Comparison of Frontline Therapies in Older Adults Age ≥ 80 Years With Chronic Lymphocytic Leukemia (CLL): A Mayo Clinic and Danish Nation-Wide Study

People with chronic lymphocytic leukemia (CLL) have a median age at diagnosis of around 70 years [1]. With an average time to first treatment of 4–5 years from diagnosis, it is estimated that a third of all people with CLL will be ≥ 80 years old when they become symptomatic and receive frontline therapy [1]. People ≥ 80 years old often have comorbidities and unfavorable disease characteristics (like unmutated IGHV and 17p deletion) associated with lower survival [2]. Despite efforts to include advanced age in the eligibility criteria, landmark CLL clinical trials that led to the approval of ibrutinib and venetoclax primarily included participants < 75 years old, making it challenging to apply their findings to older patients. To help close this knowledge gap, we conducted a multi-center retrospective cohort study with data spanning nearly three decades to describe treatment outcomes for patients treated for CLL in the frontline setting at ages ≥ 80 years.

Using the Mayo Clinic CLL Database and the Danish Lymphoid-lineage Cancer Research (DALY-CARE) data (Danish Cohort), we conducted a retrospective cohort study of patients who were ≥ 80 years old at the time of first CLL treatment between January 1995 and November 2022 (Mayo Clinic) and January 2008 and March 2022 (Denmark). The Mayo Clinic CLL Database contains records of consenting patients diagnosed with CLL who were seen at Mayo Clinic in Rochester, MN. The Danish Cohort includes all Danish citizens diagnosed with CLL, with data available from time of birth or immigration, enabled by linkage of personal-level data across national administrative and clinical quality registers using a unique personal identification number. We extracted information about clinical characteristics, patient demographics, and all lines of treatment (Table S1). We calculated overall survival (OS) from the date of first treatment to the date of death or date last known to be alive. OS was displayed using the Kaplan–Meier method. Time to next treatment (TTNT) was calculated from the date of first treatment to earliest date of (a) date of second treatment, (b) date of death, or (c) date last known to be alive. TTNT was visualized using the cumulative incidence method with competing risk of death. Univariable Cox proportional hazards models were used to analyze the association between type of treatment and time to events (i.e., OS or TTNT); results were reported as hazard ratios (HR) and 95% confidence intervals (CI). Analyses were conducted using statistical software SAS 9.4 (SAS Institute Inc., Cary, NC, USA) and R statistical software versions 4.2.0 and 4.3.2. The Mayo Clinic Institutional Review Board and the Danish Health Data Authority and National Ethics Committee (approvals P-2020-561 and 1804410, respectively) approved this study.

We identified a total of 653 patients aged ≥ 80 years at the time of frontline CLL treatment (216 from Mayo Clinic and 437 from Denmark). The patients received the following types of frontline treatment: alkylators (such as chlorambucil or cyclophosphamide, n = 354), anti-CD20 monoclonal antibody only (such as rituximab or obinutuzumab, n = 90), purine analogues (such as fludarabine and pentostatin, n = 55), novel agents (n = 82), and other/unknown (n = 72). Overall, the median OS was 3.1 years from the start of first CLL treatment. The 2-, 5-, and 10-year OS rates were 63% (95% CI 60%–67%), 29% (95% CI 25%–33%), and 8% (95% CI 6%–12%), respectively. By treatment category, the median OS was 2.5 years for patients who received alkylators, 3.9 years for anti-CD20 monoclonal antibody alone, 3.5 years for purine analogues, not reached for novel agents, and 2.6 years for other/unknown. (Figure 1). In univariable analysis, use of non-novel agent-based therapy (HR 2.8, 95% CI 1.8–4.4, p < 0.001) was associated with a shorter OS. Sensitivity analyses were conducted by analyzing each cohort separately to ensure that no important differences were obscured when combining the cohorts (Figures S2–S4, Tables S1 and S2).

FIGURE 1

Open in figure viewerPowerPoint

Overall survival of older adults Age ≥ 80, stratified by frontline treatment.

We conducted a subgroup analysis to compare Bruton's tyrosine kinase inhibitors (BTKi) and venetoclax-based regimens on overall survival. Among the 82 patients who received novel agents, 41 received ibrutinib ± anti-CD20 monoclonal antibody, 16 received acalabrutinib ± anti-CD20 monoclonal antibody, 1 received orelabrutinib, and 18 received venetoclax ± anti-CD20 monoclonal antibody. We excluded 4 patients who were treated with venetoclax and BTKi and 2 who received idelalisib ± anti-CD20 monoclonal antibody from our sub-analysis of patients treated with novel agents. The 2- and 5-year OS of patients who received BTKi-based treatment (n = 58) was 83% (95% CI 73%–94%) and 48% (95% CI 26%–89%), respectively (Figure 2). Although there was a suggestion of a longer OS in the acalabrutinib vs. ibrutinib group (HR 6.3, 95% CI 0.8–48.1, p = 0.08), our study was underpowered to detect statistically significant differences in OS estimates between specific BTKi. The 2- and 5-year survival of patients who received venetoclax ± anti-CD20 monoclonal antibody (n = 18) was 100% (95% CI 100%–100%) and 75% (95% CI 43%–100%). All patients treated with venetoclax and anti-CD20 monoclonal antibodies were alive at the time of this study, although their follow-up was shorter compared to other treatment groups. The median OS was 4.8 years (95% CI 3.7–not evaluable) for BTKi ± monoclonal antibody, while not reached for venetoclax (95% CI 4.4–not evaluable), with a HR of 5.51 (95% CI 0.73–41.79, p = 0.099) in univariable analyses.

FIGURE 2

Open in figure viewerPowerPoint

Overall survival comparison between BTKi- and venetoclax-based regimens.

Overall, the median TTNT for the entire cohort was 4.9 years, with 126 patients requiring second-line therapy. The median TTNT was 4.4 years for alkylators, 4.6 years for anti-CD20 monoclonal antibodies, 5.4 years for purine analogues, 4.7 years for novel agents, and not reached for other/unknown (Figure S1). In univariable analysis, the use of non-novel agent-based therapy was significantly associated with shorter TTNT (HR 2.9, 95% CI 1.5–5.7, p = 0.002).

Our study illustrates that historical frontline CLL management in adults ≥ 80 years old primarily included alkylators, purine analogues, and monoclonal antibodies. Treatment with a non-novel agent was associated with an inferior OS. These results propose that novel agents should be considered for frontline therapy in patients ≥ 80 years old. There was a suggestion that venetoclax-based treatment may have a longer OS compared with BTKi. However, this finding should be interpreted with caution due to our small subgroups and potential selection bias, such as frailty or presence of 17p deletion.

The underrepresentation of adults ≥ 80 years in CLL clinical trials over the last 25 years is notable. Studies evaluating adults ≥ 80 years were mostly conducted before the routine use of targeted agents [2, 3]. Simon et al. compared the outcomes of patients with CLL ≥ 80 years old and < 80 years old who were treated with targeted agents in the frontline setting [4]. They included patients with CLL ≥ 80 years old who were treated by the German CLL Study Group (GCLLSG) in 6 clinical trials (n = 95) and routine practice outside of trials through their GCLLSG registry (n = 192) [4]. For trial participants (n = 95), the median OS from the start of frontline treatment was not reached, and the 4-year OS rate was 67.8% in people ≥ 80 years old, compared to 91.7% in participants < 80 years. For patients within the GCLLSG registry (n = 192), the median OS was 35.6 months for patients ≥ 80 years old who were treated with BTKi, whereas the median OS was not reached for patients < 80 years old [4].

Our study is one of few studies that compares survival and TTNT across CLL frontline therapies over 30 years in the U.S. and Europe. Consistent with prior studies, [4-6] we found that novel agents are associated with an improved survival in adults ≥ 80 years old. Participants ≥ 80 years old who received with BTK inhibitors in the six GCLLSG trials (n = 95) had a median OS that was not reached, which is likely due to a sampling bias of trial-eligible patients [4]. The median OS in our study was slightly longer than that of the GCLLSG registry cohort (n = 192) who were treated with BTK inhibitors, [4] but shorter than the median OS of 4.5 years reported in a retrospective study of 79 Italian patients with CLL ≥ 80 years old [5]. For venetoclax-based treatment, our 2-year OS was longer than what was reported in a retrospective study of 110 CLL patients in Italy ≥ 80 years (OS 81%) who received venetoclax as any line of therapy, [6] but our study exclusively reports outcomes of venetoclax in the frontline setting.

The strengths of this study include similar data collection and analysis approaches from Mayo Clinic and Denmark, which provide highly reliable information on patient characteristics and outcomes that are not subject to risks like recall and nonresponse biases. Our study has several limitations. Both cohorts were limited in granular data about any treatment modifications, such as dose reductions or treatment delays, that could impact outcomes. In addition, comparing longitudinal outcomes with variable therapies is challenging; hence, our study must be interpreted with the caveats of any retrospective study.

In conclusion, our study identified and compared OS and TTNT across CLL therapies in adults ≥ 80 years old to close an important knowledge gap. We show that survival after CLL treatment in adults age ≥ 80 has improved with novel agents. This study informs the management of an understudied population of patients with CLL, guides prognostication and shared decision-making, and stresses the importance of recruiting older adults in future clinical trials.