In-hospital mortality, hemorrhagic complications, and sociodemographic disparities in acute promyelocytic leukemia: a nationwide analysis of the United States, 2016 to 2023

Abstract Acute promyelocytic leukemia (APL) is highly curable with all-trans retinoic acid and arsenic trioxide, yet early in-hospital death from hemorrhage and organ failure continues to limit outcomes in routine practice. Contemporary national data on inpatient mortality, hemorrhagic complications, and sociodemographic disparities in the United States remain limited. We performed a retrospective cross-sectional analysis of adult APL hospitalizations in the National Inpatient Sample. The primary cohort spanned seven calendar years (2016 to 2022); temporal trend and non-race sensitivity analyses extended to 2023 (an eight-year window). APL was identified by International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) code C92.4. The primary outcome was in-hospital mortality. Secondary outcomes were intracranial hemorrhage and disseminated intravascular coagulation (DIC). Survey-weighted multivariable logistic regression models adjusted for age, sex, race, primary payer, household income quartile, hospital characteristics, admission type, and acute comorbidities. Annual outcome rates from 2016 to 2023 were modeled by survey-weighted linear regression. Four prespecified sensitivity analyses were performed. We identified 4,591 unweighted APL hospitalizations representing 22,955 (95% CI, 22,006 to 23,904) weighted admissions from 2016 to 2022. In-hospital mortality was 8.60%, intracranial hemorrhage occurred in 6.36%, and DIC in 18.33%. Annual in-hospital mortality did not change significantly across the eight-year trend window (β = +0.20% per year, p = 0.243). In the multivariable model, the strongest factors associated with in-hospital mortality were acute respiratory failure (adjusted odds ratio [aOR] 11.94, 95% CI 9.04 to 15.78), acute kidney injury (aOR 2.93, 95% CI 2.22 to 3.87), uninsured status (aOR 2.94, 95% CI 1.49 to 5.79), age 65 years or older (aOR 2.55, 95% CI 1.57 to 4.16), and severe liver disease (aOR 1.90, 95% CI 1.22 to 2.94). Model discrimination was high (C-statistic 0.856). The uninsured association persisted across all four sensitivity analyses. Compared with other AML subtypes, APL admissions carried higher adjusted odds of intracranial hemorrhage (aOR 3.04, 95% CI 2.66 to 3.49) and DIC (aOR 10.69, 95% CI 9.64 to 11.87). Despite a curative paradigm, in-hospital APL mortality has not improved in the United States from 2016 to 2023. Mortality is concentrated in patients with acute respiratory failure, kidney injury, severe liver disease, and sepsis, and falls disproportionately on the uninsured. The APL bleeding phenotype remains distinct from other AML subtypes. Systems-level interventions targeting early recognition, immediate empiric ATRA, supportive care, and access for the uninsured are needed to close the gap with trial-level outcomes.