In a previous study, we evaluated efficacy of repeated antilymphocyte globulin (ALG) treatment for patients with severe aplastic anemia not responding to an initial ALG treatment or relapsing after initial response to ALG. We now searched in the same cohort of patients for differences between patients who responded to treatment and remained free of complications and those who relapsed or developed a clonal complication. From 107 patients surviving for more than 1 year after immunosuppression, 34 remained free from complications after the first course of ALG, and 73 presented an event defined as relapse of aplastic anemia, development of a clonal complication such as paroxysmal nocturnal hemoglobinuria, myelodysplastic syndrome or leukemia, or appearance of a solid tumor. We compared these two groups for survival, clinical performance and blood counts during follow-up. Survival probability was 93% for the event-free patients, and 55% for the patients with a complication event (p = 0.0003). Event-free patients had a higher incidence of complete remission (71%), were more often free of immunosuppressive treatment (79%) and independent of transfusions (100%), and had a higher Karnofsky score (91% with a score ≥90%) as compared to the group with events (29, 37, 67, 48%; p ≤ 0.0002). At 1 and 3 years, event-free patients had significantly higher leukocyte and neutrophil counts, as compared to patients with a complication (p < 0.05). However, at 3 and 5 years, event-free patients had borderline higher platelet counts (p = 0.056, p = 0.078) and hemoglobin (p = 0.097, p = 0.061). The coefficient of variation as an expression of the variability of the results in each group was systematically lower at 3, 5 and 10 years in the group of event-free patients. Despite some differences between the two groups, our data support the hypothesis that patients with long-lasting remissions should not be considered as definitively cured of aplastic anemia.