Geriatric-8 Assessment is Predictive of Hospitalization in Patients with Glioblastoma
Wade Whitt1, Lauryn Hemminger2, Sarah Cawley1, Nimish Mohile1, Andrea Wasilewski3, Sara Hardy1
1University of Rochester, 2Strong Memorial Hospital, 3Givens Brain Tumor Center
Objective:
To describe predictors of adverse outcomes in older patients with glioblastoma (GBM).
Background:
Older adults with GBM have poor survival and are vulnerable to treatment toxicities. Optimal therapy is not defined and decision making depends on identifying risk factors associated with poor
outcomes. Design/Methods:
We prospectively enrolled patients over age 65 with newly-diagnosed GBM to undergo a comprehensive geriatric assessment (CGA) and geriatric-8 (G8) prior to disease-directed therapy. The G8 is a brief, valid and feasible tool that consists of eight items assessing age, food intake, weight loss, mobility, neuropsychology, body mass index, prescription drug, and self-perception of health. CGA consisted of several tests including Short Physical Performance Battery (
SPPB), Geriatric depression scale (GDS-15), and Montreal Cognitive Assessment (MoCA). Patients were followed for adverse events, hospitalizations, and survival. Logistic regression modeled contributions of MGMT methylation status, extent of resection (EOR), G8 score and elements of CGA. Cox regression was also used to model impact of these factors on survival time.
Results:
27 patients were enrolled with median age of 73.5. 20 patients were male and 11 were methylated. 21 patients are deceased (median survival=11.6 months). 44% of patients (n=12) had unplanned hospitalizations. Most common causes were for seizures (n=3), focal weakness (n=5), and cognitive decline (n=3). Poor
G8 score (low) predicted unplanned hospitalizations (OR=3.38, p=0.0386). Age,
EOR, MGMT methylation, GDS-15, and CGA elements were not significantly associated with hospitalization. Only
MoCA score was significantly associated with improved survival (HR=0.7, p=0.02).
Conclusions:
Nearly half of older patients with GBM had unplanned hospitalizations. Poor G8 score predicted hospitalizations, suggesting that G8 may reveal underlying vulnerability to treatment toxicity and should be studied prospectively to validate these findings. Future studies should evaluate if treatment decisions based on G8 can reduce hospitalizations, improve patient quality of life, and decrease healthcare cost.