Adherence-enhancing intervention and relapse in childhood acute lymphoblastic leukemia: results from the Children's Oncology Group randomized trial ACCL1033

Durable remissions in children with acute lymphoblastic leukemia (ALL) require an 18-to-24-month-long maintenance phase with daily self-administered oral mercaptopurine (6MP) [ 1 , 2 , 3 , 4 ]. In a previous multicenter study, we showed that 6MP adherence rates <95% were associated with a 2.7-fold higher relapse risk [ 4 ]. Over 40% of relapses in children entering maintenance in first remission were attributable to suboptimal 6MP adherence [ 4 , 5 , 6 ]. The most common reason for missing 6MP was forgetfulness [ 5 ]; adherent patients endorsed parental vigilance to overcome forgetfulness [ 7 ]. These findings prompted a multicenter phase III randomized clinical trial (RCT) to test the efficacy of adherence-enhancing strategies in children and adolescents with ALL receiving 6MP during maintenance [ 8 ]. Parents of children with ALL who were <12 years old (yo) at enrollment, and parents and adolescents with ALL who were ≥12yo were randomized to receive education alone (EDU) or an intervention package (IP) that addressed facilitators/barriers to adherence (i.e., forgetfulness, parental vigilance) [ 5 , 7 ]. The primary aim of the trial was to determine the impact of EDU vs . IP on 6MP adherence; the results of this trial varied by patient age [ 8 ]. While post-intervention 6MP adherence rates were comparably high for EDU vs . IP in the <12yo participants ( p = 0.53), they were significantly higher for IP ( p = 0.04) among the ≥12yo [ 8 ]. In this report we describe results of the exploratory aim that examined relapse risk by intervention arm. We tested the hypothesis that EDU will result in higher relapse risk vs . IP among the ≥12yo participants but will have comparable relapse risks among <12yo participants. This investigator-initiated, unblinded, parallel-group, RCT (ACCL1033; NCT01503632) was conducted at 59 Children’s Oncology Group (COG) sites in the US. Enrollment occurred between 2012 and 2018; participants were followed until August 2025. Eligibility criteria (detailed in the primary report) [ 8 ] included diagnosis of ALL at age ≤21, in first clinical remission at enrollment in adherence trial, receiving 6MP for maintenance. Therapeutic trials are detailed in Supplementary Table S1 . Participants were randomly assigned 1:1 to IP or EDU using stratified block randomization with age at enrollment (<12 y; ≥12 y) and race/ethnicity (non-Hispanic whites, Hispanics, African Americans, others) as stratification factors. The education program included video vignettes of ALL patients/parents addressing perceived susceptibility/severity of ALL, perceived benefits/barriers to 6MP ingestion, and examples of how patients/parents overcame such barriers. Study participants (EDU and IP) viewed the video in clinic at start of intervention, self-tailoring viewing to fit linguistic preferences (English/Spanish). Intervention components for IP arm included (i) education (as in EDU), (ii) daily personalized text-messages from healthcare provider to parent (of <12yo patient) or patient and parent (of ≥12yo patient) using a HIPAA-compliant web-based application ( MedActionPlan/CT ), and (iii) daily directly supervised therapy by designated parent [ 8 ]. For the first 28 days after enrollment, participants in both arms received 6MP from a Medication Event Monitoring system device (MEMS) without intervention to calculate baseline adherence rates. MEMS uses microelectronic technology to record date and time of 6MP bottle openings [ 9 ]. Adherence rate was defined as the ratio of days with MEMS openings to days 6MP was prescribed for each patient. Intervention lasted 16 weeks. Participating sites submitted annual reports, detailing dates of last visit, relapse or death for up to 10 y after enrollment. An intention-to-treat analysis compared relapse risk between intervention arms using sub-distribution hazards regression analysis, treating non-relapse death as competing risk [ 10 ]. Analyses were conducted using SAS statistical software v9.4 (SAS Institute). Mean (SD) adherence rates were higher among those in complete continuous remission (92%[12%]) vs. those who relapsed (86%[17%], p = 0.059). The 5 y cumulative incidence of relapse was higher among non-adherent patients (adherence rate <95%) vs . adherent patients (adherence rate ≥95%): 13.4% (95% CI = 6.4–18.4) vs. 6.0% (95% CI = 3.1–8.5), p = 0.009 (Supplementary Table S2 ). Analyses were conducted in the entire cohort by age group (<12 y; ≥12 y). As a post-hoc analysis, we examined the risk of relapse in a subgroup of participants with end-of-induction minimal residual disease (EOI MRD) ≤ 1% to minimize the influence of disease biology. EOI MRD status was not available for 71 patients (16%); in such cases, those with favorable cytogenetics were arbitrarily assigned to the EOI MRD ≤ 1% group. Baseline characteristics were comparable between participants on EDU ( n = 143) and IP ( n = 147) (Table 1 ). A median follow-up of 7.7 y (0.2–10.3) yielded a 5 y cumulative incidence of relapse of 4.4% vs . 7.3% for EDU vs . IP, respectively, p = 0.31 (Supplementary Table S3 ). Multivariable analysis adjusted for age at enrollment, race/ethnicity, time from start of maintenance to study enrollment, and EOI MRD revealed no difference in relapse risk between the two arms (EDU: Hazard Ratio [HR] = 0.63, 95% CI = 0.2–1.7, p = 0.35; reference: IP) (Fig. 1A , Supplementary Table S4 ). A Hazard of relapse from start of intervention to date of relapse, death, or last follow-up among those randomized to EDU (reference: IP) in the two age groups (entire cohort). Adjusted for age at trial enrollment, race/ethnicity, time from start of maintenance to study enrollment and EOI MRD. B Hazard of relapse from start of intervention to date of relapse, death, or last follow-up among those randomized to EDU (reference: IP) in the two age groups (patients with EOI MRD ≤1%). Adjusted for age at trial enrollment, race/ethnicity, and time from start of maintenance to study enrollment. EDU denotes Education alone; IP denotes the Intervention Package. EOI MRD denotes end of induction MRD. Baseline characteristics were comparable between participants on EDU ( n = 71) and IP ( n = 83) (Table 1 ), except for a lower prevalence of patients with EOI MRD ≤ 1% on EDU vs . IP arms (14% vs . 28%, p = 0.04). Median follow-up of 4.86 y (0.08–9.45) yielded a 5 y cumulative incidence of relapse of 15.1% vs . 10.5% for EDU vs . IP, respectively, p = 0.47 (Supplementary Table S3 ). Multivariable analysis adjusted for factors listed above revealed a higher relapse risk among patients on EDU (HR = 1.84, 95% CI = 0.7–4.9, p = 0.22; reference: IP arm) (Fig. 1A , Supplementary Table S4 ); however, this difference did not reach statistical significance. Baseline characteristics of participants on EDU ( n = 124) and IP ( n = 129) (Table 1 ) were comparable. A median follow-up of 8.1 y (0.2–11.0) yielded a 5 y cumulative incidence of relapse for EDU vs . IP was 3.5% vs . 5.3%, p = 0.48 (Supplementary Table S3 ). Multivariable analysis adjusted for age at study, race/ethnicity, and time from start of maintenance to enrollment revealed no difference in relapse risk between the two arms (EDU: HR = 0.7, 95% CI = 0.2–2.1; p = 0.52; reference: IP arm) (Fig. 1B , Supplementary Table S5 ). Baseline characteristics were comparable between participants on EDU ( n = 61) and IP ( n = 60) (Table 1 ). A median follow-up of 5.5 y (0.2–9.5) yielded a 5 y cumulative incidence of relapse of 14.8% vs. 3.9% for EDU vs . IP, respectively, p = 0.07 (Supplementary Table S3 ). Multivariable analysis revealed a higher risk of relapse among patients on the EDU arm (HR = 5.1, 95% CI = 1.1–24.4, p = 0.04; reference: IP arm) (Fig. 1B , Supplementary Table S5 ). We have previously reported the primary results of our multicenter RCT [ 8 ], showing that the adherence-enhancing intervention resulted in higher adherence in ≥12yo patients with ALL, but comparably high adherence rates for both arms among the <12yo. Adolescents are especially likely to have suboptimal adherence [ 5 , 6 ], because of increasing assumption of independence and decreasing parental supervision, resulting in this subgroup drawing benefit from our adherence-enhancing intervention [ 8 ]. In the current report, we describe whether the benefit from improving 6MP adherence resulted in reduction in relapse risk, testing the hypothesis that the hazard of relapse would be comparable between the two arms in the younger age group, but would be higher in the older age group randomized to EDU vs . IP. We tested this hypothesis in the entire cohort and also among those with EOI MRD ≤ 1% to ensure that we were examining the impact of the intervention on relapses likely attributable to poor adherence and not to disease biology. Among <12yo participants, we observed no difference in relapse risk between the two arms in the entire cohort, as well as among those with end of induction MRD was ≤1%. However, among ≥12yo participants, we observed a two-fold greater hazard of relapse in the entire cohort, and a 5-fold greater hazard of relapse in the sub-cohort with EOI MRD ≤ 1% in the EDU vs. IP arm. These findings suggest that improving 6MP adherence in the older age group may have helped prevent disease relapse. Findings of this study should be interpreted in the context of certain limitations. In a post hoc analysis, we excluded patients with EOI MRD > 1% to focus on relapses that would be more likely to be attributable to non-adherence to therapy rather than disease biology. EOI MRD (a measure of clone-specific leukemic markers) is a strong prognostic marker and is highly predictive of relapse in ALL. EOI MRD cut points employed to make clinical decisions include <0.01% (most favorable), 0.01% to <1.0% (intermediate), and ≥1.0% (unfavorable) [ 11 ]. Our results were comparable when patients with the subcohort were restricted to EOI MRD < 0.01% (Supplementary Table S6 ). Furthermore, in the cohort with EOI MRD ≤ 1%, there was no difference in other high-risk features between those in remission vs . those who had relapsed (Supplementary Table S7 ). The relatively small number of relapse events as well as limitations associated with any post hoc analyses need to be acknowledged. These limitations notwithstanding, to our knowledge, this study is the first RCT in pediatric oncology that examined the impact of a 6MP adherence-enhancing intervention on relapse risk, showing that improvement in adherence may have reduced relapse risk. These findings support continued focus on adherence-enhancing strategies in patients at risk for poor adherence. Along these lines, we are currently conducting two parallel RCTs tailored to baseline adherence, with intensification of the intervention for the baseline non-adherers and a parent/patient-empowered intervention for baseline adherers. The datasets generated and/or analyzed during the current study are not publicly available due to restrictions related to participant and protocol-wide study permissions, but are available from the corresponding author on reasonable request and with permission from the Children’s Oncology Group. We would like to thank the patients and families who participated in this study. This study was supported in part by the Children’s Oncology Group, the St Baldrick’s Foundation, and the National Institutes of Health under award numbers R01CA174683, U10CA098543, U10CA098413, U10CA095861, U10CA180886, U10CA180899, and UG1CA189955. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Institute for Cancer Outcomes and Survivorship, University of Alabama at Birmingham, Birmingham, AL, USA Smita Bhatia, Lindsey Hageman, Yanjun Chen, Aman Wadhwa & Wendy Landier Department of Population Sciences, City of Hope, Duarte, CA, USA F. Lennie Wong Department of Psychiatry and Human Behavior, Brown University, Providence, RI, USA Elizabeth L. McQuaid Cancer and Blood Disease Institute, Division of Hematology/Oncology, Children’s Hospital Los Angeles, Los Angeles, CA, USA David R. Freyer Department of Pediatric Hematology/Oncology, Driscoll Children’s Hospital, Corpus Christi, TX, USA Nkechi Mba Department of Pediatrics, University of California, San Diego, CA, USA Paula Aristizabal Department of Pediatrics, NYU Langone Medical Center, New York, NY, USA Elizabeth Raetz Search author on: PubMed Google Scholar Search author on: PubMed Google Scholar Search author on: PubMed Google Scholar Search author on: PubMed Google Scholar Search author on: PubMed Google Scholar Search author on: PubMed Google Scholar Search author on: PubMed Google Scholar Search author on: PubMed Google Scholar Search author on: PubMed Google Scholar Search author on: PubMed Google Scholar Search author on: PubMed Google Scholar SB, WL, FLW, and ELM designed the study. SB, WL, and LH supervised the study. YC, SB, and LH analyzed the data. SB and WL wrote the first version of the manuscript. AW, DRF, NB, PA, and ER served as investigators, collected data, and edited the manuscript. All authors approved the submitted manuscript. Correspondence to Smita Bhatia . The authors declare no competing interests. Approval for the Children’s Oncology Group ACCL1033 trial was obtained from the Pediatric Central Institutional Review Board of the National Cancer Institute, by the University of Alabama at Birmingham Institutional Review Board (IRB-120904013), and by institutional review boards at all participating institutions. Written informed consent/assent was obtained from participants and/or parents. All procedures adhered to relevant institutional and governmental guidelines and regulations. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . Reprints and permissions Received : 06 November 2025 Revised : 05 March 2026 Accepted : 27 March 2026 Published : 13 April 2026 Version of record : 13 April 2026 DOI : https://doi.org/10.1038/s41375-026-02951-0 Anyone you share the following link with will be able to read this content: Sorry, a shareable link is not currently available for this article. Provided by the Springer Nature SharedIt content-sharing initiative