Unexpected Toxicity with Intensified Induction in Infant Acute Lymphoid Leukemia.

Infants with Acute Lymphoid Leukemia (ALL) have a poor event free survival (EFS) characterized by early relapse typically occurring less than 6–9 months after diagnosis. More than 70% of these infants have evidence of the MLL gene rearrangement in their leukemia blasts and for that high risk group the published EFS ranges from 20–30%. To improve this overall dismal outcome, infants less than 1 year old with ALL on Children’s Oncology Group (COG) 9407 received shortened, intensified therapy including early Induction Intensification (I.I.). Initially, all drugs, except vincristine and intrathecal therapy, were dosed per m2 without traditional infant dose reductions. Infants were treated in 3 cohorts: 1997–2000, Cohort 1 (n=16) and Cohort 2 (n=55); 2001–2006, Cohort 3 (n=142). In Cohort 1 + 2, the three week induction included dexamethasone × 21 days, vincristine weekly × 3, L-asparaginase twice weekly × 6 doses, daunorubicin by continuous infusion (CI) days 1+2, and cyclophosphamide days 3+4. An early I.I. immediately followed induction and included high dose methotrexate (4 gm/m2 weekly × 2) followed by a five day pulse of daily cyclophosphamide with etoposide. Due to the excessive skin breakdown, myelosuppression and induction / I.I. deaths observed in Cohort 1 (notably in infants less than 90 days of age), the daunorubicin dosing was initially changed from mg/m2 CI over 48 hours to mg/kg/age over 48 hours by CI (2 mg/kg/day for age < 6 months; 2.5 mg/kg/day for age 6–9 months; 3 mg/kg/day for age > 9 months). With no reduction in death rate, daunorubicin was then changed to IV push daily × two from the previous 48-hour CI. Toxicity remained unacceptable and in Cohort 3 prednisone replaced dexamethasone as the induction steroid with the remainder of induction and I.I. unchanged. The toxic deaths during induction / I.I. by cohort were: Cohort 1– 3/16 (19%); Cohort 2– 13/55 (24%); Cohort 3– 8/142 (6%). Infection was the most common cause of death: Cohort 1+2 (16 induction / I.I. deaths)- bacterial-8; fungal-2; viral-2; Pneumocystis-1; non-infectious pulmonary disease-3. Cohort 3 (8 induction / I.I. deaths)- bacterial 2; fungal-1; viral-2; tumor lysis/renal failure-2; veno-occlusive disease-1. Infants < 90 days of age were at greatest risk for death. In Cohort 1 there were 3 deaths in the 3 infants < 90 days and in Coh 2 there were 7 deaths in 11 infants < 90 days. The effect of age was much less pronounced in Cohort 3 with only 4 deaths in 27 infants < 90 days of age. Modifications of daunorubicin dose and schedule had no impact on mortality (Cohort 1 vs. 2), but the substitution of prednisone for dexamethasone significantly reduced the number of deaths in Cohort 3 (p=.0002). Age less than 90 days was an important risk factor for induction / I.I. death in Cohort 1+2 vs. Cohort 3 (p=.006). Based on these observations, prednisone is the preferred steroid for induction in infants also receiving anthracycline.