Comparison of Apnea Identified by Respiratory Inductance Plethysmography with That Detected by End-tidal CO 2 or Thermistor

As part of the Collaborative Home Infant Monitoring Evaluation (CHIME) we compared apnea identified by a customized home monitor using respiratory inductance plethysmography (RIP) with simultaneously recorded polysomnography-acquired nasal end-tidal CO(2) (PET(CO(2))) and nasal/oral thermistor in 422 infants during overnight laboratory recordings to determine concordance between techniques, sources of disagreement, and capacity of RIP to detect obstructed breaths within an apnea. Among 233 episodes of apnea identified by at least one method as >/= 16 s, 120 were observed by the CHIME monitor, 219 by PET(CO(2)), and 163 by thermistor. The positive predictive value of the CHIME-identified apnea was 89.2% (95% CI 83, 95) and 73% (95% CI 65, 81) for PET(CO(2)) and thermistor, respectively. However, the sensitivity of the CHIME monitor in identifying events detected by the other methods was only approximately 50%. Among 87 apnea events identified by all three techniques, no two methods showed high agreement in measurement of apnea duration: RIP and PET(CO(2)) (ICC = 0.54), RIP and thermistor (ICC = 0.13), PET(CO(2)) and nasal thermistor (ICC = 0.41). Among the 179 breaths identified by RIP as obstructed, 79.9% were judged to be obstructed on the PET(CO(2)) and 80.4% were judged to be obstructed on the thermistor channel. Among 238 breaths identified on PET(CO(2)) as obstructed, 54.2% were determined to be obstructed by RIP. Among 204 breaths identified on thermistor as obstructed, 55. 4% were determined to be obstructed by RIP. Reasons for discrepancies in apnea detection among channels included body movement, partial airway obstruction, and obstructed breaths. Despite these limitations the CHIME monitor provides an opportunity to record physiological data previously unavailable in the home.