BACKGROUND: Major respiratory problems during spinal anesthesia occur due to several causes, particulary, high spinal block, use of sedatives or opioids, and underlying cardiopulmonary diseases. Pulse oximetry has prevented most of these problems, but has not provided rapid and accurate information of the patient's ventilation. We measured end-tidal CO2 tension via the side-stream capnometer with a small rubber cannula and investigated its effectiveness in ventilatory monitoring under spinal anesthesia. METHODS: Nineteen patients were involved in this study. We performed spinal anesthesia with 0.5% heavy marcaine 12 mg (L3-4 interspace). After fixation of spinal sensory blockade level, 0.035 mg/kg of midazolam was administered intravenously to sedate the patient. A polyvinylchloride catheter with a rubber extending nasal cannula was used for sampling of respiratory gas. PETCO2 was measured at 15 minutes after spinal anesthetic injection (before sedation), and at 5 minutes after midazolam injection (after sadation). Arterial CO2 tension was also measured during PETCO2 measurement. RESULTS: There was no correlation of spinal maximal sensory blockade level with repiratory rate, PaCO2 and PETCO2. Linear regression analysis of arterial vs. end-tidal CO2 yielded a slope of 0.92, r=0.81 and p<0.001 (before sedation), and a slope of 0.98, r=0.79 and p<0.01 (after sedation). Arterial to end-tidal differences were 4.2+/-2.8 mmHg (before sedation) and 4.3+/-3.0 (after sedation), but there was no significant difference in comparing them with each other. Conclusions : We conclude that this form of PETCO2 measurement is useful in continuous, noninvasive monitoring of ventilation in patients under spinal anesthesia.