Transcutaneous carbon dioxide during episodes of apnea and upper airway flow-limitation

Abstract

Falling asleep is associated with decreasing chemosensitivity and increasing transcutaneous carbon dioxide (tcCO2). We hypothesize that the ventilatory gain for respiratory stimuli can be estimated by measuring the level of tcCO2. Therefore, the level of tcCO2 measured during repetitive episodes of various respiratory events should follow the order of decreasing ventilatory gain (central apnea, mixed apnea, obstructive apnea, hypopnea and flow limitation). To test our hypothesis, we analyzed the tcCO2 levels during wakefulness, steady breathing during sleep and during the above-mentioned respiratory abnormalities. A total of 174 consecutive cardio-respiratory sleep recordings were retrospectively analyzed. Recordings were included if three of the following episodes and corresponding tcCO2 levels could be determined: (1) wakefulness, (2) highest steady level of tcCO2 during steady breathing (to exclude REM), and (3) repetitive respiratory event (central, mixed, obstructive apnea or hypopnea). Inspiratory flow limitation was included when present. Minimum duration of an episode was either five minutes or at least ten similar respiratory events. TcCO2 during the respiratory episodes were tested against those measured during wakefulness and steady breathing during sleep (Wilcoxon signed ranks test). Thirty-one recordings (29 m/2f) with 69 episodes were included to the study. Mean age was 51 (range 32–71) and mean AHI was 25.8/h (range 8.6–52.9/h). TcCO2 while awake (5.21 kPa) did not differ from that during central (5.12 kPa, p = 0.128, n = 7) or mixed apnea (5.14 kPa, p = 0.068, n = 4), whereas it did differ from obstructive apnea, hypopnea, steady breathing and flow-limitation (p ⩽ 0.002). TcCO2 during steady breathing (5.63 kPa) was higher during episodes of central apnea (5.12 kPa, p = 0.018) and hypopnea (5.59 kPa, p = 0.014, n = 20) but lower during flow limitation (5.64 kPa, p < 0.000, n = 26). TcCO2 during steady breathing was at similar level as recorded during episodes of mixed (5.14 kPa, p = 0.273, n = 4) or obstructive apnea (5.32 kPa, p = 0.136, n = 12). As predicted by our hypothesis, tcCO2 levels reflect the ventilatory gain: respiratory events associated with high ventilatory gain are accompanied by tcCO2 levels higher than wakefulness but lower than steady breathing. In contrast, flow limitation with low ventilatory gain appears with high tcCO2. TcCO2 signal during sleep could be used to rank respiratory events according to their ventilatory gain. The authors would like to thank Aaro Salminen for providing valuable assistance in the preparation of this work.