Impact of Comorbidity on Initial Treatment and Overall Survival in Elderly Head and Neck Cancer Patients

Discussion

Concurrent comorbid conditions in patients with cancer have been shown to affect disease course, treatment selection, and overall survival (4). Moreover, numerous comorbidity indices have been developed to measure and evaluate the presence of comorbidity in patients with various malignancies and its effects on patient survival. Hall et al. demonstrated that patients with HNC were at an increased risk of dying from comorbid diseases associated with HNC compared to the general population and that after 7.5 years this risk of death from comorbid diseases surpassed the same risk from HNC itself (11). Singh et al. revealed comorbidity to be a prognostic factor in a small study (n=88) of HNC patients (6). Similarly, Reid et al. conducted a larger study of 9,386 elderly (>65 years of age) patients with HNC and found comorbidity to be an “independent, clinically relevant determinant of survival” (12). More recently, a Danish study of 12,623 HNC patients with a median age of 62 years offered further evidence that comorbidity had a significant independent effect on overall survival but not on cancer specific death (suggesting patients died from comorbidity rather than their cancer) (13). All three of these latter studies utilized the CCI as a marker of comorbidity.

Several studies have also been conducted to examine the relationship between comorbidities and treatment selection in HNC patients. The standard treatment of HNC involves a combination of surgery, radiotherapy and/or chemotherapy. However, the feasibility of these options is often questioned in the elderly cancer patient population. The apparent disparities of treatment selection for and clinical trial involvement of elderly patients are demonstrated throughout almost every malignancy (14-16). The fears among trial investigators and clinicians are that elderly patients will not tolerate aggressive measures, suffer increased toxicities and complications, and have resultant poor qualities of life – consequences not solely due to patient age but comorbidity as well. Derks et al. wrote that severe comorbidity, more so than age alone, impacted on treatment selection and post-operative complication rates in elderly HNC patients (17). Similarly Bernardi et al. found that underlying severe comorbidities, age-related frailty and psychosocial impairments were hindrances to intensive treatment plans (18). Ferrier et al. concluded that underlying comorbidity was a major predictor of complications in head and neck surgeries (19). Finally, Hathaway et al. demonstrated in a re-evaluation of a prior study on adjuvant chemoradiation in squamous HNC that concurrent comorbidity was one of the reasons patients declined adjuvant therapy (the study (20) went on to conclude that adjuvant chemoradiation conferred a survival advantage when offered in conjunction with surgery) – reiterating the prognostic importance of comorbidity (21). Despite these studies assessing the overall impact of comorbidity on treatment selection and outcome, there has yet to be a study that assesses the predictive value of a specific index on selecting patients for initial treatment. Our study chose to test the distinct hypothesis that the CCI could be predictive of initial treatment selection and outcomes in elderly patients with HNC.

Our study, in congruence with prior similar studies, confirmed that comorbidity in itself is a predictor of overall survival. We demonstrated that patients with low comorbidity status (CCI score 0-2) had longer survival than patients with high comorbidity status (CCI >2). However, in multivariate analysis, only clinical stage was found to be a significant independent predictor of overall survival. In that analysis, CCI score >2 was not a significant predictor for a worse overall survival in the entire cohort or in the subgroups that received curative intent surgery or chemoradiation.

The strengths of this study include having one centralized electronic medical record, which allowed for the tracking of every patient over the 7-year time period. This aided in acquiring survival data on virtually every patient (even if a patient left DVAMC and was treated at a different VAMC). In addition, the fact that patients with a CCI >2 did have a worse survival in the univariate analysis suggests that the CCI score was being applied accurately, albeit retrospectively, in this population. Moreover, as HNC is increasing in prevalence in patients >65 years of age, our study's target population was appropriate for clinical practice application.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Kaplan-Meier analysis of patients with CCI 0-2 and CCI >2.

Our study had certain limitations as well. The population included was entirely male. The retrospective nature of the study limited us to the patient profile, as documented in the CPRS, as well as to the treatment plan selected for reasons deemed appropriate at the time of assessment. Our study also did not take into consideration the performance statuses of the patients. Such a qualifier might have had an effect on the patients' treatment selections and overall outcomes. Lastly, the number of participants included in the study is modest in comparison to some of the other comorbidity trials.