Safe Tension: What the Latest Data Say About Waiting to Induce in AML

Recently, a fellow on the inpatient leukemia service said, “I don’t know why you all get so excited about ‘acute leuks’; all we do is sit around and wait.” As the attending of record, I was a bit taken aback by the comment at first, but then realized this fellow was accurate in recollections from medical school and residency that the recommendation is to start chemotherapy in acute leukemia patients as soon as the diagnosis is made. This rapid intervention approach is based on the observation that acute leukemia carries a grave prognosis if untreated. Additionally, there are existing literature analyses demonstrating less favorable outcomes in patients with delayed treatment initiation.^1,2  Waiting multiple inpatient days for a plan for induction chemotherapy is uncomfortable for the patient and the team after a new diagnosis of acute myeloid leukemia (AML). There are patient, disease, and team expectations, as well as acute medical issues, that all require urgent attention. I wanted to reassure the fellow that we “wait” to safely begin treatment, and with good reason.

In a recent issue of Blood, Dr. Christoph Röllig and colleagues conducted a retrospective analysis of more than 2,200 patients with AML (patients with acute promyelocytic leukemia and those treated beyond 50 days from diagnosis were excluded) who received therapy in one of 46 centers in Germany. Each of these centers logged their patients’ data in the AML registry of the Study Alliance Leukemia (SAL) Study Group. This robust registry catalogs non-trial patients’ disease characteristics, treatments, and outcomes for leukemia. All patients included in the current analysis received intensive induction treatment between 2011 and 2019. The induction regimens were reasonably homogeneous, utilizing cytarabine backbones with an anthracycline or mitoxantrone. All patients had follow-up data for at least one year. Their baseline data included age, sex, white blood cell count (WBC), European LeukemiaNet (ELN) 2017 risk, de novo versus secondary AML, and use of cytoreductive hydroxyurea prior to start of intensive induction.

The cohort was subdivided into four groups based on time to treatment of zero to five days, six to 10 days, 11 to 15 days, and more than 15 days. The median time to treatment overall was three days (interquartile range [IQR], 2-7 days). The median age across all groups was 59 years, and favorable, intermediate, and adverse genetic risks according to ELN2017 were present in 27 percent, 54 percent, and 19 percent of patients, respectively. Three-fourths of the cohort was categorized as de novo AML, and the remaining 25 percent as secondary or treatment-related AML. The median WBC was 8 × 10⁹/L across all groups, but, not unexpectedly, the zero- to five-day treatment group had a higher median WBC at presentation (14 × 10⁹/L) compared with the other three. While the data are not shown, the authors allude to no adverse effect from the use of cytoreductive hydroxyurea. The early death rate was less than 5 percent for all four groups. In a multivariate analysis of overall survival (OS; further stratified by age ≤ 60 and > 60 years, and for high vs. lower WBC at presentation), no significant differences between timing groups were found. Also reassuring were the very similar unadjusted complete remission (CR) rates of 79 percent, 76 percent, 72 percent, and 77 percent, respectively, across the cohorts. After a median follow-up of 42 months (IQR, 25-60 months), the two-year OS rate for all patients was 50 percent. This did include patients who proceeded in CR1 or CR2 to transplantation.

The manuscript’s discussion is masterfully composed by clinician scientists who are clearly adept at data analysis, but also proficient at real-world management of patients with AML. Their conclusions center around two key features: Their first point is that credit is due to bedside physicians who know which patients need urgent treatment and which patients require diagnostic and therapeutic optimization prior to induction initiation. This ties into the lack of harm from cytoreduction and benefit of antibiotics and transfusion support. The second discussion topic, and perhaps the more relevant one in today’s therapeutic climate, is the recognition that biology supersedes much of what we manage proactively in human beings. Ultimately, the stronger determinants of response to treatment, relapse, and OS are the patient-specific factors such as age or organ function and leukemia-specific factors including genetic alterations. The authors summarize the data presented with the conclusion that waiting for genetic data to guide induction chemotherapy is reasonably safe and likely prudent.