BACKGROUND: Tourniquet is associated with tourniquet induced hypertension (T-HTN) and with occasional circulatory collapse. These changes can be well tolerated by young healthy patients; however, in elderly patients with reduced physiologic cardiovascular reserve and blunted baroreflex, these alterations could be significantly detrimental. So we studied that whether there was any difference in hemodynamic changes in elderly patients receiving general or epidural anesthesia. METHODS: One hundred twenty patients underwent total knee arthroplasty were classified into four groups; general anesthesia in control (<65 years, group I, n=30), general anesthesia in elderly (> or = 65 years, group II, n=30), epidural anesthesia in control (group III, n=30), epidural anesthesia in elderly (group IV, n=30). Mean blood pressure (MAP), heart rate (HR) and PETCO2 were recorded throughout the experiments. RESULTS: MAP in the general anesthesia was higher than that in the epidural anesthesia after tourniquet inflation and there was pronounced occurrence of T-HTN in the group II (43.3%). After tourniquet deflation, MAP was more reduction in the group II than that in the group IV (23.3+/-10.9% vs 17.4+/-9.4%, p<0.05), and heart rate was increased in all groups except group II. Tourniquet time did not correlate with the magnitude of the changes in MAP and HR associated with tourniquet application. CONCLUSIONS: In elderly patients associated with general anesthesia, occurrence of "tourniquet hypertension" is more frequent and reduction in MAP after deflation is more marked without compensatory increasing of HR than epidural.