Platelet count, fibrinogen and ROTEM® parameters predict deep vein thrombosis in elderly patients with femoral fracture

Background: Deep vein thrombosis (DVT) is a frequent complication in elderly patients undergoing surgery for proximal femoral fractures. Identifying early coagulation and viscoelastic changes may help to personalize thromboprophylaxis strategies. This study evaluated the temporal evolution of coagulation and ROTEM parameters and explored whether baseline values and early changes could predict the development of asymptomatic DVT. Methods: We conducted a prospective observational study including patients ≥ 80 years undergoing surgical repair of proximal femoral fractures. Conventional coagulation tests and ROTEM were performed at admission (T0) and on postoperative day 3 (T3). Compression ultrasonography was used to screen for DVT on day 5 (T5). Longitudinal changes were assessed with linear mixed-effects models. Results: Among 40 enrolled patients, 7 (17.5%) developed asymptomatic DVT. Platelet count was consistently higher in DVT patients (baseline: 275.0 ± 47.1 vs 186.5 ± 54.9 × 109/L; p < 0.001). ROTEM revealed increased EXTEM maximum clot firmness (67.6 ± 1.7 vs 62.2 ± 5.7 mm; p = 0.024) and shorter clot formation time (92.0 ± 9.3 vs 116.1 ± 32.0 s; p = 0.041). Fibrinogen levels diverged over time: while baseline values were comparable, DVT patients showed progressive increases (+ 43.7 mg/dL/day) versus reductions in non-DVT patients (-9.5 mg/dL/day), yielding significant differences at T3 (525.3 ± 177.0 vs 372.1 ± 177.8 mg/dL; p = 0.015). FIBTEM-MCF did not differ between groups. A baseline platelet count ≥ 277 × 109/L predicted DVT with 71% sensitivity and 97% specificity (AUC 0.896). Conclusion: In elderly patients with femoral fractures, asymptomatic DVT was associated with elevated baseline platelet counts, enhanced clot firmness, and progressive fibrinogen elevation over time. These findings suggest multifaceted hypercoagulable abnormalities with temporal monitoring potentially superior to single-timepoint assessments for DVT prediction.