In addition, 42% of the patients in the study had GVHD involving the skin only, which again, is easier to treat compared with GVHD affecting the GI tract or internal organs. Osiris used a liberal definition of response to measure Prochymal's efficacy. In the study, a patient was measured for response on day 28 after treatment. Osiris says the complete response rate at that time point was 74%, double the historical rate. But the study Osiris cites to make that comparison, published in 2002, defined complete response as total clinical resolution of GVHD at day 28, maintained for another 28 days without requiring any additional treatment. There were nine patients in the Osiris study who required additional treatment after Prochymal to eliminate their GVHD, including four patients who Osiris deemed complete responders. These patients would not have counted as responders had Osiris used a stricter and more typical definition of efficacy. Mills defended the design and conduct of the Prochymal study. "We used an incredible efficacy standard for GVHD. The fact that we had 74% of patients with a complete response at day 28 is remarkable," he says. And he says the study design did not allow patients to be retreated with Prochymal, which explains why some patients required other, additional treatments even after they were deemed responders. "We have re-treatment data with Prochymal if GVHD returns, and it works very well, but not in this trial," he says.