Important questions remain regarding how best to monitor patients during procedural sedation and analgesia (PSA). Capnometry can detect hypoventilation and apnea, yet it is rarely used in emergency patients. Even the routine practice of performing preoxygenation in low-risk patients is controversial, as supplementary oxygen can delay the detection of respiratory depression by pulse oximetry. The purpose of this study was to determine whether the capnometer or the pulse oximeter would first detect respiratory events in adults breathing room air.

During a randomized clinical trial comparing fentanyl with low-dose ketamine for PSA with titrated propofol, patients were monitored using pulse oximetry and continuous oral–nasal sampled capnography. Supplemental oxygen was administered only for oxygen desaturation. Sedating physicians identified prespecified respiratory events, including hypoventilation (end-tidal carbon dioxide > 50 mm Hg, rise of 10 mm Hg from baseline or loss of waveform) and oxygen desaturation (pulse oximetry < 92%). These events and their timing were corroborated by memory data retrieved from the monitors.

Of 63 patients enrolled, 57% (36) developed brief oxygen desaturation at some point during the sedation. All responded to oxygen, stimulation or interruption of propofol. Measurements of end-tidal carbon dioxide varied substantially between and within patients before study intervention. Hypoventilation (19 patients, 30%) was only weakly associated with oxygen desaturation (crude odds ratio 1.4 [95% confidence interval 0.47 to 4.3]), and preceded oxygen desaturation in none of the 12 patients in whom both events occurred (median lag 1:50 m:ss [interquartile range 0:01 to 3:24 m:ss]).

During PSA in adults breathing room air, desaturation detectable by pulse oximeter usually occurs before overt changes in capnometry are identified.