The promise of urinary extracellular vesicles (uEVs) in biomarker discovery is emerging. However, the characteristics and compositions of different uEV subpopulations across normal physiological and pathological states require rigorous explication. We recently reported proteomic signatures of small (s)‐uEVs (<200 nm membranous nanoparticles) and described putative biomarkers corresponding to the diagnosis, tumour burden and recurrence of the lethal adult primary brain tumour, glioblastoma. Here, we comprehensively characterise uEV populations with significantly different mean and mode particle sizes obtained by differential centrifugation at 100,000 × g (100K‐uEVs; smaller) and 17,000 × g (17K‐uEVs; larger) using Fourier‐transform infrared spectroscopy and quantitative data‐independent acquisition mass spectrometry. We show distinct differences in protein and lipid content, prominent protein secondary structures, and proteome distributions between uEV populations that can distinguish glioblastoma patients from healthy controls and correspond to clinically relevant tumour changes (i.e., recurrence and treatment resistance). Among the key findings is a putative seven‐protein biomarker panel associated with 17K‐uEVs that could distinguish all glioblastoma patients from healthy controls and accurately classify 98.2% of glioblastoma samples. These novel, significant findings demonstrate that both uEV populations offer individual and combined biomarker potential. Further research is warranted to elucidate the complete diagnostic, prognostic, and predictive capabilities of often‐neglected 17K‐uEV populations.