ABSTRACT
Endoscopic procedures performed under conscious sedation require careful monitoring of respiratory status to prevent adverse outcomes. This study utilizes a non-invasive respiratory volume monitor (RVM) that provides continuous real-time measurements of minute ventilation (MV), tidal volume and respiratory rate (RR) to assess the adequacy of ventilation during endoscopy. Digital respiratory traces were collected from 51 patients undergoing upper endoscopy with propofol sedation using an impedance-based RVM. Baseline MV for each patient was derived from a 30 s period of quiet breathing prior to sedation (MVBASELINE). Capnography data were also collected. Because RR from capnography was frequently unavailable, the RVM RR's were used for analysis. RR rate values were compared the MV measurements and sensitivity and specificity of RR to predict inadequate ventilation (MV <40 % MVBASELINE) were calculated. Initial analysis revealed that there is a weak correlation between an MV measurement and its corresponding RR measurement (r = 0.05). If MV is an actual indictor of respiratory performance, using RR as a proxy is grossly inadequate. Simulating a variety of RR alarm conditions [4-8 breaths/min (bpm)] showed that a substantial fraction of low MV measurements (MV <40 % MVBASELINE) went undetected (at 8 bpm, >70 % low MV measurements were missed; at 6 bpm, >82 % were missed; and at 4 bpm, >90 % were missed). A cut-off of 6 bpm had a sensitivity of only 18.2 %; while <40 % of all RR alarms would have coincided with a low MV (39.4 % PPV). Low RR measurements alone do not reflect episodes of low MV and are not sufficient for accurate assessment of respiratory status. RVM provides a new way to collect MV measurements which provide more comprehensive data than RR alone. Further work is ongoing to evaluate the use of MV data during procedural sedation.
