Background: Aldehyde dehydrogenase-1A1 (ALDH1A1) mediates therapy resistance and poor prognosis in multiple cancers. In GBM, ALDH1A1 expression above vs. below the mean is associated with a significant reduction in survival.¹ ALDH enzymes activate transcriptional programs that induce the cancer stem cell phenotype, detoxify anticancer drugs, and function as antioxidants by reversing lipid peroxidation and etheno-DNA adducts, thus conferring chemotherapy and radiation resistance. PPI have been shown to increase enzyme activity of ALDH1A1 up to 6-fold, suggesting a potentially hazardous effect on tumor biology and survival. However, no prospective randomized data are available to address these concerns. Thus, we turned to XCELSIOR, an observational database that prospectively collects medical records including medication usage. Methods: Patients with GBM were e-consented to XCELSIOR (NCT03793088), a nationwide observational research protocol that leverages patient consent to permit gathering of medical records from all sites of care. Clinical data were abstracted from medical documents into a central EDC system, source-verified, and standardized to generate comprehensive longitudinal histories for aggregate analysis. Identified patient records permitted the definitive identification of dates of diagnosis and death. Overall survival (OS) was calculated and stratified based for PPI exposure, extent of surgical resection (EOR), and MGMT methylation status. Results: XCELSIOR contained 320 GBM patients from 222 treatment sites. We stratified patients based on exposure to PPI and assessed median OS (mOS). For the entire population, mOS was substantially better for patients not receiving PPI. (TABLE) The 5-yr survival was 29.9% [21.0 – 42] and 6.6% [1.4 – 32.0] for patients not receiving vs. receiving PPI. In a multivariate analysis considering EOR and MGMT methylation, PPI exposure yielded a hazard ratio (HR) of 1.57 (p=0.009). For the interaction of m-MGMT and PPI exposure together, the HR is 2.63 (p=0.043), significant as the HR of m-MGMT alone is 0.40 (p=0.0028). Conclusions: In this RWE dataset, PPI use in GBM appeared detrimental to survival. This is strongly consistent with the hypothesis that PPI antagonizes chemotherapy. Additionally, use of PPI appeared to completely abrogate the survival benefit of chemotherapy associated with methylated-MGMT. While we recognize the utility of PPI in managing acute UGI bleeding, these data suggest avoiding PPIs in GBM patients. 1) Schmidt-Graf F, et al. Neuro Oncol. 2012;14(12):1452–1464.

*GTR: gross total resection, NTR: near-total resection.