical records and did not involve a prospective clinical trial. This study was approved by the Institutional Review Board of the First Affiliated Hospital of Soochow University (No.2021140). As the data were retrieved retrospectively from medical records, informed consent was not required.

Competing interests

The authors declare no competing interests.