Approach to Outcome Measurement in the Prevention of Thrombosis in Surgical and Medical Patients

Abstract

CHEST Supplement ANTITHROMBOTIC THERAPY AND PREVENTION OF THROMBOSIS, 9TH ED: ACCP GUIDELINES Approach to Outcome Measurement in the Prevention of Thrombosis in Surgical and Medical Patients Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines Gordon H. Guyatt , MD , FCCP ; John W. Eikelboom , MBBS ; Michael K. Gould , MD , FCCP ; David A. Garcia , MD ; Mark Crowther , MD ; M. Hassan Murad , MD , MPH ; Susan R. Kahn , MD ; Yngve Falck-Ytter , MD ; Charles W. Francis , MD ; Maarten G. Lansberg , MD , PhD ; Elie A. Akl , MD , MPH , PhD ; and Jack Hirsh , MD , FCCP This article provides the rationale for the approach to making recommendations primarily used in four articles of the Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines: orthopedic surgery, nonorthopedic surgery, nonsurgical patients, and stroke. Some of the early clinical trials of anti- thrombotic prophylaxis with a placebo or no treatment group used symptomatic VTE and fatal PE to measure effi cacy of the treatment. These trials suggest a benefi t of thromboprophylaxis in reducing fatal PE. In contrast, most of the recent clinical trials comparing the effi cacy of alterna- tive anticoagulants used a surrogate outcome, asymptomatic DVT detected at mandatory venog- raphy. This outcome is fundamentally unsatisfactory because it does not allow a trade-off with serious bleeding; that trade-off requires knowledge of the number of symptomatic events that thromboprophylaxis prevents. In this article, we review the merits and limitations of four approaches to estimating reduction in symptomatic thrombosis: (1) direct measurement of symp- tomatic thrombosis, (2) use of asymptomatic events for relative risks and symptomatic events from randomized controlled trials for baseline risk, (3) use of baseline risk estimates from studies that did not perform surveillance and relative effect from asymptomatic events in randomized controlled trials, and (4) use of available data to estimate the proportion of asymptomatic events that will become symptomatic. All approaches have their limitations. The optimal choice of approach depends on the nature of the evidence available. CHEST 2012; 141(2)(Suppl):e185S–e194S