In-hospital deep vein thrombosis after tibial plateau fractures: incidence, laterality/anatomy, and risk factors in a multicenter retrospective cohort of 3366 patients.

backgroundWe aimed to determine the cumulative in-hospital incidence, anatomic distribution, and predictors of duplex ultrasonography-detected lower-extremity deep vein thrombosis (DVT) in patients hospitalized with tibial plateau fractures under routine thromboprophylaxis and protocolized surveillance.

methodsWe retrospectively analyzed consecutive adults (≥ 18 years) admitted to two level-I trauma centers (November 2014-January 2024). Bilateral duplex ultrasonography was performed on admission/preoperatively (as soon as feasible after arrival and prior to surgery) and repeated serially during hospitalization (approximately every 5-7 days and when clinically suspected). The primary outcome was cumulative in-hospital DVT from admission to discharge; isolated intermuscular calf-muscle vein thrombosis (soleal/gastrocnemius) was recorded descriptively but excluded from the primary endpoint. Multiple injuries (polytrauma) were defined as ≥ 1 additional acute traumatic lesion beyond the index tibial plateau fracture documented on admission. Pulmonary embolism (PE) was assessed only when clinically suspected and confirmed by computed tomography pulmonary angiography (CTPA). Multivariable logistic regression identified independent predictors; receiver operating characteristic (ROC) analysis evaluated discrimination for key continuous predictors.

resultsAmong 3366 patients, 675 developed in-hospital DVT (20.0%). First detection occurred preoperatively in 432 (64.0%) patients and postoperatively in 243 (36.0%). Most events were distal (584/675, 86.5%) and ipsilateral (617/675, 91.4%); proximal DVT occurred in 91 patients (2.7% of the cohort). Independent predictors included age, multiple injuries, anemia, alcohol use, and inverse associations with activated partial thromboplastin time and serum sodium. ROC analysis of age showed modest discrimination (area under the curve [AUC] 0.593) with a cohort-derived Youden threshold of 42 years (age ≥ 42 years). Symptomatic PE occurred in five patients (0.15% of the cohort) and was confirmed by CTPA.

conclusionsUnder protocolized inpatient duplex surveillance, patients with tibial plateau fracture had a substantial in-hospital burden of predominantly distal and ipsilateral DVT. Increasing age was independently associated with DVT; however, age alone showed limited discriminatory ability (AUC 0.593), with a cohort-derived Youden threshold of 42 years (age ≥ 42 years). Multiple injuries, anemia, alcohol use, shorter activated partial thromboplastin time (APTT), and lower serum sodium were independently associated with DVT, while the discrimination of age alone was modest. These findings may help inform risk stratification and the consideration of systematic or risk-adapted surveillance and individualized perioperative management; prospective studies are needed to determine whether risk-tailored strategies improve clinically meaningful outcomes.