Association of preoperative serum 25-hydroxyvitamin D levels with deep vein thrombosis after total knee arthroplasty: a hypothesis-generating analysis of coagulation function mechanisms

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Objective This hypothesis-generating study aimed to explore the association of preoperative serum 25-hydroxyvitamin D [25(OH)D] levels with deep vein thrombosis (DVT) risk after total knee arthroplasty (TKA) and clarify the potential coagulation-related mechanisms. Methods A total of 193 patients undergoing primary unilateral TKA in a single center were stratified into 25(OH)D deficiency (<20 ng/mL, n = 106), insufficiency (20–29 ng/mL, n = 47), and sufficiency (≥30 ng/mL, n = 40) groups. Postoperative DVT (detected via day 5–7 venous ultrasonography) and coagulation indices [D-dimer, prothrombin time (PT)] were compared across groups. Univariate and multivariate logistic regression were used to screen DVT-related factors, and Firth's penalized logistic regression was applied for a sparse-data sensitivity analysis. Results The overall postoperative DVT incidence was 13.5% (26/193). A bivariate analysis revealed a markedly higher DVT rate in the deficiency group (18.9%) than in the sufficiency group (2.5%), with a crude 2 × 2 odds ratio (OR) of 9.070 [95% confidence interval (CI): 1.198–68.542, P = 0.015]; this estimate was unstable due to only one DVT case in the sufficiency group. To provide a more stable estimate, a secondary binary logistic regression model (deficiency vs. non-deficiency) was performed, which identified preoperative 25(OH)D deficiency as significantly correlated with postoperative DVT (OR = 3.140, 95% CI: 1.200–8.212, P = 0.020). After analysis of covariance adjustment for baseline levels, the deficiency group maintained significantly higher postoperative D-dimer and shorter PT than the sufficiency group (both P < 0.001). A univariate analysis identified 25(OH)D deficiency and elevated postoperative D-dimer as DVT-related risk factors. In the secondary exploratory model adjusting for postoperative D-dimer (a potential mediator), the association between vitamin D deficiency and DVT was attenuated (OR = 2.576, 95% CI: 0.941–7.049, P = 0.065), indicating a possible mediating effect rather than confounding bias. Firth's sensitivity analysis on the binary model further verified the robustness of this finding (OR = 2.918, 95% CI: 1.142–7.456, P = 0.025). Conclusion Preoperative 25(OH)D deficiency is associated with elevated postoperative DVT risk after TKA in univariate analyses. Abnormal postoperative D-dimer and PT levels suggest that vitamin D status may modulate coagulation-fibrinolysis balance. The attenuated association after D-dimer adjustment supports a potential mediating mechanism. As a hypothesis-generating analysis, these preliminary findings require further validation in large-scale prospective studies to confirm the independent predictive value of preoperative vitamin D deficiency for post-TKA DVT.