AFFILIATIONS
1 University of Minnesota, Minneapolis, Minnesota, United States
2 Cleveland Clinic, Weston, Florida, United States
3 Crestwood Medical Center, Huntsville, Alabama, United States
ABSTRACT
Introduction
Bruton tyrosine kinase inhibitors (BTKi) are standard therapy in chronic lymphocytic leukemia (CLL), but ibrutinib carries well-described cardiovascular and bleeding toxicity. Acalabrutinib demonstrated fewer cardiac events than ibrutinib in ELEVATE-RR; however, that trial excluded patients requiring anticoagulation, leaving comparative safety unresolved in the population where the choice between agents is most consequential: patients with pre-existing atrial fibrillation (AF). We compared bleeding, cardiovascular, and mortality outcomes between acalabrutinib and ibrutinib in CLL patients with AF documented prior to treatment initiation.
Methods
We conducted a retrospective cohort study using the TriNetX US Collaborative Network, a federated EHR database of approximately 128 million patients across 70 healthcare organizations. Adults with CLL (ICD-10-CM C91.1) and atrial fibrillation or flutter (I48) documented prior to index were identified. Cohort 1 initiated acalabrutinib; Cohort 2 initiated ibrutinib. Patients with any exposure to the comparator BTKi, zanubrutinib, or pirtobrutinib were excluded, as were patients with mantle cell lymphoma, Waldenstrom macroglobulinemia, extranodal marginal zone lymphoma, follicular lymphoma, or graft-versus-host disease. The index event was first administration of the cohort-defining BTKi. Cohorts underwent 1:1 propensity score matching (greedy nearest neighbor, caliper 0.1 SD) on age, sex, race, ethnicity, hypertension, heart failure, ischemic heart disease, type 2 diabetes, chronic kidney disease, cerebral infarction, transient ischemic attack, tobacco use, obesity, long-term antithrombotic use, anticoagulants, rituximab, obinutuzumab, venetoclax, and lactate dehydrogenase. Outcomes were assessed 1 to 1,095 days after index, excluding patients with the outcome prior to the window. Risk ratios (RR) and Kaplan-Meier hazard ratios (HR) with 95% confidence intervals were calculated.
Results
Prior to matching, 773 acalabrutinib and 1,709 ibrutinib patients met criteria. After 1:1 propensity score matching, 624 patients remained per cohort, with all standardized mean differences below 0.1. All-cause mortality was significantly lower with acalabrutinib (26.3% vs 32.6%; RR 0.81, 95% CI 0.68-0.96, p=0.015), a difference confirmed in time-to-event analysis (HR 0.79, 95% CI 0.64-0.97; log-rank p=0.021), with 3-year survival probability 68.0% versus 62.3%. No bleeding endpoint differed significantly between cohorts. Gastrointestinal hemorrhage occurred in 3.9% versus 4.9% (RR 0.80, 95% CI 0.47-1.37; HR 0.78, log-rank p=0.38), melena in 4.4% versus 4.0% (RR 1.09, 95% CI 0.64-1.88; HR 1.07, p=0.82), acute posthemorrhagic anemia in 7.3% versus 8.5% (RR 0.86, 95% CI 0.58-1.28; HR 0.84, p=0.41), and other nontraumatic intracranial hemorrhage in 2.2% versus 3.5% (RR 0.64, 95% CI 0.33-1.23; HR 0.74, p=0.42). Subarachnoid and intracerebral hemorrhage counts fell below the reporting threshold. Cardiac endpoints also did not differ: ventricular tachycardia 4.6% versus 5.5% (RR 0.83, 95% CI 0.50-1.37; HR 0.81, p=0.42) and heart failure 21.2% versus 23.0% (RR 0.92, 95% CI 0.72-1.18; HR 0.91, p=0.53).
Conclusions
In this real-world cohort of CLL patients with pre-existing atrial fibrillation, a population underrepresented in randomized comparison, acalabrutinib was associated with significantly lower all-cause mortality than ibrutinib, consistent across both risk-ratio and time-to-event analyses. Notably, this survival advantage occurred without any measurable difference in bleeding, including gastrointestinal and intracranial hemorrhage, despite the anticoagulant burden expected in this population. Rates of ventricular arrhythmia and heart failure were also comparable. These findings suggest that the survival benefit is not mediated by reduced hemorrhagic events and may instead reflect differences in treatment persistence, non-hemorrhagic toxicity, or residual confounding by unmeasured disease severity, although a prespecified negative control outcome argued against gross residual confounding. Limitations include retrospective design, reliance on diagnostic coding, incomplete overlap of propensity distributions before matching, potential differential follow-up given the later approval of acalabrutinib, and unmeasured confounding by indication. These results support acalabrutinib as a reasonable choice in CLL patients with pre-existing AF and warrant prospective evaluation.
DISCLOSURES
Maida Sarfraz Chaudhry, MD: No relevant conflicts of interest to declare.