Per- and polyfluoroalkyl substances (PFAS) were components of firefighting foams used by the U.S. military for decades. Perfluorooctanoic acid (PFOA) has been identified as a carcinogen, and perfluorooctanesulfonic acid (PFOS) as a possible carcinogen. Additionally, some PFAS can cross the blood-brain barrier, though effects on the brain are not well studied. Therefore, we evaluated associations between serum PFAS levels, including PFOA, PFOS, and perfluorohexanesulfonic acid (PFHxS), and glioma risk in U.S. military service members. This nested study included 457 case-control pairs with pre-diagnostic serum samples, ascertained from the Department of Defense (DoD) Automated Central Tumor Registry, Defense Medical Surveillance System, and the DoD Serum Repository. Cases had a diagnosis of a first primary brain glioma (International Classification of Diseases for Oncology, 3rd edition codes 938-948) or a first primary brain cancer (International Classification of Diseases, 9th revision code 191) supported by relevant diagnostic procedure codes (Current Procedural Terminology codes for nervous system surgery: 61000-64999, cytopathology: 88104-88199, cytogenetics: 88230-88299, or surgical pathology: 88300-88399). Pairs were individually-matched on sex, race/ethnicity, birth year, number of pre-diagnostic serum samples (1 - 3), and the date(s) of serum collection. High-resolution metabolomics analysis was performed on the serum by the Clinical Biomarkers Laboratory at Emory University using dual-column liquid chromatography and electrospray ionization. Features were extracted from the data, with Combat batch correction. PFAS concentrations (in ng/mL) were based on reference standardization. All statistical analyses were conducted across four time periods: 10 - <20 (T4), 5 - <10 (T3), 2 - <5 (T2), and <2 (T1) years pre-diagnosis. For subjects with multiple samples in a period, the sample closest to diagnosis was included. We conducted t-tests to compare levels between cases and controls and conditional linear regression (CLR) to evaluate associations between PFAS levels and glioma risk. For CLR, PFAS levels were evaluated as tertiles, with below the detection limit as the reference group for PFHxS. The highest PFAS levels were for PFOS, followed by PFHxS and PFOA; there were not statistically significant differences in PFAS levels between cases and controls at any period. There was evidence of an association between PFHxS and increased glioma risk across all time periods: T4 (Odds Ratio (OR)Tertile 3 vs ≤LOQ = 2.22; 95% Confidence Interval (CI), 0.91 – 5.42; p-trend=0.08), T3 (ORTertile 3 vs ≤LOQ, 95%CI = 2.06, 0.95 – 4.44; p-trend=0.04), T2 (ORTertile 3 vs ≤LOQ, 95%CI = 1.76, 0.87 – 3.56; p-trend=0.07), and T1 (ORTertile 3 vs ≤LOQ, 95%CI = 2.59, 1.21 – 5.54; p-trend=<0.01). Results for PFOA and PFOS were largely null, without clear trends by increasing serum levels or time period. There is potential impact for PFAS accumulation in the brain due to its lipophilic nature. These findings provide evidence of positive associations between serum PFHxS and glioma risk up to twenty years prior to diagnosis within this military population. These consistent findings across time periods for PFHxS only may be in part due to the environmental and biological persistence of PFHxS. Further study is needed to confirm these findings and to understand mechanisms of action.