Allogeneic haematopoietic cell transplantation (HCT) is the only therapeutic modality capable of correcting the haematologic manifestations of Fanconi anaemia (FA). However, HCT from alternative donors has been associated with poor survival. Between June 1993 and July 1998, 29 FA patients (median age 12.1 years; range 3.7-48.5 years) were enrolled in a prospective phase I-II dose escalation study. All patients were treated with cyclophosphamide 40 mg/kg, total body irradiation (TBI) 450 cGy or 600 cGy and antithymocyte globulin (ATG), followed by HCT from an alternative donor. Graft-versus-host disease (GVHD) prophylaxis consisted of cyclosporin A for 6 months, short course methylprednisolone (2 mg/kg/day) between days +5 and +19 and marrow T-cell depletion by counterflow elutriation. The probability of developing grade III-IV toxicity was 17% (95% CI 3-31%). For the 25 marrow recipients, the probability of neutrophil engraftment (ANC 0.5 x 109/l by day 45) was 63% (95% CI 42-82%). Probabilities of grade II-IV acute GVHD and chronic GVHD were 32% (95%CI 10-54%) and 0% respectively. With a median follow-up of 18 months, the probability of survival for the entire cohort at 1 year was 34% (95% CI 17-51%). The presence of lymphocyte somatic mosaicism [i.e. the presence of diepoxybutane (DEB)-insensitive cells] was associated with a significantly increased risk of graft failure. Disappointingly, the use of higher dose TBI and post-transplant ATG did not improve engraftment. More effective peritransplant immunosuppression, especially in FA patients with somatic mosaicism, was required to overcome the barrier of graft rejection. New conditioning regimens adapted to each individual's alkylator sensitivity are needed to improve the outcome of alternative donor HCT for FA.