Context: Craniopharyngiomas are often associated with an unfavorable prognosis, but data on their long-term consequences are sparse. Objective: The aim of the study was to identify markers of recurrence and factors associated with compromised social rehabilitation and altered quality of life in a large cohort of patients with either childhood-onset (CO) or adult-onset craniopharyngioma. Methods: Retrospective analysis was performed for 171 patients treated for craniopharyngioma in two academic centers in France between 1972 and 2009. For each subject, data were collected concerning clinical presentation, imaging features, visual sequelae, endocrine and metabolic impact, treatment modalities (surgery, radiotherapy), recurrence-free survival rate, and social insertion, as well as answers to the WHO-QOL BREF questionnaire. Results: A total of 65 CO and 106 adult-onset patients were reviewed. If CO was diagnosed before the age of 10 yr, this was associated with a higher incidence of obesity, blindness, and panhypopituitarism, and only 40.7% of subjects had adequate work or school attendance compared to 72.4% of patients with later disease onset. Initial symptoms of intracranial hypertension (SIHT), pterional surgery, and multiple surgery were associated with obesity and poorer social insertion. No determinant of quality of life was identified. In the subgroup of patients treated in the 1990s and later, the progression rate was 59.4% in patients with residual tumor on magnetic resonance imaging compared with a 19.8% recurrence rate in the group with apparently complete resection. Recurrence/progression correlates significantly with male gender, early onset (before 10 yr), and SIHT, but only SIHT at presentation remained a significant predictor with multivariate analysis. Conclusions: Craniopharyngioma continues to be associated with severe outcomes. Higher morbidity rates are found in patients with early-onset disease (before 10 yr), initial SIHT, or in whom pterional surgery was required. Markers of recurrence are difficult to identify, with SIHT being the most powerful predictor.