Forum Neural transplantation in putaminal 18F-dopa uptake have also been highly variable, which suggests that differences in survival and growth of the grafted dopaminergic neurons may be an important factor in the recovery of patients. A systematic patient-by-patient analysis in the larger double-blind trials could help to elucidate the factors underlying this variability. Another challenge is that the reported clinical improvements after transplantation may be due to investigator bias, placebo effects, or both. In this regard, the results of the two NIH-sponsored double-blind studies are reassuring in that none of the patients who underwent sham surgery showed any significant, long-term benefit of the placebo treatment. However, these studies have raised two major concerns: only modest efficacy was observed in the transplantation groups, and a significant number of patients developed off-state dyskinesias after transplantation.

It is clear from the disappointing outcome of these two trials that we still do not fully understand how to use fetal-cell transplantation to achieve consistent, optimum results. They also highlight several serious issues that must be addressed. First, the mechanism underlying the off-state dyskinesias developed by some patients needs to be investigated. Are some patients particularly sensitive5 and how can this side-effect be avoided? Second, how do we select the patients who are likely to derive the greatest clinical benefit? Are patients with advanced disease unsuitable candidates for fetal-cell transplantation and is age a negative factor? Third, where is the best site for graft placement? Does the transplantation surgery need to be tailored to each patient’s disease profile? Fourth, do we need to standardise preparation of cells for transplantation? Is it feasible to store tissue for days or weeks in vitro—as done in the two NIH trials—without losing functional efficacy? Fifth, what about immunosuppression? Can we really exclude or reduce immunosupressive treatment when multiple allogeneic donors are used and surgery is done in stages? Finally, we need to consider the risk–benefit analysis. What is the balance between clinical benefit and side-effects, and how does it compare with the other treatment options available? Now that the NIH trials have been completed and the results reported, we are entering a new phase in the development of cell-replacement therapy for PD. In our opinon, transplantation of fetal dopaminergic neurons is based on a solid rationale and open-label trials have provided convincing proof of principle that transplants can work—and work well—in some patients. However, in the light of the NIH trials, it is clear that we need carefully to reassess critical aspects of the transplantation procedure and further analyse the issues raised in well-designed open-label studies. Further carefully planned open-label trials are required to answer various major outstanding issues—such as those relating to tissue collection and preparation, patient selection, and surgical placement and