Analysis of Treatments and Outcomes for Patients with De Novo AML, Therapy-Related AML, and Secondary AML (Prior MDS and CMML) Diagnosed in England between 2011 and 2016 Using Hospital Episode Statistics®

Introduction: Historically, patients with therapy-related acute myeloid leukemia (t-AML) or secondary AML (s-AML; prior myelodysplastic syndrome [_MDSAML] or chronic myelomonocytic leukemia [_CMMLAML]) have poor outcomes when treated with conventional intensive chemotherapy (IC). Alternative, less-intensive therapies include azacitidine and low-dose cytarabine (LDAC). The objective of this retrospective study was to utilize the Hospital Episode Statistics^© (HES) database to describe the historical treatment patterns in England from 2011 to 2016 and to analyze long-term outcomes for patients with de novo AML, t-AML, or s-AML.

Methods: The HES database contains details of all admissions, accident and emergency attendances, and outpatient appointments at National Health Service (NHS) hospitals in England. Adult patients (³18 years) diagnosed with AML between NHS years (April-March) 2011/2012 to 2016/2017 were identified through International Classification of Diseases, Tenth Revision (ICD-10) codes (C92*, excluding C921, C922, and C927). There are no ICD-10 codes specific to t-AML or s-AML, so these patients were identified through a history of relevant cancer and specific OPCS-4 procedure codes for prior chemotherapy (X7*) or radiotherapy (X65*), or a history of MDS (ICD-10: D46*, excluding D463) or CMML (ICD-10: C931). Patients with AML who did not fall into these criteria were classified as de novo AML. Based on OPCS-4 codes, patients were allocated to the following AML treatment pathways: IC ± hematopoietic cell transplantation (HCT), azacitidine, or LDAC. Patients who did not receive active systemic therapy (ie, best supportive care alone) were excluded. Median patient follow-up was 5.3 years (100% range: 2.5, 8.8).

Results: In total, 9,758 patients with AML were identified, comprising 7,499 (77%) with de novo AML, 803 (8%) with t-AML, 1,305 (13%) with _MDSAML, and 151 (2%) with _CMMLAML. Patients with de novo AML were younger (median [interquartile range] age: 63 years [50, 72]) compared to those with t-AML (67 years [56, 73]), _MDSAML (70 years [63, 76]), and _CMMLAML (68 years [62, 74]). Patients with de novo AML were more likely to receive IC ± HCT overall and across all age groups. In patients 60 to 69 and 70 to 79 years, 76% and 36% of those with de novo AML received IC ± HCT, respectively, compared to 54% to 56% and 15% to 22% of patients with t-AML/s-AML. The lower use of IC ± HCT in older patients with t-AML/s-AML versus de novo AML was due to the preferential use of azacitidine. In patients 60 to 69 and 70 to 79 years, 35% to 38% and 54% to 57% of patients with t-AML/s-AML received azacitidine, respectively, compared to 17% and 32% with de novo AML.

Patients with de novo AML had longer Kaplan-Meier–estimated overall survival (OS) versus those with t-AML/s-AML combined (median: 1.79 vs 0.94 years; unadjusted hazard ratio = 1.66 [95% confidence interval: 1.58, 1.75]; nominal P <0.0001). For patients treated with IC ± HCT (all ages), the 5-year Kaplan-Meier OS estimates were higher for patients with de novo AML versus those with t-AML or s-AML (46% vs 36% and 24%, respectively). Compared to azacitidine or LDAC, 5-year OS estimates were consistently higher with IC ± HCT across all AML subgroups (see Table).

Conclusions: This large, retrospective analysis has confirmed the poor historical outcomes for patients diagnosed with t-AML or s-AML in England. Compared to patients with de novo AML, those with t-AML or s-AML were less likely to receive IC ± HCT across all age groups. In addition, when treated with IC ± HCT, their outcomes were inferior to those of patients with de novo AML. However, despite this poor prognosis, 5-year OS estimates for patients with t-AML or s-AML were improved with IC ± HCT versus both azacitidine and LDAC therapy. Copyright^© 2020. Re-used with the permission of NHS Digital. All rights reserved.

Abstract previously published by EHA in HemaSphere, 2020;4:S1.