The role of geriatric assessment and frailty measurements in predicting surgical risk and survival in elderly patients with colorectal cancer : A prospective observational cohort study

Abstract

Background: Colorectal cancer (CRC) can be considered a disease of the elderly, with a median age of diagnosis of 72 years. Surgery is the main treatment for colorectal cancer. As chronological age does not accurately reflect physiological reserves in the heterogeneous elderly population, it has been suggested that older cancer patients may benefit from a comprehensive geriatric assessment (CGA) before treatment decisions are made. A CGA is a systematic approach aiming to assess physical function, comorbidity, polypharmacy, nutritional status, cognitive function, and emotional status in older patients. Based on a CGA, patients may be divided into three groups: Fit, intermediate, or frail. Few prospective studies have investigated the associations between elements of a CGA and surgical outcomes in elderly patients. Furthermore, the definition of “frailty” derived from a CGA is controversial. In geriatric medicine, frailty is more commonly defined as a cluster of physical impairments (often called the physical phenotype of frailty – PF).

Aims: To study the association between a categorization of patients into the groups fit, intermediate, or frail based on a pre-operative CGA and the risk of post-operative complications in elderly CRC patients who were electively operated; to identify independent predictors of post-operative complications and survival from a CGA and Eastern Cooperative Oncology Group performance status (ECOG PS); to compare a pre-operative multi-domain frailty measurement based on a CGA to a modified version of PF in older CRC patients, and to analyze the ability of the two measurements to predict post-operative complications and overall survival; to compare levels of inflammatory biomarkers (CRP, IL-6, TNF-α), and Ddimer in older CRC patients classified according to the two frailty definitions.

Methods: Patients ≥ 70 years electively operated for all stages of CRC from 2006 to 2008 in three Norwegian hospitals (Ullevaal University Hospital, Aker University Hospital, and Akershus University Hospital) were consecutively included. A pre-operative CGA, an assessment of self-reported health, measurements of grip strength and gait speed were performed, and blood samples were drawn within 14 days of surgery. CGA-frailty was defined as fulfilling one or more of the following criteria: Dependency in personal activities of daily living, severe comorbidity, cognitive dysfunction, depression, malnutrition, or >7 daily medications. PF was defined as the presence of three or more of the following criteria: Unintentional weight loss, exhaustion, low physical activity, impaired grip strength, and slow gait speed. Outcome measures were post-operative complications (any complication and severe complications) and overall survival.

Results: Patients (182) with a median age of 80 years (range, 70-94 years) were included. For the categorization into the three CGA-groups, 178 patients were available for analyses, while 176 patients were available for the comparison between the two frailty classifications. Twenty-one patients (12%) patients were categorized as fit, 81 (46%) as intermediate, and 76 (43%) as frail. Eighty-three patients experienced severe complications, including three deaths; 7/21 (33%) of fit patients, 29/81 (36%) of intermediate patients, and 47/76 (62%) of frail patients (p=0.002). Increasing age and ASA class were not associated with complications. Severe comorbidity was an independent predictor of severe complications (odds ratio [OR] 5.62; 95% CI 2.18 to 14.50) and early mortality (hazard ratio [HR] 2.78; 95% CI 1.50 to 5.17). Dependency in instrumental activities of daily living (IADL) and depression were predictors of any complication (OR 4.02; 95% CI 1.24 to 13.09 and OR 3.68; 95% CI 0.96 to 14.08, respectively) while impaired nutrition predicted early mortality (HR 2.39, 95% CI 1.24 to 4.61). When added to the models, ECOG PS independently predicted both morbidity and early mortality, and ECOG PS was a more powerful predictor than IADL-dependency, depression, and impaired nutrition. The agreement between the two frailty classifications was poor. CGA-frailty was identified in 75 (43%) patients, while PF was identified in 22 (13%) patients. Only CGA-frailty predicted post-operative complications (p= 0.001). Both CGAfrailty and PF predicted survival. Levels of CRP and IL-6 were significantly higher in frail compared with non-frail patients within both measures.

Conclusions: CGA can identify frail patients who have a significantly increased risk of developing post-operative complications after elective surgery for CRC. This multi-domain frailty measurement appears to be more useful than frailty identified from a modified version of the physical phenotype of frailty criteria in predicting morbidity, but for long-term outcomes such as overall survival, both measurements are predictive. Severe comorbidity, IADL-dependency, depression, and impaired nutrition seem to be the most important CGAelements predictive of post-operative complications and overall survival. As ECOG PS predicts all outcomes, a consistent use of ECOG PS in studies of cancer surgery is recommended.