Local versus central assessment of venographies in a multicenter trial on the prevention of deep vein thrombosis in neurosurgery

Venography is the diagnostic method of choice for end-point measurement in multicenter trials on the prevention of postoperative deep vein thrombosis (DVT). The aim of the study was to determine the inter-observer agreement b etween the local and central assessment of venographies in a multicenter tr ial comparing enoxaparin and placebo in the prevention of DVT after electiv e neurosurgery. The study was run in seven centers experienced in venography trials on DVT prevention. The central and local adjudication panels were both blind with respect to the assigned treatment. The central panel was unaware of the loc al adjudication. Venographies were adjudicated as positive, negative or ina dequate for adjudication and positive venographies as proximal or distal DV T. Inter-observer agreement was assess ed according to the Cohen's inter-ob server variability index (K index). All 266 venographies (8 monolateral) were considered adequate for adjudicat ion by both the central and local panels. A disagreement was found in 25 ea ses; K index = 0.75. Fourteen venographies adjudicated as negative centrall y were considered positive locally (3 were proximal DVT). Eleven venographi es adjudicated as positive centrally (1 was a proximal DVT) were considered negative locally. Enoxaparin was found to be more effective than placebo a ccording to both the central and local adjudication: 16.9% versus 32.6% (Re lative risk, RR = 0.52; CI95%: 0.33-0.82) according to central adjudication ; 18.5% versus 33.3% (RR = 0.56; CI95% 0.36-0.87) according to local adjudi cation. We conclude that a good inter-observer agreement in the assessment of venog raphy was observed between the central and local adjudication in a study un DVT prevention run in a restricted experienced study framework. The cost a nd work overloading of central assessment of venographies in this study fra mework seems not to be justified.