Impact of Comorbidity Burden in Treatment and Outcome of Older Patients with Indolent Non-Hodgkin Lymphomas

Introduction: Despite the growing interest about the optimal treatment of older patients with cancer, only few data are available for indolent non-Hodgkin lymphomas (NHL). Aim of the study was to evaluate the impact of comorbidities on treatment choice and survival in a cohort of older indolent NHL patients.

Material and Methods: In an IRB-approved protocol, we reviewed the records of 742 patients with age≥60 years diagnosed of indolent NHL between January 1990 to December 2012 at our Center. Patients not receiving a treatment (N=177) and those treated with local therapy for stage I disease (N=138) were excluded from the analysis. Clinical information was gathered from the electronic charts, comorbidity was assessed using the cumulative illness rating scale-geriatrics (CIRS-G). Association between categorical variables was assessed with chi-squared test or Fisher's exact test. Overall survival (OS) and progression free survival (PFS) were the end-point of the study, defined respectively as the time from diagnosis to death or last visit and time to death or relapse/progression, which one occurred first, or last visit. Differences among groups in terms of OS and PFS were verified with both log-rank test in univariate setting and Cox proportional hazard model in multivariable analysis. 

Results: A total of 427 patients with a median follow-up of 4.04 years (range 0.1-20) were evaluated in the study. Patients were diagnosed as follicular NHL (n=207), marginal zone NHL (MALT N=93, splenic N=32, nodal N=38) and lymphocytic NHL (N=57). Median age was 69.9 years (range 60-93.9) with 11% of the patients older than 80 years, M/F ratio was 0.81. At least one comorbidity were present in 363 patients (85%), CIRS total score <6 was present in 321 (75%) and a severity index>2 was recorded in 161 patients (38%). At least one score 3-4 comorbidity was present in 176 (41%) and the most frequent categories involved were vascular (N=191, 45%), metabolic/endocrine/breast (N=102, 24%) and heart (N=71, 17%). Patients were further categorized in fit (N=224, 52%) or unfit/frail (N=203, 48%) on the basis of age (≥80 years) and CIRS parameters (>1 grade 3-4 or >4 grade 2 categories). Patients categorized according these comorbidity scores were balanced for main clinical factors, with the exception of age. First line treatment was CHOP/CHOP-like (N=179, 42%), CVP (N=112, 26%), chlorambucil (N=104, 24%) or other (N=32, 7%) with the addition of rituximab in 53% of the cases. Severity index>2 (P=0.0146), presence of >1 score 3-4 category (P=0.0167) and frail/unfit status (P<0.0001) were negatively associated with rituximab use, while CIRS score<6 (P=0.0243) and frail/unfit status (P=0.0109) significantly influenced the choice of chemotherapy. Early death (<6 months from diagnosis) was not associated to comorbidities. Several multivariate models were built to analyze the effect of comorbidity scores on OS and PFS, controlling for age, stage, rituximab use and chemotherapy type. Comorbidity considered as CIRS score<6 (P=0.0091), Severity index>2 (P=0.0310), presence of >1 score 3-4 category (P=0.0239) and frail/unfit status (P<0.0001) independently influenced PFS as well as age, clinical advanced stage and rituximab use (Figure 1A). Overall survival was deeply influenced by age<80 (P=0.000), rituximab use (P=0.000) and frail/unfit status (P<0.0001, Figure 1B).

Conclusions: Age and comorbidity scores influenced treatment intensity as well as the decision to add rituximab, this was associated to a significant impact on overall outcome of indolent NHL.

Figure 1