Prospective surveillance for perioperative venous thrombosis. Experience in 2643 patients.

BACKGROUND: Patients who undergo neurosurgical procedures are at high risk for perioperative deep vein thrombosis (DVT) and pulmonary embolism (PE), which have been reported in 6% to 43% of these patients.

OBJECTIVES: To (1) determine the utility of prospective DVT surveillance in patients who undergo neurosurgical procedures by using venous duplex ultrasound scanning (VDUS), (2) assess the efficacy of DVT prophylaxis (elastic stockings and intermittent pneumatic compression), (3) identify subgroups of patients who are at higher risk, and (4) determine whether DVT surveillance would reduce the incidence of fatal PE.

DESIGN: All patients had undergone preoperative VDUS of both lower extremities, and postoperative VDUS was performed on days 3 and 7, and weekly thereafter until patients were ambulatory or discharged.

PATIENTS: During a 5-year period, 2643 patients who underwent neurosurgical procedures were enrolled in prospective DVT surveillance.

SETTING: University-affiliated community hospital.

RESULTS: Acute DVT was diagnosed in 147 (5.6%) of the 2643 patients. Eighty-one percent of the patients with acute DVT were asymptomatic at the time of diagnosis. Deep vein thrombosis developed de novo in the proximal veins in 98% of the patients. Patients in whom a craniotomy was done had a significantly higher risk for DVT (7.7%, P = .006), and patients who underwent cervical or lumbar spinal surgical procedures had a significantly lower risk (1.5%, P < .001). Among those patients in whom a craniotomy was performed for treatment of a tumor and who had DVT, 87% had malignant neoplasms. Significant lower-extremity neuromotor dysfunction was present in 69% of all patients with DVT, and this finding predominated among patients with DVT in the subgroups with a lower risk. A PE was diagnosed in 5 patients (0.19%) while they were hospitalized, and a PE was fatal in 2 (0.07% of all patients).

CONCLUSIONS: Most perioperative DVTs were clinically silent and formed spontaneously in proximal venous segments where there would be a risk for a PE. The overall incidence of DVT (5.6%) was low, suggesting effective DVT prophylaxis. Patients who underwent spinal surgical procedures were at a significantly lower risk for DVT, and future surveillance is not indicated in this patient group unless other conditions exist (paralysis, malignancy). Patients in whom a craniotomy was performed had a significantly higher risk of DVT, particularly when other risk factors existed. The low incidence of a fatal PE (0.07%) reflected that early detection and treatment of proximal DVT were facilitated by prospective VDUS surveillance in these patients.