This retrospective observational cohort study used the 100% de-identified Medicare FFS database. Medicare is a federal health insurance program for those aged 65 and older, certain people under 65 with disabilities, and people of any age with end-stage renal disease. Medicare covers about 96% of all US citizens aged 65 and older [21]. Medicare Fee-For-Service Data, which composes about half of the total enrolled Medicare beneficiaries, include claims for inpatient, skilled care nursing facility, and hospice care (Part A) as well as outpatient care (Part B) and prescription drugs (Part D) [22]. Eligible patients were identified as patients aged ≥ 65 years with CLL who initiated cBTK inhibitor monotherapy at 1L. The index was defined as the start date of the treatment initiation with 1L cBTK inhibitor monotherapy. The baseline period was defined as the 12-month period prior to the index date, unless otherwise specified. The follow-up period was defined as the time from the index date to the end of record (date of death, end of enrollment, or data cutoff). Data cutoff was December 31, 2025.

Eligible patients had ≥ 2 Medicare Part A, B, or C medical claims with diagnostic codes (International Classification of Disease, 10th Revision [ICD-10]) for CLL as a primary diagnosis on separate dates ≥ 14 days apart on or after 2010; received 1L zanubrutinib, acalabrutinib, or ibrutinib between January 1, 2020 and September 30, 2025; and had continuous enrollment in Medicare Part A, B, and D during the baseline period. Patients were excluded if they received any other CLL treatment before index, including combination or concurrent therapy with other agents, or had clinical trial participation during the study period (Fig. 1).

Flowchart of inclusion and exclusion for patient cohort. CLL chronic lymphocytic leukemia, SLL small lymphocytic lymphoma, BTK Bruton tyrosine kinase

Patients were classified as receiving 1L cBTK inhibitor monotherapy if no additional therapy was initiated within 90 days of the index date. Any additional CLL regimens received more than 90 days from the index date are considered second-line (2L) treatment, with the exception of re-initiation of the same cBTK inhibitor, which is classified as 1L treatment if it occurs within 180 days after the end of the prior cBTK inhibitor supply does not advance to 2L.

Baseline demographic and clinical characteristics included age at the index date (continuous and categorical), race/ethnicity, sex, region, socioeconomic status (dual Medicare and Medicaid eligibility, low income subsidy), rural–urban community area (RUCA; urban, rural; defined by the US Census Bureau based on geographical zones), year of index date, time from initial diagnosis to index date, Charlson Comorbidity Index (CCI, excluding any B cell malignancies: 0, 1–2, 3–4, ≥ 5), and other comorbidities of interest. Baseline demographic and clinical characteristics were summarized descriptively. Sex was defined based on beneficiary’s biological sex at birth, sourced from the CMS enrollment database. “Other” racial category is used by the Social Security Administration’s Master Beneficiary Record to identify individuals of races other than White, Black, Asian/Pacific Islander, or American Indian/Alaska Native and when an individual does not fall into these specific, mutually exclusive categories.

Real-world time to treatment discontinuation (rwTTD) was defined as the time from index date to the earliest treatment discontinuation (i.e., last prescription fill date and supply), death, or censored at the last confirmed activity (end of enrollment or end of study). Discontinuation was defined as having a subsequent systemic therapy after the current line of therapy (LOT), having a gap of > 180 days with no systemic therapy after the last drug episode of the current LOT, or having a date of death. A gap of > 180 days was selected to help distinguish temporary treatment interruptions (e.g., due to surgery, travel, or other non-disease-related reasons) from true treatment discontinuation followed by re-initiation upon disease progression. Given the treatments of interest were oral therapies and this was an older patient population, this threshold was intentionally chosen to be conservative, with 180 days representing a relatively liberal interval to minimize the risk of misclassifying short-term treatment holds as discontinuation. Patients were censored at the end of the current LOT if they did not have treatment discontinuation. Real-world time to next treatment (rwTTNT) was defined as the time from index date to the earliest of either the start of the next LOT, death, or censoring at the last confirmed activity. Real-world overall survival (rwOS) was defined as the time from the index date to the date of death or censoring on the last confirmed activity.

Descriptive statistics were used to summarize baseline and clinical characteristics including median, range, and interquartile range (IQR) for continuous variables, and relative frequency and percentage for categorical variables. Patient characteristics were reported for the overall 1L cBTK inhibitor population and by each index therapy.

The Kaplan–Meier method was used to generate survival curves for rwTTD, rwTTNT, and rwOS. Median time to event and survival and treatment probabilities with corresponding 95% CIs were calculated at 12 and 24 months after the index date by treatment group. Kaplan–Meier curves were truncated when patient counts were ≤ 10. Unadjusted and multivariate Cox proportional hazard models were used to estimate HRs and corresponding 95% CIs for treatment comparisons (zanubrutinib versus ibrutinib or acalabrutinib). Adjusted covariates included categorical age (65–74, 75–84, and ≥ 85), sex, race/ethnicity (non-Hispanic White versus other), CCI score (0, 1–2, 3–4, ≥ 5), and year of index. Stratified analyses were also performed by categorical age group. A P value less than 0.05 was the cutoff for statistical significance.

No institutional review board/independent ethics committee review was required for this secondary analysis of de-identified existing data. All work performed by ADVI Health was subject to their respective requirements for access. The data used in this study are derived from Medicare claims files maintained by the Centers for Medicare & Medicaid Services (CMS). ADVI Health has a Data Use Agreement (DUA) agreement and access to these claims through a CMS Virtual Research Data Center (VRDC) license. Data are not publicly available because of federal restrictions protecting beneficiary privacy. Qualified researchers interested in accessing these data must apply directly to CMS through the Research Data Assistance Center (ResDAC).

All data presented are aggregated in accordance with CMS privacy standards. All source data were de-identified by CMS. Individual patient consent was not required, as the data were anonymized and reported in aggregated.