In 1993 we initiated an International study in newly diagnosed adults with ALL (aged 15 to 55 years). 1508 eligible patients were entered before 31 st October 2003 and achieved a complete remission (CR) rate of 91%. 551 patients underwent transplant in 1 st CR: 321 matched family donor (MFD), 149 autograft, 67 matched unrelated donor (MUD) and 14 other/unknown. The auto transplant group include all transplants, both randomised and elected. Median time from the start of induction to transplant was 168 days (range 79–1726 days). There were few cases of graft failure; 3.7% MFD, 3% auto, 2% MUD (data available in 82% of cases). Overall survival at five years (5y OS) from transplant was 55% (95% confidence interval 49–61%) for MFD patients, 39% (30–47%) for autograft and 46% (33–58%) for MUD. Age related outcome was significant - for MFD 5y OS st CR transplants (3 MFD, 1 auto). Forty patients had MFD, 56 MUD and 13 autograft post relapse with 5y OS from relapse of 22% (8–36%), 19% (8–30%) and 19% (0–41%) respectively. Median time from relapse to transplant was 102 days (range 1–344 days). Ten patients with other forms of transplant (1 haplo-identical, 6 mismatch and 3 unknown) all died within one year of relapse. 158 relapsed patients had transplant prior to relapse (153 in 1 st CR and 5 pre-1 st CR). 5y OS in the 282 who never received a transplant (or were transplanted only after second relapse) was 2% (0–4%). If the 112 patients who died or relapsed again within 102 days of first relapse are excluded 5y OS from relapse was 5% (0–9%), but a comparison of transplant versus chemotherapy would be biased by patient selection factors. Conclusions Transplant related mortality remains significant in the allogeneic setting and is age related. Auto transplant is a safe procedure and remains under serious consideration for 1 st CR patients without a MFD. There may be some survival advantage for patients well enough to receive a transplant post-relapse, but this is only likely to benefit a tiny minority of patients. Data from this large ongoing trial of over 1500 patients indicates that it is more important to prevent initial relapse rather than to rely on salvage treatment after relapse.