We thank Drs. Casner and Hilty for their thoughtful letters. Dr. Casner inquires about the value of functional assessment before surgery and the conflicting recommendations of 2 national guidelines. The ACP guideline was published in 1997. At that time, no studies correlated exercise capacity with perioperative outcomes. In 1999, Reilly and colleagues (1) tested the hypothesis that self-reported exercise capacity would predict postoperative complications. The authors defined good exercise capacity as the self-reported ability to walk 4 blocks and climb 2 flights of stairs. Among 600 consecutive patients undergoing major noncardiac surgery, cardiovascular complications (relative risk, 0.54; P = 0.04) and total serious complications (relative risk, 0.51; P = 0.001) were both significantly less common in patients with good exercise tolerance. There was a nonsignificant trend toward fewer pulmonary complications (relative risk, 0.70; P > 0.2).

In our recent Update, we cited a study by Girish and colleagues (2), which demonstrated that directly observed stair climbing was the strongest predictor of major cardiopulmonary complications after high-risk surgery and outperformed clinical variables (2). However, this test had modest sensitivity and specificity (71% and 77%, respectively) when good exercise capacity was defined as the ability to climb 4 flights of stairs. We believe that the literature now supports the use of functional capacity (either self-reported or directly observed) as an important component of preoperative risk stratification. It complements, but does not replace, existing cardiovascular risk indexes.

Regarding the question of extended-duration thromboprophylaxis posed by Dr. Hilty: while short-duration warfarin and enoxaparin are equally effective by 3 months after hospital discharge (3), few studies have evaluated the efficacy of extended-duration oral anticoagulants. Prandoni and colleagues (4) studied 360 patients undergoing total hip arthroplasty and demonstrated superiority of extended-duration warfarin therapy (4 weeks after hospital discharge) compared with short-term prophylaxis (4). Venous thromboembolism rates were 5.1% and 0.5%, respectively (absolute difference, 4.57 percentage points [95% CI, 1.15 to 7.99 percentage points]). Samama and colleagues (5) compared extended prophylaxis using fixed-dose reviparin (a low-molecular-weight heparin) with adjusted-dose acenocoumarol in 1279 patients undergoing total hip replacement (5). The failure rate (the combination of symptomatic thromboembolism, major hemorrhage, or death) was 3.7% with low-molecular-weight heparin prophylaxis and 8.3% with oral anticoagulants (P = 0.001). Most of this difference was due to a higher bleeding rate among acenocoumarol-treated patients. On the basis of these limited data, extended-duration prophylaxis with warfarin may be inferior to low-molecular-weight heparin. Individual institutions must create policies on the cost-effectiveness of extended prophylaxis by examining actual local costs associated with medications and excess hospitalizations due to bleeding complications.