A binary logistic regression analysis of deep venous thrombosis after hip fracture surgery

Abstract

Objective To explore the incidence, etiological and protective factors, and preventive countermeasures of deep venous thrombosis (DVT) in patients undergoing hip surgery. Methods Four hundred and forty-one patients who had hip fracture surgery in our department from January, 2005 to May, 2010were analyzed retrospectively. They were 223 males and 218 females, aged 65.7 years. (range, 22 to 87). A binary logistic regression was performed, using gender, age, duration of surgery, surgery type, anesthesia type, blood transfusion, physical therapy, thrombo-prophylaxis and complication as covariates, taking DVT incidence as the dependant variable. Results The patients obtained a median follow-up of 17.2 months (range, 3 months to 4 years and 7 months). Sixty-nine patients developed DVT (incidence, 15.6%).Analysis of multiple variables showed that duration of surgery, anesthesia type, blood transfusion, hypertension and diabetes were risk factors to DVT. Physical therapy and thrombo-prophylaxis were protective factors.Age, gender and surgery type had no statistically significant influence on DVT (P ＞ 0.05). Conclusions The DVT incidence after hip fracture surgery is high. In order to decrease the DVT incidence, surgeons should try their best to estimate risk factors for every patient perioperatively, avoid general anesthesia or blood transfusion, shorten operating duration, effectively control complications, actively perform physical therapy or thrombo-prophylaxis, and prolong the anticoagulant therapy for patients with higher risk factors. Key words: Venous thrombosis; Hip fractures; Regression analysis