Lack of consistency in the relationship between asymptomatic DVT detected by venography and symptomatic VTE in thromboprophylaxis trials

Asymptomatic deep-vein thrombosis (DVT) detected by mandatory venography, a surrogate outcome, comprises most of the efficacy outcome events in recent thromboprophylaxis trials. The validity of this surrogate to estimate trade-off between thrombotic and bleeding events in these clinical trials requires a consistent relationship between asymptomatic DVT and symptomatic venous thromboembolism (VTE). In this systematic review of high quality VTE prevention trials, we examined the consistency of the ratios of asymptomatic DVT to symptomatic VTE across a broad range of indications. Studies were identified from citations listed in the chapters on VTE prevention in the antithrombotic guidelines by the American College of Chest Physicians, 2012. A study was eligible if it: 1) was a randomised trial comparing an anticoagulant with standard of care; 2) included at least 500 participants; 3) reported asymptomatic or all DVT rates; and 4) reported symptomatic VTE rates. Of the 26 eligible trials, 19 trials were conducted in orthopaedic patients, five in general surgery patients and two in general medical patients. The overall median rates (ranges) for asymptomatic DVT and symptomatic VTE were 9.11 % (0.75 to 54.87 %) and 0.49 % (0.00 to 3.10 %), respectively. The median ratio was 14.53, with a wide range (2.75 to 103.86). Wide variability in the ratios persisted despite indication- and anticoagulant-specific analyses. In VTE prevention trials of alternative anticoagulants, the wide variability in the ratios of asymptomatic DVT to symptomatic VTE precludes judging the trade-off between thrombotic and bleeding events on the basis of composite outcomes dominated by venographic DVT.