Abstract
Introduction: The majority of patients with chronic lymphocytic leukemia (CLL) present with multiple comorbidities at diagnosis or when treatment is needed. However, impact of comorbidities on CLL treatment outcomes remains understudied, as such patients are typically excluded from clinical trials. While multiple regimens have been approved for treatment of CLL, only few have been evaluated in this patient population. Furthermore, assessment of comorbidities in CLL has not been standardized. Several recent studies have assessed Cumulative Illness Rating Scale (CIRS) score at enrollment. Thus, in this retrospective analysis, we evaluate CIRS score and analyze whether it is predictive of treatment outcomes in CLL.
Methods: We performed a retrospective analysis of CLL patients treated at Dartmouth between 2000 and 2013. CIRS score was calculated as in Salvi et al, 2008. Overall survival (OS) and progression-free survival (PFS) were assessed by using Cox proportional hazards models adjusting for performance status, age group, and chemotherapy treatments. In addition, the impact of severe organ dysfunction (CIRS 3-4 in ≥1 organ system, referred to as "CIRS-3+") was assessed.
Results:133 patients received a total of 255 treatment regimens. Median age at the start of therapy was 65. Fludarabine (Flu)-Rituximab (FR, N=52), Fludarabine-Cyclophosphamide-Rituximab (FCR, N=30), Rituximab-Cyclophosphamide-Vincristine-Prednisone (R-CVP, N=35), Bendamustine-Rituximab (BR, N=30) and chlorambucil (Chl; N=25) were the most commonly used treatment regimens. 65.6% of treatments were administered in a frontline setting. Specifically, 77%, 73%, 63%, 33%, and 76% of patients received the above regimens as frontline therapy, correspondingly. Median OS among all patients was 126.1±12.2 months. 53% of patients had CIRS≥7 and 57% had CIRS-3+. Their respective median OS were 106.1±12.2 and 108.5±3.7 months (p<0.05 compared to patients with CIRS<7, at 128.2±1.4 months).
In this retrospective analysis, CIRS score did not predict the choice of therapy. Age over 70 was associated with a decreased likelihood of receiving Flu (OR=0.2, p=0.017), while Chl was used more frequently in this patient population (OR=45.26, p<0.001).
Multivariate analysis of a total CIRS score, age and performance status (PS) demonstrated that OS after CLL therapy shortened with each increase in total CIRS score by one point (HR=1.12; p=0.02). Multivariate analysis of CIRS-3+, PS, and age demonstrated that OS after therapy worsened with an increase in number of organ systems affected by a severe (grade 3-4) comorbidity (HR=1.58, p=0.04). Patients>70 year old had worse OS in multivariate analysis of CIRS≥7, CIRS-3+, age and PS. Interestingly, higher total CIRS, presence of a severe organ dysfunction (CIRS-3+) as well as CIRS≥7 were all significant predictors of decreased PFS.
Among therapies used, Flu and cyclophosphamide-based regimens were associated with decreased OS in patients with CIRS≥7 (HR=6.5, p=0.01 and HR=10.4, p=0.001). Shortened OS following Flu treatment was observed among patients in 70-79 age group (HR=19.22, p=0.005), and in patients older than 80 years (HR=47.94; p<0.001). Similarly, shortened OS after Flu therapy was also found in patients who were older than 70 y.o and had either a higher total CIRS score (HR=4.73,p=0.025) or severe organ dysfunction (CIRS3+; HR=5.4, p=0.015). Interestingly, multivariate analysis including age, PS and CIRS, treatment with cyclophosphamide-based regimens (excluding FCR) was also predictive of shortened OS after therapy in patients with CIRS≥7 (HR=10.4, p=0.001), CIRS-3+ (HR=7.6, p=0.001) and total CIRS (HR=8.2, p=0.001). Notably, multivariate analysis with Chl- based regimens demonstrated improved OS for patients with CIRS≥7 (HR=0.175, p=0.019), CIRS3+ (HR=0.125, p=0.005) and overall high total CIRS (HR=0.142, p=0.006)
Conclusion: In summary, CIRS is a valuable prognostic indicator that can be used to estimate OS and PFS in patients with CLL who have comorbidities, both alone and in combination with patient's age. Our data also suggest that patients above 70 years with high CIRS score may be poor Flu candidates, and tend to do better with Chl- based regimens. However, prospective studies are needed to validate CIRS effect on overall survival and to optimize treatment strategies in this high-risk population, including in a setting of novel "targeted" therapies.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.