On May 17, Health Canada approved its first stem cell therapy, Prochymal, produced by Columbia, Maryland–based Osiris Therapeutics. Only a few weeks later, Medsafe, the medical regulatory agency in New Zealand, approved it, too. In both countries, the drug, a preparation of bone marrow–derived mesenchymal stem cells, will be used to treat graft-versus-host disease (GVHD), in children under 18. The approval is surprising given that in 2009, Prochymal missed two phase 3 trial endpoints in severe refractory GVHD despite startling phase 2 results, prompting the company to abandon the indication in the United States (Nat. Biotechnol. 27, 966–967, 2009). The Canadian go-ahead was based on a subset of data, in which Osiris claimed a clinically meaningful response in some 64% of children. GVHD, a life-threatening immune system reaction, affects more than 40% of children who receive bone marrow transplants. But Health Canada itself says the efficacy data are not conclusive and, as a condition of its approval, Osiris must carry out an international trial within five years. Edwin Horwitz, a pediatric oncologist at the Children's Hospital of Philadelphia, found the approval of a subset “mind boggling,” especially because Osiris has not published its data. “The ultimate success depends on doctors' willingness to use it,” he says. “Data would lend confidence.” Willem Fibbe, who also researches the clinical use of mesenchymal stem cells, says it is not clear whether the study was blinded, whether the post hoc analysis resulted in a biased interpretation of data, or whether the study was sufficiently powered. “Altogether, I have some pretty serious methodological concerns related to design and interpretation of the data,” says Fibbe. Lee Buckler, managing director of the Bellingham, Washington–based consulting firm, Cell Therapy Group, admitted that publishing the data would be useful, but he says the Canadian government's willingness to look—where the US Food & Drug Administration does not—at retrospective data, is valid. “They have a different risk-benefit analysis of the drug,” says Buckler.