• ABSTRACT
    • Deep venous thrombosis (DVT) complicates 60% of knee and 80% of hip arthroplasties performed without prophylactic therapy. Routine postoperative duplex ultrasound surveillance has been proposed for the detection of venous thrombosis following arthroplasty. In order to determine whether surveillance represents an effective strategy to detect postoperative DVT when prophylaxis is used, surveillance duplex exams obtained after primary or revision hip or knee arthroplasty were analyzed using decision analysis techniques. DVT was suspected clinically after 95 of 738 (13%) arthroplasties, with no symptoms suggestive of DVT after the remaining 643 procedures. Surveillance duplex scans were performed within 2 weeks of 371 procedures, while no surveillance studies were performed after the remaining 272 procedures. In these asymptomatic patients only 2 (0.5%) surveillance duplex studies were positive for DVT. In contrast, 4 of 37 (11%) duplex exams and 5 of 62 (8%) contrast phlebograms performed among symptomatic patients were positive for DVT. The overall incidence of DVT after arthroplasty in the entire population was 1.4% (10/738) with no pulmonary emboli. Patient follow-up averaged 162 +/- 285 days. Using the 1995 Medicare reimbursement of $163 for venous duplex, the incremental cost was $35,000 to detect 1 additional unsuspected DVT and $110,000 per additional quality-adjusted life-year gained. The low incidence of clinically significant DVT and pulmonary emboli with current prophylaxis does not justify an aggressive screening program. Decision analysis suggests that a greater incidence of DVT is required for screening to be worthwhile.