Because survival of pediatric acute lymphoblastic leukaemia (ALL) has improved in the last years, adverse effects acquire increasing importance. Avascular osteonecrosis (ON) is a long term complication of ALL treatment affecting the patients' quality of life. Its pathogenesis recognizes multiple factors (age, race, genetics, chemotherapies and steroids). The different roles of dexamethasone (DXM) versus prednisone (PDN) is still under discussion, as dosages and rate of administration.The prevalence of ON varies in different series (0.43-17.6%). It affects mostly the weight-bearing epiphyseal bone, usually knee and hip. Multiple joint involvement is common. ON treatment is not established: efficacious drugs are not available; surgery is strongly invasive for pediatric patients (pts); among physical treatments, the use of hyperbaric oxygen therapy (HBO) is uncommon.We report our experience in children treated consecutively from 2000 to 2015 for ALL at the Pediatric Hematology Clinic of Padua, who developed ON. HBO was used in some pts with variable efficacy, suggesting its possible role in early stages.Materials and methods Prevalence ofON was assessed retrospectively in 469 pts with ALL enrolled in AIEOP 2000 Protocol (P 2000, 327 pts) and AIEOP BFM ALL 2009 Protocol (P 2009, 142 pts); median age was 5 years (yrs, range 1-17). Girls were 53% and boys 47% (Table 1).ON was defined as persistent pain in one or more major joints (hip, knee, elbow, shoulder, ankle) and peculiar radiological findings (plain XR or MRI). Severity was graded radiologically by the ARCO (Association of Research Circulation Osseous) staging system; clinically by the CTCAE v3.0.Statistical analysis was done using z- and t-tests.Results 19/469 pts (4%) developed ON of CTCAE grade 2-4. Age at leukaemia diagnosis was over 10 yrs in 17/19 pts. ON always presented during maintenance or off therapy. Median follow up after ON development in P 2000 was 54 months (range 21-170) and 28.5 months (8-48) in P 2009. The most frequent sites were knee, hip and ankle. Multiple joint involvement was frequent (11/19, 58%); 14/19 (74%) had also functional limitation at diagnosis (grade 3-4).Nine ON pts (3 in P 2000, 6 in P 2009) were at high risk, receiving higher steroid dosages than low risk (p<0.001). Four pts underwent TCSE in first remission (1 in P 2000 and 3 in P 2009).As for ON treatment, 14/19 (74%) pts were treated conservatively; 5/19 (26%) pts underwent joint replacement; 7/19 (37%) pts received HBO (25-40 sessions, 2.5 ATA).Four children showed complete recovery or only limited disturbances; 14 pts had more severe impairment (grade 3-4, 5 joint replacements). One patient developed osteosarcoma as secondary malignancy after first line chemotherapy, TCSE and HBO.HBO results: pain improvement was not obtained in any patient. The 5 pts with ON ARCO grade II-III showed lesion stability without joint collapse at control MRI. The 2 pts with very advanced ON (grade IV) placed implants.Comparing the two consecutive Protocols (P 2000 vs P 2009), we found different prevalence of ON (0,3 % vs 8,4% respectively, p <0.001), but no difference in gender nor steroid dosage. Notably, children aged over 10 yrs were more frequent in P 2000 than in P 2009 (p <0.001); however, older children had 3.1% of ON prevalence in P 2000 and 25% in P2009 (p 0.001, Table 2).Conclusion In our experience the prevalence of symptomatic ON in pts treated with two consecutive and similar ALL protocols was 4% (19/469) as reported by literature. No asymptomatic cases were reported, probably due to lack of screening. Boys were affected similarly than girls; age was significantly higher in patients with ON (p 0.001).A marked different prevalence between the two groups (0.3% vs 8.4%) was found, without apparent role of age, gender or steroid dosages. However, a possible underdiagnosis in P 2000 can be considered, due to higher attention given recently to this problem.High risk pts developed ON more frequently (p<0.001). The role of different steroids can not be inferred from our study, since both PDN and DXM were administered to all pts. As suggested by DFCI reports, the difference between the two Protocols could be due to the extensive use of L-Asparaginase, administered in its pegylate form in P 2009.HBO seems to stabilize the progression of early stage of ON, but it is not efficacious in advanced ON. Screening MRI could be useful for early diagnosis and to understand the utility of HBO in these cases. [Display omitted] [Display omitted] DisclosuresNo relevant conflicts of interest to declare.