Glucocorticoid selection for pediatric ALL

In this issue of Blood, Möricke et al report findings from the Associazione Italiana di Ematologia e Oncologia Pediatrica–Berlin-Frankfurt-Münster (AIEOP-BFM) acute lymphoblastic leukemia (ALL) 2000 clinical trial that substituting dexamethasone for prednisone during induction chemotherapy for children and adolescents with ALL improved event-free survival (EFS), but had no impact on overall survival (OS).¹ 

More than 50 years ago, treatment with prednisone alone for 4 weeks induced complete remission in almost 60% of children with ALL, yet a cure was largely unattainable.²  Today, almost 90% of children with ALL are cured using complex chemotherapy regimens delivered in sequential phases over 2 to 3 years.³  Glucocorticoids remain a cornerstone of ALL induction therapy. Prednisone and dexamethasone are both glucocorticoids, but have important differences in metabolism, central nervous system penetration, and toxicity. Dexamethasone generally is found to have more antileukemic efficacy at doses that have equal anti-inflammatory effects, but has also been associated with more toxicity, including fatal infections during induction therapy.^4-7  These 2 glucocorticoids have been compared directly in several randomized clinical trials, but there is no consensus regarding which one is optimal during induction therapy, at which dose, and for how long.⁸ 

The AIEOP-BFM ALL 2000 clinical trial was designed to determine if induction chemotherapy that included 3 agents plus dexamethasone improved EFS over the baseline expected with the same chemotherapy with prednisone. Over 3500 children and adolescents with ALL received a 7-day prednisone prophase, and then were randomized to receive either 21 days of dexamethasone or prednisone for the remainder of the first month of therapy.¹ 

The results are impressive and add important new data for ALL clinical researchers to consider. Substituting dexamethasone for prednisone, one glucocorticoid for another, for 3 weeks in the context of 2 years of multiagent chemotherapy reduced the relapse risk by one-third (see figure). This is the largest reduction in relapse risk seen in BFM trials in the past 40 years. There was a significant increase in early infectious deaths in the dexamethasone arm, but there was still a statistically significant improvement in EFS (see figure), the primary end point of the trial. So, on balance, dexamethasone is the drug of choice and the investigators must have popped the cork on a bottle of champagne, right?

Not so fast. The devil is in the details, and the results of this study have complex implications for how we think about ALL therapy in the 21st century. Although dexamethasone improved EFS, there was absolutely no difference in OS, as shown by the identical survival curves (see figure). With a median follow-up of 8.8 years, these results won’t change with additional follow-up. Why was there an EFS difference, but no survival difference? More patients treated initially with prednisone were salvaged following relapse. When one examines the data closely, this makes sense for two reasons. First, dexamethasone had its greatest effect on preventing extramedullary relapses, which are much more salvageable than marrow relapses. Second, many of the relapses prevented by dexamethasone were those that occurred after completion of therapy, because the EFS curves do not diverge at all for the first 2.5 years, 6 months after treatment was completed. Children with relapsed ALL are much more curable when relapse, even marrow relapse, occurs after completion of therapy than when relapse occurs during primary therapy. Thus, one major lesson of this trial is that investigators must strive to identify interventions that prevent relapses that are unlikely to be salvaged, especially marrow relapses occurring on therapy.

However, this is not the only message of the AIEOP-BFM ALL 2000 clinical trial. The authors carefully examined the outcome of different patient subsets and found one for which dexamethasone clearly provided a lasting benefit. About 15% of children with ALL have T-cell ALL (T-ALL), about one-third of whom have poor initial responses to therapy, as measured by the peripheral blood blast count after treatment with 7 days of prednisone and a single dose of intrathecal methotrexate. These prednisone poor-responder patients showed no benefit from dexamethasone. In contrast, the T-ALL patients with a good prednisone response benefitted dramatically from dexamethasone with a 60% reduction in relapse risk, a 45% reduction in EFS events, and a 50% reduction in deaths. In this subset the EFS curves diverge early, before the end of the first year of therapy. So, the relapses that were prevented were the ones that could not be salvaged easily with current therapeutic approaches. For this group, dexamethasone is clearly better than prednisone, and current BFM clinical trials use dexamethasone during induction for this subgroup of patients and this subgroup only.

For these and many other reasons, the AIEOP-BFM ALL 2000 clinical trial has had a substantial impact on contemporary ALL therapy. A parallel randomized comparison of dexamethasone vs prednisone during induction for children with ALL and high-risk clinical features was conducted by the Children’s Oncology Group in North America, Australia, and New Zealand, also with important and interesting results showing that subjects <10 years old benefitted from a 14-day course of dexamethasone during induction, whereas older children did just as well with 28 days of prednisone and had a much lower risk of developing osteonecrosis.⁹  So although questions still remain concerning how to optimally use glucocorticoids in pediatric ALL therapy, a fundamental lesson of this trial is the need to identify interventions that prevent early bone marrow relapses. Such interventions will likely improve survival, which is the ultimate goal of ALL therapy.