Toxicity during induction of pulsed versus continuous prednisolone in children with acute lymphoblastic leukaemia: a multi-centre, open label, randomised, phase 3 trial from India (2016–2022)

Abstract

Background

Despite the use of similar treatment protocols, induction treatment-related mortality (TRM) in children with ALL in India is 5%, compared to <1% in high income countries. This study evaluated pulsed prednisolone during induction to reduce toxicity.

Methods

ICiCLe-ALL-14 (CTRI/2015/12/006434), an open-label, randomised, multicentre trial. Children ≤ 10-years with newly diagnosed B-cell precursor ALL, risk-stratified to standard- or intermediate-risk, were randomised (1:1) to receive prednisolone 60 mg/m²/day, for 4 weeks with taper (R1A) or on days 1–14 and 22–28 (R1B) during induction. Primary outcomes were grade 3–5 toxicities. Secondary outcomes were complete remission (CR) rates, measurable residual disease (MRD) at end of induction (EoI), event-free (EFS), and overall survival (OS).

Findings

Of 3315 patients screened at 6 centres, 2505 were eligible and 1246 randomised (623 per group). Thirty (2.4%) patients died during induction: 22/623 (3.5%) and 8/623 (1.3%) in R1A and R1B respectively (absolute risk difference 2.3%, 95% CI 0.54–4.1; p = 0.0149). Relative risk of induction death was 2.75 (95% CI 1.2–6.1) for R1A versus R1B. Grade 3–5 toxicities occurred in 46.4% (289/623) and 45.4% (283/623) of patients (p = 0.7762). CR rates were 98.8% (581/588) and 98.0% (594/606); MRD ≥ 0.01% at EoI was 26.2% and 27.9%; 3-year EFS 72.7% (95% CI 68.2–76.7) and 72.2% (67.6–76.3); OS 85.0% (81.5–87.8) and 87.0% (83.6–89.3) in R1A and R1B respectively. Multivariable analysis confirmed higher risk of induction death with continuous prednisolone (HR 3.06, 95% CI 1.4–6.9, p = 0.0069) and anthracycline use (HR 5.44, 95% CI 1.6–18.9 p = 0.0077).

Interpretation

Pulsed prednisolone (R1B) during induction significantly reduced treatment-related deaths, particularly early deaths from infections compared to continuous dosing (R1A), without compromising CR rates, MRD clearance, EFS, or OS. Anthracycline use was an independent risk factor for induction mortality. Higher white cell count and anthracycline use independently predicted grade 3–4 toxicity.

Funding

This study was funded partly by the Indian Council of Medical Research (Reference 79/159/2015/NCD-III), the National Cancer Grid (Reference 2016/001), DBT-Wellcome India Alliance (Margdarshi IA/M/12/1/500261) and a centre grant from the Tata Consultancy Services Foundation to the trial centre.