Abstract Background Gemistocytes (GCs) are unique phenotype of reactive astrocytes. In diffuse low-grade gliomas, they are associated with grim prognosis. Less is known regarding their pathogenesis and prognostic significance in glioblastoma (GBM). Previous data on-neoplastic central nervous system (CNS) lesions has linked the presence of GCs with states of uncontrolled hyperglycaemia in diabetes mellitus type 2 patients. study purpose is to find a correlation between systemic metabolic features and the presence of GCs in patients with GBM.Also, to try and correlate between the presence of GCs, molecular changes, and use of anti-diabetic drugs on overall survival of GBM patients. Methods Electronic medical records from newly diagnosed GBM patients were retrospectively reviewed and extracted for demographic, clinical, metabolic, radiological and pathological variables with emphasis on different metabolic - related features. A statistic-based comparison was made between GBM patients with and without GCs for the same variables. We compared DM2 GC-GBM patients to DM2 non-GC GBM patients in a cohort of poorly controlled DM2 (i.e., HBA1C ≥ 8.0). A possible influence of metformin administration on OS of DM2 GBM patients, with and without GCs, and with and without TP53 mutations was investigated. Results A total of 220 patients with newly diagnosed GBM were included in the study. 14.5% were defined as GC-GBM (group 1, n=32) and 85.5% were defined as non-GC GBM (group 2, n=188). The incidence of DM2 was 25% in both groups. The rate of poorly-controlled DM2 was nearly as twice in group 1 than in group 2 (18.75% vs. 9.5%; p=0.130). In the DM2 GBM group, significant differences were found between DM2 GC-GBM (group 1a, n=9) and DM2 non-GC GBM (group 2a, n=49) patients with male gender predominance (89% vs. 50%, p=0.073) and obesity. Patients in group 1a were more likely to have weight ≥85kg (OR 6.16; 95% CI 1.3336 to 28.5147, p=0.0019) and had higher BMI values (mean BMI of 34.1±11.42 vs. 28.7±5.44 for group 1a and 2a respectively; OR= 1.095, 95%CI 1.053-1.207;p=0.034). Similar significant findings were in the poorly-controlled DM2 subgroup. None of the poorly-controlled DM2 GC-GBM patients were using insulin prior to diagnosis (0% vs. 61.10%, p=0.016). Survival analyses have shown trends for differences in overall survival between subgroups in correlation with presence of GCs, p53 mutations and use of metformin. Conclusion GBM patients with systemic metabolic factors that are associated with marked insulin resistance and are not using insulin at time of diagnosis are significantly more likely to have GCs in their initial pathology. This is a novel finding that may add to data on glucose metabolism in astrocytes and in astrocytes-associated tumors. The presence of GCs, molecular changes in the tumor, and under the influence of anti-diabetic drugs such as metformin, may affect the prognosis of different GBM subgroups.