The Chronic Lymphocytic Leukemia Comorbidity Index (CLL-CI) Predicts Survival and Tolerance of Ibrutinib Therapy in Patients with CLL: A Multicenter Retrospective Cohort Study

Introduction: Medical comorbidities influence survival in CLL. We previously reported on a simplified CLL-specific comorbidity scale, the CLL-CI (Gordon et al. 2019), which required assessment of only three organ systems and was predictive of outcomes in a heterogeneous retrospective patient cohort. Herein we analyzed CLL-CI among patients treated with ibrutinib.

Methods: This retrospective study included 339 CLL patients treated with ibrutinib at 9 academic centers between 2014-2019. Vascular, endocrine and upper-gastrointestinal organ systems were assessed at the time of ibrutinib initiation. Each was scored from 0 to 3, in order of increasing severity of dysfunction to generate the CLL-CI score (range, 0-9; Figure A). As established previously, CLL-CI≥2 was deemed high-risk. Event free survival (EFS) was measured from start of ibrutinib to development of new CLL-related symptoms, disease progression, start of a new therapy or death. Overall survival (OS) was measured from treatment initiation to death. Patients with no EFS or OS events were censored at last follow up. The Kaplan-Meier method and log-rank test were used to estimate and compare survival. Multivariable Cox regression was utilized to model EFS and OS. Differences between CLL-CI groups were evaluated with Wilcoxon rank sum and Fisher’s exact tests.

Results: Median age was 68 years (range, 30-91), 240 (71%) were treated in the relapsed/refractory setting (range of prior therapies, 0-10). Advanced Rai stage (3-4) was present in 206 (61%) and TP53 aberrancy was present in 93 (27%) patients at the start of ibrutinib therapy. Median follow up was 23 months (range, 1-71).

CLL-CI score was high (≥2) in 202/339 patients (60%). Patient characteristics were well balanced between subgroups (CLL-CI <2 vs ≥2). The median number of prior therapies was 1 in both groups. The distribution of Rai stage, TP53 aberrancy and IGHV mutational status (available in 48% of patients), were similar.

In multivariate models adjusted for age, del(17p) and relapsed disease, high CLL-CI was associated with shortened EFS (HR=1.65; p=0.014, Figure ) and OS (HR=1.73; p=0.1). CIRS score≥7 also correlated with EFS (HR=1.91; p=0.002) and OS (HR=2.78; p=0.006). Ibrutinib discontinuation rates due to adverse events were more frequent in patients with CLL-CI ≥2 (25% vs 14%; p=0.014). However, dose reduction rates were similar (24% vs 20%; p=0.51).

Fifty-two deaths occurred: 40 in the high CLL-CI subgroup and 12 in the low CLL-CI subgroup. Cause of death was known in 31 patients. Death due to disease progression was more frequent in the high CLL-CI subgroup (28% vs 8%; p<0.001). Infection was the second most common cause of death and occurred at a similar rate in both groups (13% vs 16%; p=0.66). Two deaths occurred due to cardiac causes, both in the high CLL-CI group.

Since some of the key ibrutinib toxicities (atrial fibrillation, hypertension) may not have been accounted for in the CLL-CI we further elucidated their possible impact. Cardiac disease was significantly more prevalent among patients with high CLL-CI (37% vs 16%, p<0.001; severe - 13% vs 5%; p=0.013) as was hypertension (63% vs 35%; p<0.001; severe - 27% vs 13%; p=0.004). When computing pairwise correlation coefficients, we found that vascular and endocrine comorbidities significantly correlated with both cardiac disease and hypertension (Spearman ρ = 0.15 to 0.27). However, re-computing CLL-CI score upon inclusion of those categories did not influence the EFS concordance probability (CLL-CI, 0.669 vs CLL-CI plus cardiac and hypertension, 0.664).

Conclusions: Here we present the largest cohort of CLL patients treated with ibrutinib in whom comorbidities have been systematically assessed. We find that the CLL-CI (which assesses endocrine, vascular and upper gastrointestinal conditions) correlates with survival and tolerance of therapy in this population. Unexpectedly, we found that hypertension and cardiac comorbidities did not improve CLL-CI’s discriminatory power. This result combined with the simplicity of scoring the CLL-CI makes it an attractive tool for clinical practice. CLL-CI needs to be explored prospectively in patients treated with ibrutinib and other targeted therapies.