We studied 1162 consecutive total hip replacements (THR) to establish the incidence of fatal pulmonary embolism (PE), clinical non-fatal PE and deep-vein thrombosis (DVT) in the six months after surgery. Chemical thromboprophylaxis had not been routinely used. We used a validated questionnaire supplemented by post-mortem records and a review of the clinical notes. Follow-up was 100%. The death rate from PE was 0.34% (95% CI 0.09 to 0.88), with one fatal PE after discharge 40 days after operation. The clinical PE rate confirmed by imaging was 1.20% (CI 0.65 to 2.02), with 0.7% of patients readmitted. The venographically-confirmed clinical DVT rate was 1.89% (CI 1.11 to 2.76), with 1.13% readmitted. The total thromboembolic morbidity was 3.4% (95% CI 2.5% to 4.7%); prophylaxis to reduce this would be justifiable if the complications of such prophylaxis did not produce an alternative morbidity. The fatal PE rate after THR without routine chemical prophylaxis was low; a very large randomised clinical trial would be required to demonstrate directly whether any prophylactic measure could reduce this. There is a large discrepancy between the high DVT rate reported in clinical trials using universal screening venography and the symptomatic DVT rate shown in this study. We found insufficient evidence to recommend continuing thromboprophylaxis after discharge from hospital.

We examined ten femoral veins with duplex ultrasound during total hip replacement to demonstrate the operative manoeuvres which cause venous obstruction and to assess prophylactic measures which may overcome it. Exposure of the acetabulum by distraction of the femur with a hook was less likely to occlude flow than retraction with bone levers. Adequate exposure of the femoral shaft by adduction, flexion and either internal or external rotation caused cessation of flow in all cases. In four cases an A-V Impulse System foot pump was activated during periods of stasis. In each case it overcame the obstruction and produced peak velocities which were twice that of the resting state. In five cases, towards the end of the procedure, debris was seen travelling proximally through the femoral vein.