The authors reply:

We do not question the validity of the measurements presented by Butenas and Mann, particularly because it had been previously shown that tissue factor antibodies did not prolong blood clotting times1. Nor do we dispute their statement that alternatively spliced tissue factor, which does circulate2, has minimal activity, although the tested preparations were derived from Escherichia coli, perhaps thereby underestimating its activity. Moreover, this protein is found in all thrombi, thus supporting its further study.

Whole-blood clotting times are performed using nonflow conditions and therefore do not address the participation of blood-borne tissue factor in thrombogenesis, for which there is considerable evidence, apparently ignored by Butenas and Mann. Giesen et al.3 showed that ex vivo deposition of fibrin-containing thrombi on a collagen surface was virtually eliminated by a monoclonal antibody to tissue factor. Inasmuch as these experiments used no tissues other than blood, one must conclude that blood contributed the tissue factor. An in vivo model developed by Himber et al.4, in which thrombus growth within rabbit jugular veins or within a silastic jugular-jugular shunt was evaluated by radiolabeled fibrin accretion, rose linearly with time until a monoclonal antibody to rabbit tissue factor was perfused, after which there was essentially no thrombus growth. Further, a mouse transgenic model supported the role of circulating tissue factor in thrombogenesis5.

A fundamental difference between the cited experiments and the current data of Butenas and Mann is that the latter measured in vitro blood coagulation under nonflow conditions whereas the former studied thrombus formation in flowing blood. It is abundantly clear that shear forces caused by laminar flow are involved in thrombogenesis6. A potentially critical observation emphasizing the role of shear force is that fresh platelets do not stain for tissue factor; however, following perfusion with tissue factor–positive cells under physiologically relevant flow conditions, platelet aggregates were markedly positive for tissue factor7.

In summary, we do not dispute the results presented by Butenas and Mann; we do, however, seriously question the applicability of data derived from static assays to thrombogenesis, as well as the underlying rationale for this extrapolation.