Glioblastoma is a fatal and devastating primary brain tumour that poses significant challenges to accurate tumour monitoring and patient treatment. Substantial efforts are underway to develop accurate non-invasive liquid biopsy methods that can routinely assess glioblastoma markers in patient body fluids. We have recently shown that patient urine offers ready access to glioblastoma biomarkers within small extracellular vesicle populations (EVs; <200 nm membrane particles) and can accurately diagnose glioblastoma and reflect changes in glioblastoma pathophysiology. However, little is known about other urinary-EV populations in glioblastoma, and their capacity as biomarkers. In this study, we compared the biochemical properties of small (< 200 nm) and large (200-1000 nm) urinary-EVs using fourier-transform infrared spectroscopy and data independent acquisition mass spectrometry and identified significantly changing molecular distributions across the two EV populations in glioblastoma states. We also determined putative glioblastoma biomarkers associated with large urinary-EV biomarkers by ranking significant proteins with a logistic regression model. Our analysis revealed seven proteins capable of perfectly distinguishing all glioblastoma patients from healthy controls (AUC=1) and accurately classifying 96.7% of all cases as glioblastomas. Many significant large urinary-EV proteins also displayed consistent trends at different glioblastoma clinical timepoints, with protein abundance shifts post-surgery that return to pre-surgical levels at recurrence. Our findings demonstrate that both EV populations facilitate easy access to glioblastoma biomarkers for clinical diagnostics. Putative urinary-EV biomarkers described here merit further validation using comprehensive sets of longitudinal cohorts of glioblastoma urine.