Purpose Our focus was on patients with pediatric acute lymphoblastic leukemia (ALL) who experienced relapse or died without becoming transplantation candidates. The purpose was to outline measures needed to improve the outcome. Patients and Methods We analyzed our population-based 20-year data on 3,385 Nordic children with ALL treated on Nordic Society for Pediatric Hematology and Oncology ALL protocols, and described the flow of these patients through relapses, remissions, and deaths as a result of toxicity, demonstrating where major patient losses occurred. Results In total, 854 patients (25%) had a first and 274 patients (8%) had a second ALL relapse. P for survival after the first relapse was .35 +/- .02. The induction mortality (2.2%, primary; 10.3%, first relapse; 26.3%, second relapse) and remission mortality (1%, first complete remission [1CR]; 19%, second CR [2CR]) were significant; transplantation-related mortality (TRM) only represented 15% (69 of 459) of the deaths as a result of toxicity. Of the 766 patients entering 2CR, 29% underwent transplantation (P for survival,.46 +/- .04), whereas 71% continued receiving chemotherapy (P for survival, .39 +/- .02). Children with stem-cell transplantation indications in 2CR, if they did not undergo transplantation, generally died or had a second relapse. The patient groups that underwent transplantation in 1CR (n = 84), 2CR (n = 220), and >= 3CR (n = 62) represented different risk profiles. Those with allogeneic stem-cell transplantation (allo-SCT) in >= 3CR (P for survival, .37 +/- .07) had an ALL and first relapse with favorable features. Conclusion Major patient losses occurred through mortality as a result of toxicity and resistant disease during the pathways before allo-SCT. After relapse, more patients were lost to mortality as a result of toxicity during conventional chemotherapy compared with TRM. After second relapse, the chance for rescue by allo-SCT in >= 3CR was minimal. The question of whether transplantation is recommended after ALL relapse should be carefully addressed, and more efficient relapse protocols should be launched.