STUDY OBJECTIVE: To investigate the effect of propofol and midazolam on cardiac autonomic nervous system (CANS) activity during combined spinal-epidural anesthesia. DESIGN: Prospective, clinical study. SETTING: Operating room of a university hospital. PATIENTS: Forty ASA physical status I and II patients scheduled for knee surgery. INTERVENTION: Patients were randomized to receive sedation with either propofol or midazolam. MEASUREMENTS: Heart rate (HR), HR variability (HRV), systolic arterial pressure (SAP), and SAP variability (SAPV) were used for the analysis. These values were measured at the preanesthetic period, after intrathecal injection for spinal anesthesia, after sedation with propofol or midazolam, and just before the end of surgery with sedation. Cross-spectral analyses of the HR and SAP data were assessed to quantify the frequency-related coherence spectra and phase spectra. MAIN RESULTS: Spinal anesthesia itself had no effect on power spectral changes in both groups. After sedation, as for HRV, high-frequency (HF) power (HF, 0.15-0.40 Hz) did not change, whereas low-frequency (LF) power (LF, 0.04-0.15 Hz) and LF/HF, an indicator of CANS balance, significantly decreased with propofol. Further, coherence in cross-spectra presented depression in the LF band area after sedation with propofol. Before the end of surgery with sedation, LF and LF/HF in both HRV and SAPV were correlated with age in those with propofol; however, scarce relation was observed in those who received midazolam. CONCLUSIONS: Propofol was more potent than midazolam in causing CANS activity to be sympatholytic during combined spinal and epidural anesthesia and which was correlated with age only with propofol.