Promoter methylation of the O6-methylguanine-DNA methyltransferase (MGMT) gene plays a role in cellular response to alkylating agents. In the present study aimed to: (i) evaluate the concordance between MGMT promoter methylation status in tumor tissue and plasma; (ii) monitor MGMT promoter methylation status in plasma taken before and during temozolomide treatment; (iii) explore the value of MGMT promoter methylation status in plasma as a prognostic/predictive biomarker in glioma patients. We enrolled 58 patients with histologically confirmed glioma at different grades of malignancy. All patients underwent surgical resection and temozolomide treatment. Paraffin-embedded tumor tissue was available for 48 patients. Blood samples were collected from all patients before temozolomide treatment (baseline) and at each MRI examination for a 12-month period. MGMT promoter methylation status was assessed in both sample types by real time PCR with a specific probe. The frequency of MGMT promoter methylation was 60.4 % in tumor tissue and 41.38 % in plasma. MGMT promoter methylation status was concordant in the two sample types (Kappa = 0.75, 95 % confidence interval (CI) 0.57-0.93; p value <0.001). Overall and progression-free survival were longer in patients with methylated MGMT promoter. Mortality was higher in patients with unmethylated MGMT promoter, whether in tumor tissue [hazard ratio (HR) 2.21; 95 % CI 0.99-4.95] or plasma (HR 2.19; 95 % CI 1.02-4.68). Progression-free survival was shorter in patients with unmethylated MGMT promoter, whether in tissue (HR 2.30; 95 % CI 1.19-4.45) or plasma (HR 1.77; 95 % CI 0.95-3.30). The cumulative incidence of unmethylated MGMT promoter in plasma at baseline was 58 %, and reached virtually 100 % at 12 months. In conclusion MGMT promoter methylation status in tumor tissue and plasma was highly concordant, and both were associated with longer survival, supporting the role of the detection of methylated MGMT promoter in predicting treatment response. However we suggest caution in using plasma as a surrogate of tumor tissue due to possible false-negative results.