Elderly hip fracture patients with isolated calf muscle vein thrombosis are more likely to have suffered an intertrochanteric fracture and be hypertensive

Abstract

Purpose: This study aimed to characterize the preoperative deep venous thrombosis (DVT) of lower extremity by locations and evaluate the diagnostic ability of plasma D-dimer level in elderly patients with hip fracture.

Method: This retrospective study reviewed the elderly patients presenting with a hip fracture definitely undergoing surgical treatment and identified those who had preoperative DVT. Data on demographics, lifestyle habits, comorbidities and laboratory indexes were extracted and collected. Four groups were divided by presence and locations of thrombi: non-DVT (controls), isolated calf muscle vein thrombosis (CMVT), deep calf vein thrombosis (DCVT) and proximal DVT group. The comparisons were conducted between either DVT group or the non-DVT group. Receiver operating characteristic (ROC) curve and the area under the curve (AUC) were used for determining the diagnostic ability of D-dimer for each group.

Results: Among 951 eligible patients included, 298 (31.3%) were found have preoperative DVT. Compared to non-DVT group, patients with CMVT had significantly lower albumin and hemoglobin concentration, more likely an intertrochanteric fracture and a higher prevalence of hypertension (P = 0.001, 0.006, 0.002 and 0.024, respectively); DCVT group was not observed to be significantly different in terms of any variables (all P > 0.05); and proximal DVT group had older age and more patients ≥ 80 years, lower albumin and hemoglobin concentration, higher prevalence of hypertension, more likely an intertrochanteric fracture, extended time from injury to imaging examination and higher age-adjusted Charlson comorbidity index (ACCI) (all P < 0.05). D-dimer demonstrated nonsignificant diagnostic ability for CMVT and DCVT, and a significant but poor ability for proximal DVT (AUC, 0.621; 95% CI 0.542-0.710; P = 0.011).

Conclusion: Varying characteristics were found among preoperative DVT stratified by locations after elderly hip fractures, and D-dimer level demonstrated no or poor diagnostic ability for DVTs by locations. Level of evidence level III, diagnostic. Trial registration statement Not applicable.

Keywords: Clinical epidemiology; D-dimer; Deep venous thrombosis; Diagnostic performance; Hip fracture.

Fig. 1

The ROC and AUC analysis for D-dimer in diagnosis of isolated CMVT, demonstrating no ability with AUC of 0.509 (95% CI 0.467–0.552; P = 0.675)

Fig. 2

The ROC and AUC analysis for D-dimer in diagnosis of DCVT, demonstrating poor but statistically significant ability with AUC of 0.0.621 (95% CI 0.542–0.710; P = 0.011)

Fig. 3

The ROC and AUC analysis for D-dimer in diagnosis of proximal DVT, demonstrating no ability with AUC of 0.532 (95% CI 0.438–0.626; P = 0.542)