The natural history of paroxysmal nocturnal hemoglobinuria (PNH) clinical subcategories (classic PNH and aplastic anemia [AA]/PNH syndrome) is still unknown. We retrospectively studied 460 PNH patients diagnosed in 58 French hematologic centers from 1950 to 2005. The median (SE) follow-up time was 6.8 (0.5) years. The median survival time (SE) was 22 (2.5) years. We identified 113 patients with classic PNH, 224 patients with AA-PNH syndrome, and 93 (22%) intermediate patients who did not fit within these 2 categories. At presentation, classic PNH patients were older, with more frequent abdominal pain and displayed higher levels of GPI-AP-deficient granulocytes. A time-dependent improved survival was observed. In classic PNH, diagnoses before 1986 (hazard ratio [HR]: 3.6; P = .01) and increasing age (P < .001) were associated with worse survival prognoses, whereas use of androgens within the first year after diagnosis was protective (HR, 0.17; P = .01). Transfusions before 1996 (HR, 2.7; P = .007) led to lower survival rates in patients with AA-PNH syndrome, whereas immunosuppressive treatment was associated with better outcomes (HR, 0.33; P = .03). Evolution to thrombosis affected survival in both subcategories (classic PNH: HR, 7.8 [P < .001]; AA-PNH syndrome: HR, 33.0 [P < .001]). Evolution to bicytopenia or pancytopenia for classic PNH (HR, 7.3, P < .001) and malignancies for AA-PNH syndrome (HR, 48.8; P < .001) were associated with worse outcomes. Although clinical presentation and prognosis factors are different, classic PNH and AA-PNH syndrome present roughly similar outcomes, affected mainly by complications.