The Vulnerable Elders Survey‐13 Predicts Mortality in Older Adults with Later‐Stage Colorectal Cancer Receiving Chemotherapy: A Prospective Pilot Study

To the Editor: Colorectal cancer (CRC) is the second most common cause of cancer death in the Western world. Almost half of all CRC cases occur in individuals aged 70 and older,1 but the data regarding treatment are largely focused on individuals younger than 65, even though studies suggest that older adults with CRC derive benefit from adjuvant chemotherapy equivalent or nearly equivalent to that of younger individuals.2, 3 Older adults form a heterogeneous group, and physiological fitness is weakly correlated with chronological age; criteria other than age should help form the basis for treatment decisions. One solution is to derive these criteria using the tools of a comprehensive geriatric assessment (CGA).4 A prospective cohort study was conducted to evaluate baseline CGA for older adults with Stage III and IV CRC undergoing first-line chemotherapy. It was hypothesized that the presence of impairments on the CGA would be associated with poorer survival. Eligible individuals were recruited from the University of Chicago Gastrointestinal Medical Oncology Clinic between February 2006 and February 2009. Protocol approval was obtained from the institutional review board, and all participants provided written informed consent. Participants underwent assessments just before their first dose of chemotherapy. Survival data were obtained from chart review and the Social Security Death Index. The comprehensive evaluation included the Vulnerable Elders Survey (VES-13),5 a 13-item self-administered survey designed to screen for overall functional status, and assessments of cognition, emotional affect, social functioning, nutritional status, activities of daily living (ADLs), comorbidities, and physical performance. Bivariate analysis for overall survival used Cox proportional hazards models and Kaplan-Meier curves, and multivariate survival analysis was performed using logistic regression. The initial model was obtained by including Eastern Co-operative Oncology Group Performance Status (ECOG-PS), ADL, and VES-13 scores, all of which were significant (P ≤ .05) in the bivariate analysis, plus age. The models were limited to four variables to avoid overfitting, given the small data set. All statistical calculations were performed using Stata SE, version 11 (StataCorp LP, College Station, TX). Thirty-eight patients were enrolled (median age 72, range 65–89; 63% male). Seventy-nine percent had metastatic disease at presentation. Of the 33 for whom baseline ECOG-PS was assessed, 30 (91%) had a PS of 0 or 1 as assessed according to the treating physician. Bivariate analysis revealed several measures that were associated with poorer overall survival in this cohort. For the entire cohort, VES-13 of 3 or greater (hazard ratio (HR) = 5.34, P = .002), ECOG PS of 1 or greater (HR = 2.4, P = .05), ADL dependence (HR = 5.62, P = .005), and Geriatric Depression Scale (GDS) score of 5 or greater (HR = 3.95, P = .04) were correlated with poorer survival. For individuals with metastatic disease, only VES-13 of 3 of greater (HR = 4.71, P = .005) and ADL dependence (HR = 6.19, P = .01) were prognostic. Using a multivariate regression model, only abnormal VES-13 score remained significant (HR = 15.61, P = .02, Figure 1). Age was not prognostic in the bivariate (HR = 1.02, P = .56 for ≥75 vs <75) or multivariate model. Relative dose intensity of chemotherapy of greater than 85% was also not predictive of overall survival (HR = 0.61, 95% confidence interval = 0.25–1.41). Oncologists are increasingly faced with treatment decisions for older adults, who are much less likely to receive even well-established therapies.6, 7 This may derive from lack of knowledge about trials that included older adults or concern about the generalizability of trial results from younger cohorts being applied to older adults. Tools such as the VES-13 that formally and prospectively assess vulnerability may help provide important prognostic information. In the current study, baseline VES-13 scores emerged as the best prognostic tool. Competing causes of death are important to consider in this age group, and VES-13 should be incorporated into risk assessments. It adds prognostic information to ECOG-PS, a commonly used assessment of performance status in oncology practice typically used to determine fitness for chemotherapy. The cohort was small and incorporated individuals receiving chemotherapy in addition to surgery and those with metastatic disease, reducing the generalizability of the conclusions that may be drawn. Participants were also recruited from the referral base of a single academic center, and the results therefore may not be generalizable to other populations with CRC. Nevertheless, this study indicates that assessments for frailty in elderly adults with CRC, if further validated, could be incorporated into clinical practice. It highlights, in particular, the usefulness of the VES-13 in risk-stratifying individuals with cancer. This tool is strongly correlated with frailty as measured using a full CGA in older adults with cancer.8, 9 Nonphysician personnel can easily collect it, and it takes only 3 to 5 minutes to complete, and may therefore be easily incorporated into busy oncology practices. CGA measures such as the VES-13 screening instrument should be included in multicenter trials for further validation. Presented at the American Society of Clinical Oncology Annual Meeting, June 2011, Chicago, Illinois, and at the American Geriatrics Society Annual Meeting, May 2012, Seattle, Washington. Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper. Funded by a John A. Hartford Center of Excellence grant, University of Chicago Medical Center. Author Contributions: Study concept and design: BP, JH, KB, HK, SM, WD. Acquisition of participants and data: ER, BP, KB, HK, SM. Analysis and interpretation of data: ER, BP, JH, SM, WD. Preparation of manuscript: ER, BP, JH, KB, HK, SM, WD. Sponsor's Role: None.