To the editor—Horwitz and colleagues are to be congratulated on the successful engraftment of mesenchymal cells from donor marrow1. However, the clinical changes they reported require careful evaluation, particularly when considering the potentially hazardous nature of the procedure and the availability of an effective alternative treatment such as bisphosphonates.

Bone biopsy in infants with osteogenesis imperfecta (OI) is technically difficult. The fragmentation of the pre-treatment biopsies may have prevented accurate assessment of osteoblast numbers, as the cells were identified by site and morphology rather than by osteoblast-specific stains.

Bone mineral content increased rapidly. Although weight is a factor that accounts for the most variation in bone mineral content in regression analyses, weight subsumes the contemporaneous effect of increasing length in normal children in such analyses. Bone densitometry in childhood is confounded by the nature of the measurement, such that larger children of the same age will have apparently increased bone mineral content and density. Thus, the increase in length of 5% in each child may have accounted for a proportion of the increase in bone mineral content.

The fracture rate in infancy is highest in severe OI between birth and six months of age, falling thereafter. The apparent decline may simply have been part of a continuing pattern, and a comparison with the fracture rate six months before transplant rather than from birth would give a clearer picture of the effect of the procedure on fracture rates.

Finally, I understood from Dr. Horwitz that more than three children were enrolled in this program. Would it be reasonable to ask whether the other recipients did as well as the three children described in the article?