Based on this lukewarm recommendation, Canadian regulators approved Prochymal on May 17 with conditions that restricts use of the stem cell therapy and requires Osiris to provide further clinical data to prove its clinical benefit. This is not the ringing endorsement of Prochymal that Osiris has portrayed since the Canadian approval was announced last month. A read of the Prochymal label approved in Canada makes that point clear. "Protocol 280" was a double-blind, placebo-controlled phase III study that enrolled 240 patients with steroid-refractory GVHD. In 2009, Osiris announced that the "protocol 280" study failed, as did a companion phase III study in newly diagnosed GVHD patients. Prochymal provided little or no benefit over placebo. Osiris, however, sifted through the negative phase III study to find a subgroup of 28 children with all grades of GVHD --14 treated with Prochymal and 14 with a placebo. In this retrospectively defined subgroup, the Prochymal response rate at day 28 was 64% versus 36% for the placebo-treated patients -- a trend favoring Prochymal but not statistically significant. The survival data from this data-mined subgroup of children is equally weak. At day 100, 11 Prochymal patients (78.6%) were alive compared to seven placebo-treated patients (50%). The result was not statistically significant. At day 180, the survival "benefit" narrows, with nine Prochymal-treated kids alive versus seven placebo-treated kids. Health Canada also considered data from "Protocol 275," a single-arm (uncontrolled) study of 75 pediatric patients with Grades B-D acute GVHD who failed to respond to other treatments. GVHD severity is graded A-D, with A least severe, D most severe. At day 28, the patient response to Prochymal was 61%, but unlike in protocol 180, there was no control arm to which the Prochymal response could be compared. In order to examine survival in protocol 275, Osiris compared Prochymal to a historical control of "similar" GVHD patients. The probability of survival at 180 days post treatment was 51% for Grade D GVHD patients treated with Prochymal compared to 31% for the historical control group -- statistically significant. But the same survival analysis conducted for patients with milder Grades B&C GVHD failed to demonstrate favoring Prochymal. In fact, survival trended in favor of the historical control for Grade C GVHD patients.