Glioblastoma is the most common malignant cancer of the central nervous system (CNS) in adults. Glioblastoma cells show increased glucose consumption associated with poor prognosis. Since mitochondria play a crucial role in energy metabolism, mitochondrial mutations and mitochondrial DNA copy number changes may function as biomarkers. As the brain is difficult to access, analysis of mitochondria directly from the brain tissue represents a challenge for neuropathology. Exosome analysis is an alternative (still poorly explored) approach to investigate molecular changes in CNS tumors. Here we analyze characteristics of brain tissue DNA and plasma-derived exosomal DNA (exoDNA) of 44 glioblastoma patients and 40 control individuals. Quantitative real-time PCR was performed to determine mtDNA copy numbers and the Kruskal-Wallis and Mann-Whitney U test were used for statistical analysis of data. Subsequently, sequencing libraries were prepared and sequenced on the MiSeq platform to identify mtDNA point mutations. Tissue mtDNA copy number was different among controls and patients in multiple comparisons. A similar tendency was detected in exosomes. Based on NGS analysis, several mtDNA point mutations showed slightly different frequencies between cases and controls, but the clinical relevance of these observations is difficult to assess and likely less than that of overall mtDNA copy number changes. Allele frequencies of variants were used to determine the level of heteroplasmy (found to be higher in exo-mtDNA of control individuals). Despite the suggested potential, the use of such a biomarker for the screening and/or diagnosis of glioblastomas is still limited, thus further study will be required.