Osteonecrosis as a Complication of Treating Acute Lymphoblastic Leukemia in Children: A Report From the Children's Cancer Group

Osteonecrosis is a well-recognized complication of corticosteroid use. Morbidity is caused by progressive joint damage, articular collapse, and arthritis. Weight-bearing joints are commonly affected, and totaljoint replacement is often necessary to restore function. Corticosteroids are integral for the optimal management of childhood acute lymphoblastic leukemia. The authors in conjunction with pediatric oncologists from 111 pediatric oncology centers undertook an extensive retrospective study of a treatment protocol of high-risk leukemia patients (CCG 1882). Between May 1989 and June 1995, a total of 1541 patients were registered on this protocol; 1409 of these patients were included in this analysis. Eligibility of patient participation in this protocol included patients 1 to 9 years old with a white blood cell (WBC) count greater than 50,000/cm3, and children older than 10 years irrespective of the WBC count. Patients with lymphomatous features and Burkitts leukemia were excluded. A detailed osteonecrosis data questionnaire was obtained for each patient. This included symptoms, joint involvement, diagnostic evaluations, modifications of acute lymphoblastic leukemia (ALL) therapy, and orthopedic interventions. Additional information was obtained from the Children's Cancer Group data bank. In the entire study population, 111 of 1409 patients had osteonecrosis, with a 3-year life-table estimated incidence of 9.3%. There was only one occurrence of osteonecrosis (ON) 3 years after diagnosis of ALL. ON also varied by sex and age at diagnosis of ALL. There was a very low rate in the group less than 10 years old (4/516 patients) compared with the 10to 15-year-old group (85/736 patients) and the 16to 22-year-old group (22/157 patients). ON occurred significantly more often in females than males with a 3-year incidence of 12.2% and 7.7%, respectively. The relative incidence in the white population was five times greater than that in the black population.The diagnosis of ON resulted in the discontinuation of further corticosteroid therapy in 60 patients and dose modifications in six patients. Therapy was not altered in 40 patients still on treatment. Three had completed treatment at diagnosis of ON. Thirty-five patients (32%) were diagnosed with ON during year 1 of therapy, 60 (54%) were diagnosed with ON in year 2, 15 (13%) in year 3, and 1 patient was diagnosed in year 5. Based on the results of this investigation, current Children's Oncology Group ALL protocols de-