Background: Osteonecrosis (ON) has been reported as an infrequent but sometimes debilitating late effect of HL therapy. The goal of this study is to provide a detailed analysis of ON as a late effect after treatment for HL. Methods: 12,083 patients treated within first-line GHSG trials between 05/98 and 07/14 were evaluated. ON cases were identified and information on patient characteristics, localizations, interventions and outcomes was collected. Risk factors for ON were analyzed by multivariate logistic regression. Results: 66 patients had a total of 140 ON. The cumulative incidence of ON was 0.16% [CI: 0.08-0.30] in early (favorable or unfavorable) stage HL (esHL) (n=10) and 0.93% [CI: 0.71-1.21] in advanced stage HL (asHL) (n=54). 75% of ON cases were male. 71% of patients had more than one area affected and in 73% ON of the femoral head was present. 54% of patients needed surgical intervention and 66% had lasting symptoms despite treatment. Peak incidence (41%) was observed in the second year after diagnosis and 83% of cases occurred within 3 years after HL diagnosis. In a multivariate logistic regression comprising all patients, male gender (OR 2.1; CI: 1.2-3.7) and asHL (OR 3.9; CI: 2.3-6.9) were identified as risk factors for ON. Because ON was a rare complication in esHL, the following results focused on survivors of asHL. The median cumulative prednisone dose in ON cases after asHL was 8,400mg (range: 3,920-10,800) versus 7,350mg (range: 0-16,800) in not affected patients. In a multivariate logistic regression model including only asHL patients, young age (OR 0.7 for each additional 10 years; CI: 0.5-0.9) and a higher cumulative dose of prednisone during therapy (OR 1.3 for each additional 1gr; CI: 1.1-1.5) were identified as additional risk factors. Nodal pain after alcohol ingestion (p=0.78), radiotherapy (p=0.29), a large mediastinal tumor (p=0.13), international prognostic score (IPS) (p=0.09) and body-mass-index (BMI) (p=0.99) did not influence ON risk. Conclusion: We provide the most comprehensive analysis of ON after HL therapy performed so far. ON after HL therapy is infrequent but often leads to significant disease burden in affected patients. The described risk factors and peak incidence timeframe could help to identify patients at high risk for ON. Early evaluation of patients with suggestive symptoms is recommended. Decreasing the cumulative corticosteroid dose in future regimens might reduce the incidence of ON after HL therapy.