Practice patterns and real-life outcomes for patients with acute promyelocytic leukemia in the United States

Abstract

Acute promyelocytic leukemia (APL) is associated with a favorable long-term prognosis if appropriate treatment is initiated promptly. Outcomes in clinical trials and population-based registries vary; potential explanations include a delay in treatment and lower adherence to guideline-recommended therapy in real-world practice. We used the Vizient Clinical Data Base to describe demographic characteristics, baseline clinical characteristics, and treatment patterns in patients newly diagnosed with APL during the study period of April 2017 to March 2020. Baseline white blood cell count was used to assign risk status and assess treatment concordance with National Comprehensive Cancer Network guidelines. Logistic regression models examined adjusted associations between patient, hospital, disease characteristics, and adverse outcomes (in-hospital death or discharge to hospice). Among 1464 patients with APL, 205 (14.0%) experienced an adverse outcome. A substantial subset (20.6%) of patients did not receive guideline-concordant regimens. Odds of adverse outcomes increased with failure to receive guideline-concordant treatment (odds ratio [OR], 2.31; 95% confidence interval [CI], 1.43-3.75; P = .001), high-risk disease (OR, 2.48; 95% CI, 1.53-4.00; P < .001), and increasing age (≥60 years: OR, 11.13; 95% CI, 4.55-27.22; P < .001). Higher hospital acute myeloid leukemia (AML) patient volume was associated with lower odds of adverse outcome (OR, 0.44; 95% CI, 0.20-0.99 [for ≤50 vs >200 AML patients per year]; P = .046). In conclusion, in this large database analysis, 14.0% of patients newly diagnosed with APL died or were discharged to hospice. A substantial proportion of patients did not receive guideline-concordant therapy, potentially contributing to adverse outcomes.

Graphical abstract

Figure 1.

Patient selection flowchart. Patients were selected in a stepwise process. Patients had to have an ICD-10 diagnosis code of “APL, not having achieved remission” (C92.40) and received at least 1 day of ATRA for inclusion. We excluded outpatient and subsequent inpatient encounters, patients who were transferred or had an unknown disposition status, and patients without information on medication administration. The overall cohort included 1464 patients with APL. For analyses evaluating the treatment patterns, patients with LOS ≤3 days and unknown WBC were excluded as well. Data from the Vizient CDB used with permission of Vizient, Inc. All rights reserved.

Figure 2.

Treatment patterns and concordance with NCCN guidelines by baseline disease risk. Distribution of treatment patterns in all patients with a LOS >3 days and known baseline risk status. Low-risk (n = 686) and high-risk (n = 322) patients were defined as WBC ≤10 G/L and WBC >10 G/L, respectively. Treatment regimens were classified based on current NCCN recommendations. ATRA + ATO was defined as the preferred regimen for low-risk APL (blue) with ATRA + ATO + anthracycline or GO constituting the preferred regimens for patients with high-risk APL (green). ATRA + anthracycline or GO can be used independent of disease risk in patients with contraindications to ATO and was included as a risk-agnostic, guideline-concordant treatment regimen among the respective NCCN guideline–concordant regimens in the low-risk (blue) and high-risk (green) patient populations. Other treatment regimens were classified as not NCCN-recommended (red). Overall, 86.1% of low-risk patients and 64.6% of high-risk patients with APL received guideline-concordant regimens for their risk status. Among patients with high-risk APL, 18.9% were treated with ATRA + ATO, a regimen that is NCCN concordant only for patients with low-risk APL. Data from the Vizient CDB used with permission of Vizient, Inc. All rights reserved.