OBJECTIVES: Failed implant removal using a high-frequency electrosurgical device (HFED) has been reported to be less invasive than other surgical techniques. We sought to clarify the mechanism of removal torque reduction in an implant by heating with HFED. MATERIALS AND METHODS: Sixty-eight Wistar rats received titanium implants on the maxillary bone 4 weeks after extraction of the first and second molars. The control group was sacrificed 6 weeks after implant installation. In the experimental group, the implant was heated by HFED for 10 s using three different power outputs, and samples were collected at 3, 7, and 14 days after heating. Removal torque measurement and histological analysis were performed in the control and experimental groups. Implant surfaces were observed using an electron-probe microanalyzer (EPMA). Data were analyzed using Mann-Whitney U test at a significance level of 5%. RESULTS: The removal torque could not be measured in the control group due to fracture of the implant. After heating, the removal torque was measurable without fracture and decreased significantly at 14 days as compared with that at 3 days (p < .05). Heating with "min" power output resulted in a significantly smaller blank lacunae area and fewer osteoclasts at 14 days after heating (p < .05). EPMA revealed bone matrix adherence to outer surface of heated implant. CONCLUSIONS: After heating, an enlarged area of blank lacunae around the implant and an increased number of osteoclasts into the bone marrow cavity were observed, which may have contributed to the reduction in removal torque.