Purpose: Osteonecrosis (ON) is a serious complication of the treatment of childhood acute lymphoblastic leukemia (ALL). The incidence of ON varies (range, 2.0-18.0%) according to treatment strategy. Previous studies reported older age, female sex, dosage of corticosteroid, use of dexamethasone (DEX), combination therapy with DEX and L-asparaginase (L-Asp), and ethnicity were putative risk factors. However, most of these studies were mainly conducted in U.S.A. or European countries. Thus, the data relating to ON in Asian ALL patients are scarce. Herein, we performed a retrospective analysis of Japanese ALL patient cohorts uniformly treated with consecutive two clinical trials (ALL-97 and ALL-02) to clarify the incidence, clinical characteristics and risk factors of ON.

Methods: In the ALL-97 study, various data on adverse events, including ON, were collected prospectively; however, in the ALL-02 study, there was no prospective collection of ON data. Thus, a questionnaire targeting ON was administered to all JACLS investigators in 2016. A total 1662 patients (635 patients treated with ALL-97, 1027 with ALL-02 protocol) were analyzed in this study. The total numbers of patients enrolled in the ALL-97 and ALL-02 studies were 635 and 1192 patients, respectively, and the majority of patients enrolled in both studies (100% in ALL-97 and 86.2% in ALL-02) were included in these analyses. The numbers of patients aged >10 years were 167/635 (26.3%) and 209/1027 (20.4%) in the ALL-97 and ALL-02 study, respectively. Clinical diagnosis of ON was defined as a symptomatic patient and positive findings on at least one imaging study, including plain bone X-ray, CT, or magnetic resonance imaging (MRI). Patients with ON diagnosed after hematopoietic stem cell transplantation were excluded. Patients with ON identified thorough these investigations were further examined for time of onset, sites involved, severity of symptoms, treatment, and outcome by questionnaire.

Results: In total, 24 of 1662 patients suffered from ON, of which 12/635 (1.9%) and 12/1027 (1.2%) patients were treated with ALL-97 and ALL-02 protocols, respectively. Of these 24 patients, the median age at diagnosis of ALL was 13.0 ± 1.4 and 12.5 ± 3.0 in ALL-97 and ALL-02 study, respectively. Of patients with ON, only one patient was <10 years old, with the remaining 23 patients aged >10 years at onset of ALL. Most patients were diagnosed with ON by MRI, and the femoral head was the most frequent site of ON (10/12 patients in ALL-97; 7/12 patients in ALL-02). In 13 patients, ON became apparent during chemotherapy (seven patients in ALL-97; six patients in ALL-02), while nine patients (five patients in ALL-97; four patients in ALL-02) developed ON after completion of chemotherapy. On multivariate analysis, only age >10 years was a significant factor associated with ON development (hazard ratio = 27.8, 95% = CI 5.99-129.0, p < 0.001). Separate evaluation of patients >10 years indicated incidence rates of ON of 12/167 (7.2%) and 11/209 (5.3%) in the ALL-97 and ALL-02 protocols, respectively, which was lower than previous reports, despite total administration of dexamethasone (DEX) similar to that in comparable studies; however, concomitant administration of DEX and L-Asp was reduced in JACLS protocols. In JACLS ALL-97 and ALL-02, combined use of DEX and L-Asp was performed in induction therapy. DEX was administered on days 8 to 14 and L-Asp three times per week after day 15 of induction therapy in the JACLS protocol; thus we did not administer DEX and L-Asp simultaneously in any treatment period. By contrast, all previous studies administered DEX and L-Asp/Peg-Asp simultaneously during several treatment phases. Hence, only JACLS studies lacked simultaneous administration of DEX and L-Asp. In addition, the cumulative dosage of DEX in the treatment phases including both DEX and L-Asp was lower in the JACLS ALL-97 and -02 protocols than in those in other studies.

Conclusions: We identified a low frequency of ON in the JACLS ALL-97 and -02 studies. Although the sole risk factor for ON was age (>10 years), even among high-risk patients, ON incidence was significantly lower than that reported in previous studies. These results suggest that, not only the total amount of DEX, but also how DEX and L-Asp are administered, which affects the clearance of DEX, may be associated with ON incidence in ALL patients.


Hiramatsu: Novartis Pharmaceuticals Corporation: Other: Clinical trials .

Author notes


Asterisk with author names denotes non-ASH members.