Abstract. The aim of the present study was to explore variations of the displacement (D)-loop region in patients with acute myeloid leukemia (AML) and their possible associations with AML pathogenesis. Blood or bone marrow samples from 216 patients with AML (158 AML patients in the first stage, and 58 more patients with AML-M3 for further verification), and 146 healthy controls were collected. Sanger sequencing was performed for the D-loop region ranging between nucleotide (nt)15811 and nt 775. With the exception of mitochondrial microsatellite instability (mtMSI) variations, a total of 2,630 variations in 232 loci were identified with similar variation rates/person in patients with AML and controls when compared with the revised Cambridge reference sequence (8.54±2.14 vs. 8.77±2.15; P=0.366). A positive association between AML and variation-T152C was identified, which occurred more frequently in patients with AML compared with in controls [26.6 vs. 17.1%; P=0.048; odds ratio (OR), 1.752; 95% confidence interval (CI), 1.004-3.058]. Furthermore, T152C was identified to be associated with promyelocytic leukemia-retinoic acid receptor α(PML-RARα) and French-American-British AML subtypes, with a tendency to occur in patients with AML-M3. The AML-M3 sample size was extended by 58 cases, and it was identified that the T152C variation rate was significantly higher in patients with AML-M3 compared with that of controls (41.0 vs. 17.1%; P<0.001; OR, 3.228; 95% CI, 1.714-6.079). However, no association was identified between the T152C variation and clinical characteristics, or chemotherapy response in patients with AML-M3. In addition, the mtMSIs, including D310, mt514-523 (CA) n and T16189C, demonstrated no association with AML risk. Together, the results of the present study suggest that the mitochondrial DNA D-loop region is high variable, and that T152C is associated with AML risk, particularly regarding the M3 subtype. T152C mayparticipate in AML pathogenesis and may be a diagnostic biomarker; however further studies with larger sample sizes are required in order to verify its value.