Effect of Hypercapnia on Total Pulmonary Resistance during Wakefulness and during NREM Sleep1-3
M. SAFWAN BADR, JAMES B. SKATRUD, PEGGY M. SIMON, and JEROME A. DEMPSEY With the technical assistance of Dominic S. Puleo and James D. Lookabaugh
Introduction
T he activity of upper airway muscles is an important determinant of upper airway patency during inspiration. The net effect of augmented ventilatory drive on upper airway patency is determined by the balance between the activities of the inspiratory muscles and the upper airway muscles (1, 2). This is relevant to human sleep, especially in snorers and patients with obstructive sleep apnea syndrome. Previous studies comparing EMG measurements between two muscle groups have reached conflicting conclusions regarding the balance between inspiratory and upper airway muscle activity. Studies in anesthetized (3) and sleeping (4, 5) animals have suggested a higher threshold for activation of upper airway dilating muscles compared with the diaphragm, which would predispose to upper airway obstruction at low levels of chemical stimuli. In contrast, studies in awake humans (6, 7) and unanesthetized decerebrate cats (8) have demonstrated a linear relationship between the diaphragm and the genioglossal muscle recruitment in response to chemical stimuli. Thus, firm conclusions regarding upper airway patency cannot be drawn from these electromyographic studies comparing two muscle groups. Furthermore, the susceptibility of the upper airway to collapse may be an important determinant of upper airway patency in response to alterations in chemical stimuli (9). The purpose of this study was to determine the effect of increased ventilatory drive on upper airway patency. Total pulmonary resistance (RL) was used as an index of upper airway patency. Changes in RLwere measured in response to increased ventilatory drive with hypercapnia during wakefulness and NREM sleep. Methods
Subjects Thirteen healthy, nonsmoking subjects were
406
We Investigated the effect of different levels of hypercapnia on total pulmonary resistance (RL) In 13 subjects ranging from nonsnorers with low RL to snorers with high RL and dynamic narrowing of the upper airway during Inspiration. Added CO2 was adjusted to achieve a steady-state Increase In PETC0 2 of +2, +4, or +6 mm Hg. RL was measured at peak Inspiratory flow (RLpf), at maximal resistance within breath (RLmax), and at 10 equally spaced points within Inspiration In several trials. During wakefulness, hypercapnia was associated with decreased RLmlx. During steady state +6 mm Hg hypercapnia, RLmax decreased by 30% (p < 0.01). During NREM sleep, low levels of hypercapnia did not affect RL. However, +6 mm Hg hypercapnia was associated with decreased RLmax In six of eight subjects (p 0.07), especially In subjects with high RLmax during room air breathing. The effects of hypercapnia on RLpf paralleled Its effect on RLmax. We concluded that (1) the decrease In RL during awake hypercapnia suggests an Increase In upper airway dimensions and stiffness, (2) the absence of Increased RL during low level NREM hypercapnia (despite the Increase In Inspiratory flows and collapsing pressures) also suggests an Increase In upper airway dimensions and stiffness, and (3) upper airway dilating muscles appear to be recruited In a coordinated fashion with Inspiratory muscles In normal humans during NREM sleep. The Implications of these findings In patients with obstructive sleep apnea are not clear at this point. SUMMARY
=
AM REV RESPIR DIS 1991; 144:406-414
studied. Six of them werestudied during both wakefulness and NREM sleep, two werestudied during wakefulness only, and five were studied during NREM sleep only. Sleep studies were performed in stable Stage 2 and in slow-wave sleep only; no studies were performed in Stage 1 NREM sleep. All subjects werefree ofcardiopulmonary disorders. Three had a history of occasional snoring. Subjects were asked to restrict their sleep the night before the study. Sleep deprivation was variable (total sleep time, 2 to 8 h). The study was approved by the Human Subjects Committee in our institution. Signed informed consent was obtained from every subject. The study was done during regular sleep hours.
Measurements Ventilation and timing. Ventilation was measured with a flow-through hood canopy system (10).An airtight Plexiglas'" head canopy with a tight neck seal was continuously flushed with room air at 60 to 80 L/min to prevent CO 2 accumulation. Inflow and outflow to the canopy were maintained at equal rates so that pressure in the hood remained atmospheric, and changes of flow measured in the hood were produced by the subject's ventilation. Inspiratory and expiratory flows weremeasured using a pneumotachometer attached to the canopy. Tidal volume was obtained by integrating inspiratory and expiratory flows. Volume measured from the canopy (range, 0.05 to 2.0 L) showed close
correlation (r = 0.99; SE = ± 0.02 L) compared with volume measured simultaneously using a water seal spirometer (11) or with volumes measured simultaneously with a pneumotachometer placed inside the canopy. Rib cage and abdominal volumes and respiratory cycle timing were measured using inductance plethysmography (Respitraces; Ambulatory Monitoring, Ardsley, NY). The inductance plethysmography was calibrated by an isovolume technique (12) in conjunction with an Ohio 800 rolling seal spirometer (Ohio Instruments, Madison, WI). Mechanics. Esophageal pressure was measured using an esophageal balloon catheter (n = 7), or a pressure-transducer-tipped catheter (Model no. TL-500; Millar Instruments, (Received in original form May 29, 1990 and in revised form December 13, 1990) 1 From the Medical Service, William S. Middleton Memorial Veterans Hospital, Madison, Wisconsin, and the John Rankin Laboratory of Preventive Medicine, Departments of Medicine and Preventive Medicine, University of Wisconsin, Madison, Wisconsin. 2 Supported by the Medical Research Service of the Department of VeteransAffairs and by SCOR Grant POI HL-42242-01 from the National Heart, Lung, and Blood Institute. 3 Correspondence and requests for reprints should be addressed to M. Safwan Badr, M.D., H6/380 Clinical Science Center, 600 Highland Avenue, Madison, WI 53702.
HYPERCAPNIA AND AIRWAY RESISTANCE DURING WAKEFULNESS AND NREM SLEEP
Houston, TX) (n = 6). The catheter was positioned to best reflect pleural pressure (13).RL was measured throughout the breath using the technique of von Neergard and Wirz (14). Lung elastic pressure, measured at points of zero flow, was subtracted from total intrapleural pressure to obtain resistivepressure. RL was calculated at multiple points within a breath by dividing resistivepressure by flow. RLwas measured at two points in every trial: resistance at maximal negative esophageal pressure, which represented the maximal resistance within a breath (RLmax), and resistance at peak inspiratory flow (RLpf). Expiratory resistance was measured at peak expiratory flow (RLexp). During NREM sleep, RL often changed markedly within a breath because of the dynamic narrowing of the upper airway. Therefore, we measured RL at 10 equally spaced points within an inspiration in several representative trials in two subjects with high RLmax and in two subjects with low RLmax. We used a computerized system (Buxco Electronics, Sharon, CT) that had previously been validated by manual calculation (15). Respiratory muscleactivity. Genioglossus EMG (EMG gg) was obtained by two wire electrodes inserted perorally (16). Inspiratory muscle EMG (EMGinsp) was obtained by two surface electrodes (3M Red Dot; 3M Company, St. Paul, MN) placed 2 to 4 em above the right costal margin in the anterior
407
axillary line. Pairs were placed at the level of the diaphragm at FRC and at TLC as determined by chest percussion. The electrode pair with the best signal-to-noise ratio was chosen for analysis. The raw EMG signal was amplified, band-pass-filtered from 5 to 10,000 Hz (ModeI7-D polygraph; Grass Instruments, Quincy, MA), full-wave-rectified, and integrated with a Paynter filter with a time constant of 100 ms. The area beneath this moving time average was obtained by computerized integration, divided by the duration of muscle activity, and expressed in arbitrary units. Chemicalstimuli. End-tidal CO 2 (PETC02 ) was measured breath by breath (Beckman LB-2; Beckman Instruments, Fullerton, CA), and end-tidal O2 (PEToJ was measured breath by breath (oxygen analyzer Beckman OM-11). Sleep stage. Electroencephalogram, chin electromyogram, and electrooculogram were recorded (Grass modeI7-DO). Sleep stage was analyzed according to the criteria of Rechtschaffen and Kales (17).Subtle sleep changes during CO 2 breathing weredetected using the fast Fourier transformation (FFT). The EEG signal was low-pass-filtered with an antialiasing filter and sampled by an analogue-todigital conversion board into the computer at 64 samples/so The digitized EEG signal was divided into segments (epochs) of 64 samples (1 s) for the FFT analysis. Mean frequency, total power, and relative power contribution
were determined for every epoch. This analysis was performed on 10 representative +6 mm Hg trials (18).
Protocol Subjects breathed room air (RA) for 3 to 5 min. CO 2 was added to the inflow gas and adjusted to achieve a steady-state (SS)increase in PETC02 of 2, 4, or 6 mm Hg, These will be referred to as + 2, + 4, and + 6 trials, respectively. Hypercapnia was maintained for 3 to 5 min. Hypercapnia trials were done in eight subjects during wakefulness and in 11 subjects during NREM sleep. If sleep state changed during a trial, the subject was brought back to RA, and the trial was not analyzed.
Data Analysis A trial consisted of 10breaths of RA breathing immediately preceding hypercapnia, and 10 SS hypercapnia breaths. Mean values for all RA breaths and all SS breaths were obtained for ever trial. Trials of similar levels of hypercapnia were averaged in every subject. Student's paired t test was used to compare RA to SS under each level of hypercapnia (19). Results
Effects of Hypercapnia on Ventilation and Mechanics during Wakefulness The effectsof hypercapniaon ventilation
TABLE 1 EFFECT OF HYPERCAPNIA ON VENTILATION AND MECHANICS DURING WAKEFULNESS·
Variable Ventilation and timing \fE, Umin VT, L f, breaths/min TI, s TE, s VTITI, Us VTITE, Us TIITT Chemical stimuli PETC02, mm Hg PET02, mm Hg Mechanics Ppf, cm H2O Flowpf, Us RLpf, cm H2O·L-1·s Ppp, cm H2O Flowpp, Us RLmax, cm H2O·L-1.s RLexp, cm H2O·L-1·s
+ 2 mm Hg Hypercapnia
+ 4 mm Hg Hypercapnia
+ 6 mm Hg Hypercapnia
(n "" 5)
(n "" 7)
(n "" 8)
RA
7.50 0.48 15.5 1.63 2.94 0.23 0.14 0.39
± ± ± ± ± ± ± ±
SS
3.43 0.20 1.5 0.26 0.89 0.05 0.05 0.03
38.6 ± 3.1 109.9 ± 5.7 4.60 0.31 14.2 4.69 0.31 14.9 9.15
± ± ± ± ± ± ±
2.18 0.06 4.4 2.11 0.07 3.9 4.1
9.91 0.57 17.8 1.67 2.25 0.31 0.23 0.43
RA
± 3.28§ ± 0.21* ± 2.7t
± 0.41 ± 0.79* ± 0.10§ ::f: 0.08§ ± 0.04
40.9 ± 3.1§ 118.3 ± 5.0§ 4.95 ± 0.39 ± 12.3::f: 5.00 ± 0.38 ± 12.8 ± 9.16 ±
2.59 0.09* 4.1 2.56 0.11* 4.0t 4.8
7.82 0.47 16.1 1.60 2.48 0.25 0.16 0.40
± ± ± ± ± ± ± ±
2.82 0.20 1.7 0.22 0.57 0.06 0.05 0.04
38.9 ± 2.6 113.7 ± 7.3 4.02 0.34 11.9 4.03 0.33 12.3 9.61
± ± ± ± ± ± ±
RA
SS
2.03 0.11 4.2 2.01 0.11 4.2 3.7
13.83 0.72 18.4 1.62 2.18 0.38 0.29 0.43
± ± ± ± ± ± ± ±
4.04§ 0.25§ 3.4 0.35 0.65t 0.12* 0.11§ 0.03t
42.8 ± 2.4§ 124.9 ± 6.7§ 5.66 0.54 10.6 5.75 0.51 11.5 7.6
± 3.45t ± 0.23* ± 4.4t
± 3.38t ± O.24t ± 4.5 ± 3.2
9.01 0.52 14.1 1.56 2.09 0.23 0.18 0.43
± ± ± ± ± ± ± ±
SS
3.29 0.23 6.1 0.13 0.37 0.06 0.06 0.03
37.9 ± 2.4 115.0 ± 6.8 4.11 0.33 13.1 4.14 0.31 13.8 9.69
± 1.89 ± 0.12 ± 4.8
± 1.85 ± 0.12 ± 4.7 ± 4.5
16.83 0.90 18.7 1.15 1.87 0.49 0.40 0.45
± ± ± ± ±
4.23t 0.30t 3.9 0.29 0.42 ± 0.15§ ± 0.14* ± 0.02
44.3 ± 2.6* 127.1 ± 6.9§ 5.41 0.60 9.63 5.41 0.56 10.4 8.11
± 2.47t ± 0.23* ± 4.00*
± 2.48t ± 0.25* ± 4.0* ± 2.87
Definitionof abbreviations: AA ... room air control; SS ... steady state; VE • minute ventilation; VT ... tidal volume; f • frequency; TI ... inspiratory time; TE ... expiratory time; VTfTI ... mean inspiratory flow; VTiTE ... mean expiratory flow; TliTT ... inspiratory timeltotal respiratory cycle duration; Ppf ... resistive pressure measured at peak inspiratory flow; Flowpl ... inspiratory flow measured at peak inspiratory flow; ALpl • total pulmonary resistance measured at peak inspiratory flow; Ppp ... resistive pressure measured at peak negative esophageal pressure; Flowpp ... inspiratory flow measured at peak negative esophageal pressure; ALmax ... maximal total pulmonary resistance measured at peak negative esophageal pressure; ALexp ... expiratory resistance measured at peak expiratory flow. • Values are means ± SO.
t p < 0.05. *p
~ .iii
o ~1---+-+-+---4 40
1/ ,~ r~'\\ t~ /:'
2
Ql
a:: /
:
. ./"
'!
\
Subject MF +2 torr Hypercapnia
+2 torr Hypercapnia 0.45
+6 torr Hypercapnia
T
I
040t 035
~
20.:'
~./
10 10
'iii
\.:
;'
\ :
.
Ql
a::
o +--i1---+--+--+---4 o W ~ ~ Ml000
:
:
W
~
~
Inspiratory time, %
M100
f"\
u
" "
l>
.
Sd
o w
~
~
Ml000
W
~
~
Inspiratory time, %
1
0.30,
~ 025t
~
Ml00
C< 0.20
.9 u,
t
0.05
0.00
r
o
. f~
.-.~
I
~;~~t~;'
--+--t--'--t--I
2
4
6
8
10 12
a
2
4
6
Resistive Pressure, em Hp
8
10 12 Resistive Pressure, em H p
Fig. 7. A. The effect of hypercapnia on continuous within-breath RL, resistive pressure, and inspiratory flow in two nonsnorers with low RL. See figure 6 legend for explanation of symbols. B. Plots of inspiratory flow versus resistive pressure in the same two subjects.
values. Thus, measurement of RL at a who demonstrated a continuously rising fixed flow rate may not accurately reflect RL within inspiration. This corroborates previous studies demonstrating upper actual dynamic changes within the breath, airway collapsibility in snorers (9, 11,25, especially in snorers (25, 26). The anatomic and physiologic com- 26). Thus, RLmax was higher than RLpf plexity of the upper airway would argue in snorers but not in nonsnorers, suggesting dynamic narrowing of the upper for measurement of RL at more than one point within inspiration. To delineate the airway. dynamic changes of RL within a breath, Hypercapnia and Upper we measured RL at two points within inAirway Patency spiration in every trial, and at 10 equally spaced points in several representative Hypercapnic augmentation of ventilatosubjects. The first measurement of resis- ry drive was associated with decreased tance at peak inspiratory flow (RLpr) RLduring wakefulness and no significant reflects the highest attainable RLin a non- change during NREM sleep, suggesting collapsible tube with turbulent flow. preservation of upper airway patency However, measurement of RLpf does not during hypercapnic stimulation of ventake into account upper airway collapsi- tilatory drive. Our data corroborate those bility and the subsequent decrease in flow of previous studies that found a decrease and increase in RLafter peak inspiratory in RL in awake seated (27) or supine (28) flow. Therefore, we measured RLmax, . subjects, anesthetized dogs (29), anesthewhich is likely to reflect lowest upper air- tized cats (30), and sleeping infants (31). way dimensions within-breath (9, 11). The A more pronounced decrease in RL was ratio of RLmax/RLpf reflects dynamic noted in studies of anesthetized animals, narrowing of the upper airway. Finally, perhaps reflecting higher levels of hyperwe measured RL at 10 equally spaced capnia in these studies. An increase in points within inspiration in several rep- upper airway dimensions with increased resentative trials (15). This confirmed our ventilatory drive is not limited to hyperfindings from the two-point measure- capnic stimulation. Oliven and coworkment of RL,namely, the dynamic changes ers (32) found that ventilatory stimulain RLduring a breath, especially in snorers tion with salicylates was associated with
decreased RLin anesthetized dogs. Warner and colleagues (9) induced periodic breathing with sustained hypoxia. The combination of hypoxia and hypercapnia after a central apnea was associated with a high EMGinsp and low RL (9). Thus, high ventilatory drive is likely to decrease RL regardless of its mechanism. What is the mechanism of decreased RL in response to hypercapnic augmentation of ventilatory drive? A change in sleep state as a result of hypercapnia is an unlikely cause of decreased RL since sleep state remained constant by conventional criteria and spectral analysis of EEG. An increase in EELV is also an unlikely cause of decreased RL (21) since EELV remained constant. However, Van de Graaf (33) has shown that thoracic inspiratory activity provides caudal traction on the upper airway, which acts independent of upper airway dilating muscle activity to maintain upper airway patency despite increased negative intrathoracic pressures. This effect is proportional to the thoracic inspiratory activity. It is likely that this mechanism contributes to decreased RL, especially in subjects with high resistance and evidence of dynamic narrowing of the upper airway; the presence of a hypotonic
HYPERCAPNIA AND AIRWAY RESISTANCE DURING WAKEFULNESS AND NREM SLEEP
upper airway may "permit" a more effectivetransmission of traction. Our finding of a more pronounced decrease in RL in response to hypercapnia is entirely consistent with this explanation. Another likely mechanism is the augmentation of upper airway dilating muscles in response to hypercapnia. Our data do not address the role of individual muscles or the contribution of the genioglossus relative to other upper airway dilating muscles. The presence of phasic inspiratory EMG gg activity in subjects with high RLand flow limitation, combined with the lack of correlation between EMGgg augmentation and the change in RL, argue against the genioglossus as the primary upper airway dilating muscle. Conversely, the subject with the greatest decrease in RL did show the greatest recruitment of EMGgg • Thus, the augmentation of upper airway dilating muscle activity could also be responsible for decreased RL. It is likely that several muscle groups, including the genioglossus, participate in upper airway dilation. We conclude that inspiratory thoracic muscle activity and recruitment of upper airway dilating muscles contribute to decreased RL. The relative contribution of each potential mechanism is not addressed by our data. The hypercapnic effect on RLsuggests an increase in upper airway caliber during wakefulness and a smaller increase during NREM sleep. We considered several possibilities to explain the difference in hypercapnic effect on RL between wakefulness and NREM sleep. One possible explanation is the smaller upper airway diameter during NREM sleep. This is an unlikely explanation because an equal increase in diameter would represent a greater relative increase during NREM sleep, and hence lead to a greater decrease in RL. A higher threshold for activation of upper airway dilating muscle groups during NREM sleep is also unlikely; this would lead to increased RL at low levels of chemical stimuli, which did not occur. The most likely explanation is the lower tonic activity during NREM sleep (34), which renders upper airway dilating muscles less capable of transforming a stimulus into mechanical output, Le., dilatation, during NREM sleep. In other words, equal stimuli may result in greater upper airway dilatation during wakefulness. Thus, we view the difference in hypercapnic effect on RLbetween wakefulness and NREM sleep as a quantitative rather than a qualitative difference. This explanation is purely speculative in the absence of direct mea-
surements of neural output or quantitative EMG response studies. Our data corroborate previous studies, using EMG measurements, which suggested that the activation of upper airway muscles and inspiratory muscles is coordinated to maintain upper airway patency (6-8). However, our data do not address patterns of activation of upper airway inspiratory muscles versus the diaphragm since we did not obtain upper airway EMG data in every subject. Furthermore, measured activity of upper airway muscles does not necessarily reflect their shortening and/or dilating effect. Nevertheless, wedisagree with the notion that low levels of chemical stimuli would predispose the upper airway to collapse, attributed to a preferential activation of the inspiratory muscles relative to the upper airway dilating muscles (3-5). These studies were done in animals, and some were done under general anesthesia, which may account for part of the differences in findings (8). Most importantly, RL was not measured in any of these studies. Measurement of EMG activity is unlikely to be informative about airway patency, given the complex interaction between several upper airway dilating muscles. The upper airway has been shown to act as a Starling resistor (35), i.e., a passivelycollapsible segment if subjected to augmented negative pressures such as diaphragmatic pacing (36) or subatmospheric nasal pressure (37). Our data demonstrated the ability of the upper airway to maintain its patency despite augmented negative pressures. In this context, the upper airway departs from the classic Starling resistor model because it is under neural control. Teleologically, it is beneficial to possess common control mechanisms for both inspiratory muscles and upper airway muscles. Otherwise, upper airway would be subjected to collapse under conditions of high ventilatory drive. Our data also suggest that the recruitment of upper airway dilating muscles is sufficient in magnitude, and synchronized in timing, to prevent upper airway collapse in response to hypercapnic augmentation of ventilatory drive. Furthermore, our findings suggest that sleep per se does not alter the coordinated recruitment mechanisms for the two muscle groups in normal subjects and healthy snorers. However, this may not be the case in patients with obstructive sleep apnea (OSA). In a previous study during hypoxia-induced periodic breathing (9),
413
one patient with OSA developed obstructive apneas when breathing was initiated after central apnea. Nevertheless, our finding of a stabilizing role for hypercapnia in subjects with high RLmay explain, at least partially, the recent study that suggested a beneficial effect from CO 2 in patients with OSA (38). The mechanism(s) by which CO 2 may be beneficial in patients with OSA and the effects of other chemical stimuli such as hypoxia have not been studied. Finally, our data question the use of CO 2 response during sleep as an indicator of "chemosensitivity" for two reasons. First, sleep per se has a highly variable effect on airway resistance, which should by itself reduce the response to any stimulus. Second, the increase in RLis further modified during the administration of CO 2 • Thus, the ventilatory response to CO 2 depends not only on the levelof CO 2 but on baseline airway resistance and its change in response to hypercapnia. This complex interaction argues against the use of hypercapnic ventilatory response as an indicator of sleep effect on chemoresponsiveness. In summary, we have shown that hypercapnic augmentation of ventilation recruits inspiratory muscles and upper airway dilating muscles and maintains upper airway patency during wakefulness and NREM sleep. Acknowledgment The writers thank Carol Steinhart for excellent editorial assistance. References 1. Remmers JE, de Groot WJ, Sauerland EK, Anch AM. Pathogenesis of upper airway occlusion during sleep. J Appl Physiol1978; 44:931-8. 2. Brouillette RT, Thach BT. A neuromuscular mechanism maintaining extrathoracic airwaypatency. J Appl Physiol 1977; 46:772-9. 3. Weiner D, Mitra J, Salamone J, Cherniack NS. Effect of chemical stimuli on nerves supplying upper airway muscles. J Appl Physiol1982; 52:530-6. 4. Haxhiu MA, van Lunteren E, Mitra J, Cherniack NS. Comparison of the response of diaphragm and upper airway dilating muscle activity in sleeping cats. Respir Physiol1987; 70:183-93. 5. Parisi RA, Neubauer JA, Frank MM, Edelman N, Santiago TV. Correlation between genioglossal and diaphragmatic responses to hypercapnia during sleep. Am Rev Respir Dis 1987; 135:378-82. 6. Onal E, Lopata M, O'Connor TD. Diaphragmatic and genioglossal electromyogram responses to COl rebreathing in humans. J Appl Physio11981; 50:1052-5. 7. Patrick GB, Strohl KP, Rubin SB, Altose MD. Upper airway and diaphragm muscle responses to chemical stimulation and loading. J Appl Physiol 1982; 53:1133-7. 8. Hwang J-C, St John WM, Bartlett 0 Jr. Respiratory-related hypoglossal nerve activity: influence of anesthetics. J Appl Physiol 1983;
BADR, SKATRUD, SIMON, AND DEMPSEY
414 55:785-92. 9. Warner G, Skatrud JB, Dempsey JA. Effect of hypoxia-induced periodic breathing on upper airway obstruction during sleep. J Appl Physiol1987; 62:2201-11. 10. Sorkin B, Rapoport DM, Falk DB, Goldring RM. Canopy ventilation monitor for quantitative measurement of ventilation during sleep. J Appl Physiol 1980; 48:724-30. 11. Skatrud JB, Dempsey JA. Airway resistance and respiratory muscle function in snorers during NREM sleep. J Appl Physiol 1985; 59:328-35. 12. Konno K, Mead J. Measurement of the separate volume changes of rib cage and abdomen during breathing. J Appl Physiol 1967; 22:407-22. 13. Baydur A, Behrakis PK, Zin WA, Jaeger M, Milic-Emili J. A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 1982; 126:788-91. 14. von Neergaard J, Wirz K. Die Messung der StrOmungswiderstande in den Atemwegen des Menschen, insbesondere bei Asthma und Emphysem. Z Klin Med 1927; 105:51-82. 15. Henke KG, Dempsey JA, Kowitz JM, Skatrud JB. Effect of sleep-induced increases in upper airway resistance on ventilation. J Appl Physioll990; 69:617-24. 16. Sauerland EK, Harper RM. The human tongue during sleep: electromyographic activity of the genioglossus muscle. Exp Neuro11976; 51:160-70. 17. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques, and scoring system for sleep stages of human subjects. Bethesda: National Institutes of Health, 1988 (publication no. 204). 18. Perno GS. EEG monitoring. In: Tompkins WJ,
Webster JG, eds. Design of microcomputer based medical instrumentation. Englewood Cliffs: Prentice Hall, 1981; 148-56. 19. Colton T. Statistics in medicine. Boston: Little Brown and Co, 1974; 219-21. 20. Hudgel DW, Martin RJ, Johnson B, Hill P. Mechanics ofthe respiratory system and breathing pattern during sleep in normal humans. J Appl Physiol 1984; 56:133-7. 21. Begle RL, Badr S, Skatrud JB, Dempsey JA. Effect of lung inflation on pulmonary resistance during NREM sleep. Am Rev Respir Dis 1990; 141:854-60. 22. Waldron MA, Fisher JT. Differential effects of CO 2 and hypoxia on bronchomotor tone in the newborn dog. Respir Physiol 1988; 72:271-82. 23. Ingram RH Jr. Effects of airway versus arterial CO 2 changes on lung mechanics in dogs. J Appl Physiol 1975; 38:6703-7. 24. Hyatt RE, Wilcox RE. Extrathoracic airway resistance in man. J Appl Physiol1961; 16:326-30. 25. Hudgel DW, Hendricks C, Hamilton HB. Characteristics of the upper airway pressure-flow relationship during sleep. J Appl Physiol 1988; 64:1930-5. 26. Wiegand L, Zwillich CW, White DP. Collapsibility of the human upper airway during normal sleep. J Appl Physiol 1989; 55:1800-8. 27. Spann RW, Hyatt RE. Factors affecting upper airway resistance in conscious man. J Appl Physiol 1971; 31:708-12. 28. Series F, Cormier Y, Desmeules M, La Forge J. Influence of respiratory drive on upper airway resistance in normal men. J Appl Physiol 1989; 66:1242-9. 29. Oliven A, Odeh M, Gavriely N. Effect of
hypercapnia on upper airway resistance and collapsibility in anesthetized dogs. Respir Physio11989; 75:29-38. 30. Bartlett D Jr. Effects of hypercapnia and hypoxia on laryngeal resistance to airflow. Respir Physiol 1979; 37:293-302. 31. Miller MJ, DiFiore JM, Strohl KP, Carlo WA, Martin JR. Decreased supraglottic resistance in response to CO2 rebreathing in infants (abstract). Am Rev Respir Dis 1989; 139:AI82. 32. OlivenA, OdehM, GavrielyN. Effectofsalicylate on upper airway stability and pressure flow: relationship in anesthetized dogs. J Lab Clin Med 1989; 113:8-14. 33. Van de Graaf WB. Thoracic influence on upper airway patency. J Appl Physio11988;65:2124-33. 34. Sauerland EK, Orr WC, Harriston LE. EMG patterns of oropharyngeal muscles during respiration in wakefulness and sleep. Electromyogr Clin Neurophysiol 1981; 21:307-16. 35. Lambert RK, Wilson TA. Flow limitation in a collapsible tube. J Appl PhysioI1972; 33:150-3. 36. Hyland RH, Hutcheon MA, Perl A, et al. Upper airway occlusion induced by diaphragm pacing for primary alveolar hypoventilation: implications for the pathogenesis of obstructive sleep apnea. Am Rev Respir Dis 1981; 124:180-5. 37. Schwartz AR, Smith PL, Wise RA, Gold AR, Permutt S. Induction of upper airway occlusion in sleeping individuals with subatmospheric nasal pressure. J Appl Physiol 1988; 64:535-42. 38. Hudgel DW, Hendricks C, Dadley A. Alteration in obstructive apnea pattern induced by changes in oxygen- and carbon-dioxide-inspired concentrations. Am Rev Respir Dis 1988; 138:16-9.
M. SAFWAN BADR, JAMES B. SKATRUD, PEGGY M. SIMON, and JEROME A. DEMPSEY With the technical assistance of Dominic S. Puleo and James D. Lookabaugh
Introduction
T he activity of upper airway muscles is an important determinant of upper airway patency during inspiration. The net effect of augmented ventilatory drive on upper airway patency is determined by the balance between the activities of the inspiratory muscles and the upper airway muscles (1, 2). This is relevant to human sleep, especially in snorers and patients with obstructive sleep apnea syndrome. Previous studies comparing EMG measurements between two muscle groups have reached conflicting conclusions regarding the balance between inspiratory and upper airway muscle activity. Studies in anesthetized (3) and sleeping (4, 5) animals have suggested a higher threshold for activation of upper airway dilating muscles compared with the diaphragm, which would predispose to upper airway obstruction at low levels of chemical stimuli. In contrast, studies in awake humans (6, 7) and unanesthetized decerebrate cats (8) have demonstrated a linear relationship between the diaphragm and the genioglossal muscle recruitment in response to chemical stimuli. Thus, firm conclusions regarding upper airway patency cannot be drawn from these electromyographic studies comparing two muscle groups. Furthermore, the susceptibility of the upper airway to collapse may be an important determinant of upper airway patency in response to alterations in chemical stimuli (9). The purpose of this study was to determine the effect of increased ventilatory drive on upper airway patency. Total pulmonary resistance (RL) was used as an index of upper airway patency. Changes in RLwere measured in response to increased ventilatory drive with hypercapnia during wakefulness and NREM sleep. Methods
Subjects Thirteen healthy, nonsmoking subjects were
406
We Investigated the effect of different levels of hypercapnia on total pulmonary resistance (RL) In 13 subjects ranging from nonsnorers with low RL to snorers with high RL and dynamic narrowing of the upper airway during Inspiration. Added CO2 was adjusted to achieve a steady-state Increase In PETC0 2 of +2, +4, or +6 mm Hg. RL was measured at peak Inspiratory flow (RLpf), at maximal resistance within breath (RLmax), and at 10 equally spaced points within Inspiration In several trials. During wakefulness, hypercapnia was associated with decreased RLmlx. During steady state +6 mm Hg hypercapnia, RLmax decreased by 30% (p < 0.01). During NREM sleep, low levels of hypercapnia did not affect RL. However, +6 mm Hg hypercapnia was associated with decreased RLmax In six of eight subjects (p 0.07), especially In subjects with high RLmax during room air breathing. The effects of hypercapnia on RLpf paralleled Its effect on RLmax. We concluded that (1) the decrease In RL during awake hypercapnia suggests an Increase In upper airway dimensions and stiffness, (2) the absence of Increased RL during low level NREM hypercapnia (despite the Increase In Inspiratory flows and collapsing pressures) also suggests an Increase In upper airway dimensions and stiffness, and (3) upper airway dilating muscles appear to be recruited In a coordinated fashion with Inspiratory muscles In normal humans during NREM sleep. The Implications of these findings In patients with obstructive sleep apnea are not clear at this point. SUMMARY
=
AM REV RESPIR DIS 1991; 144:406-414
studied. Six of them werestudied during both wakefulness and NREM sleep, two werestudied during wakefulness only, and five were studied during NREM sleep only. Sleep studies were performed in stable Stage 2 and in slow-wave sleep only; no studies were performed in Stage 1 NREM sleep. All subjects werefree ofcardiopulmonary disorders. Three had a history of occasional snoring. Subjects were asked to restrict their sleep the night before the study. Sleep deprivation was variable (total sleep time, 2 to 8 h). The study was approved by the Human Subjects Committee in our institution. Signed informed consent was obtained from every subject. The study was done during regular sleep hours.
Measurements Ventilation and timing. Ventilation was measured with a flow-through hood canopy system (10).An airtight Plexiglas'" head canopy with a tight neck seal was continuously flushed with room air at 60 to 80 L/min to prevent CO 2 accumulation. Inflow and outflow to the canopy were maintained at equal rates so that pressure in the hood remained atmospheric, and changes of flow measured in the hood were produced by the subject's ventilation. Inspiratory and expiratory flows weremeasured using a pneumotachometer attached to the canopy. Tidal volume was obtained by integrating inspiratory and expiratory flows. Volume measured from the canopy (range, 0.05 to 2.0 L) showed close
correlation (r = 0.99; SE = ± 0.02 L) compared with volume measured simultaneously using a water seal spirometer (11) or with volumes measured simultaneously with a pneumotachometer placed inside the canopy. Rib cage and abdominal volumes and respiratory cycle timing were measured using inductance plethysmography (Respitraces; Ambulatory Monitoring, Ardsley, NY). The inductance plethysmography was calibrated by an isovolume technique (12) in conjunction with an Ohio 800 rolling seal spirometer (Ohio Instruments, Madison, WI). Mechanics. Esophageal pressure was measured using an esophageal balloon catheter (n = 7), or a pressure-transducer-tipped catheter (Model no. TL-500; Millar Instruments, (Received in original form May 29, 1990 and in revised form December 13, 1990) 1 From the Medical Service, William S. Middleton Memorial Veterans Hospital, Madison, Wisconsin, and the John Rankin Laboratory of Preventive Medicine, Departments of Medicine and Preventive Medicine, University of Wisconsin, Madison, Wisconsin. 2 Supported by the Medical Research Service of the Department of VeteransAffairs and by SCOR Grant POI HL-42242-01 from the National Heart, Lung, and Blood Institute. 3 Correspondence and requests for reprints should be addressed to M. Safwan Badr, M.D., H6/380 Clinical Science Center, 600 Highland Avenue, Madison, WI 53702.
HYPERCAPNIA AND AIRWAY RESISTANCE DURING WAKEFULNESS AND NREM SLEEP
Houston, TX) (n = 6). The catheter was positioned to best reflect pleural pressure (13).RL was measured throughout the breath using the technique of von Neergard and Wirz (14). Lung elastic pressure, measured at points of zero flow, was subtracted from total intrapleural pressure to obtain resistivepressure. RL was calculated at multiple points within a breath by dividing resistivepressure by flow. RLwas measured at two points in every trial: resistance at maximal negative esophageal pressure, which represented the maximal resistance within a breath (RLmax), and resistance at peak inspiratory flow (RLpf). Expiratory resistance was measured at peak expiratory flow (RLexp). During NREM sleep, RL often changed markedly within a breath because of the dynamic narrowing of the upper airway. Therefore, we measured RL at 10 equally spaced points within an inspiration in several representative trials in two subjects with high RLmax and in two subjects with low RLmax. We used a computerized system (Buxco Electronics, Sharon, CT) that had previously been validated by manual calculation (15). Respiratory muscleactivity. Genioglossus EMG (EMG gg) was obtained by two wire electrodes inserted perorally (16). Inspiratory muscle EMG (EMGinsp) was obtained by two surface electrodes (3M Red Dot; 3M Company, St. Paul, MN) placed 2 to 4 em above the right costal margin in the anterior
407
axillary line. Pairs were placed at the level of the diaphragm at FRC and at TLC as determined by chest percussion. The electrode pair with the best signal-to-noise ratio was chosen for analysis. The raw EMG signal was amplified, band-pass-filtered from 5 to 10,000 Hz (ModeI7-D polygraph; Grass Instruments, Quincy, MA), full-wave-rectified, and integrated with a Paynter filter with a time constant of 100 ms. The area beneath this moving time average was obtained by computerized integration, divided by the duration of muscle activity, and expressed in arbitrary units. Chemicalstimuli. End-tidal CO 2 (PETC02 ) was measured breath by breath (Beckman LB-2; Beckman Instruments, Fullerton, CA), and end-tidal O2 (PEToJ was measured breath by breath (oxygen analyzer Beckman OM-11). Sleep stage. Electroencephalogram, chin electromyogram, and electrooculogram were recorded (Grass modeI7-DO). Sleep stage was analyzed according to the criteria of Rechtschaffen and Kales (17).Subtle sleep changes during CO 2 breathing weredetected using the fast Fourier transformation (FFT). The EEG signal was low-pass-filtered with an antialiasing filter and sampled by an analogue-todigital conversion board into the computer at 64 samples/so The digitized EEG signal was divided into segments (epochs) of 64 samples (1 s) for the FFT analysis. Mean frequency, total power, and relative power contribution
were determined for every epoch. This analysis was performed on 10 representative +6 mm Hg trials (18).
Protocol Subjects breathed room air (RA) for 3 to 5 min. CO 2 was added to the inflow gas and adjusted to achieve a steady-state (SS)increase in PETC02 of 2, 4, or 6 mm Hg, These will be referred to as + 2, + 4, and + 6 trials, respectively. Hypercapnia was maintained for 3 to 5 min. Hypercapnia trials were done in eight subjects during wakefulness and in 11 subjects during NREM sleep. If sleep state changed during a trial, the subject was brought back to RA, and the trial was not analyzed.
Data Analysis A trial consisted of 10breaths of RA breathing immediately preceding hypercapnia, and 10 SS hypercapnia breaths. Mean values for all RA breaths and all SS breaths were obtained for ever trial. Trials of similar levels of hypercapnia were averaged in every subject. Student's paired t test was used to compare RA to SS under each level of hypercapnia (19). Results
Effects of Hypercapnia on Ventilation and Mechanics during Wakefulness The effectsof hypercapniaon ventilation
TABLE 1 EFFECT OF HYPERCAPNIA ON VENTILATION AND MECHANICS DURING WAKEFULNESS·
Variable Ventilation and timing \fE, Umin VT, L f, breaths/min TI, s TE, s VTITI, Us VTITE, Us TIITT Chemical stimuli PETC02, mm Hg PET02, mm Hg Mechanics Ppf, cm H2O Flowpf, Us RLpf, cm H2O·L-1·s Ppp, cm H2O Flowpp, Us RLmax, cm H2O·L-1.s RLexp, cm H2O·L-1·s
+ 2 mm Hg Hypercapnia
+ 4 mm Hg Hypercapnia
+ 6 mm Hg Hypercapnia
(n "" 5)
(n "" 7)
(n "" 8)
RA
7.50 0.48 15.5 1.63 2.94 0.23 0.14 0.39
± ± ± ± ± ± ± ±
SS
3.43 0.20 1.5 0.26 0.89 0.05 0.05 0.03
38.6 ± 3.1 109.9 ± 5.7 4.60 0.31 14.2 4.69 0.31 14.9 9.15
± ± ± ± ± ± ±
2.18 0.06 4.4 2.11 0.07 3.9 4.1
9.91 0.57 17.8 1.67 2.25 0.31 0.23 0.43
RA
± 3.28§ ± 0.21* ± 2.7t
± 0.41 ± 0.79* ± 0.10§ ::f: 0.08§ ± 0.04
40.9 ± 3.1§ 118.3 ± 5.0§ 4.95 ± 0.39 ± 12.3::f: 5.00 ± 0.38 ± 12.8 ± 9.16 ±
2.59 0.09* 4.1 2.56 0.11* 4.0t 4.8
7.82 0.47 16.1 1.60 2.48 0.25 0.16 0.40
± ± ± ± ± ± ± ±
2.82 0.20 1.7 0.22 0.57 0.06 0.05 0.04
38.9 ± 2.6 113.7 ± 7.3 4.02 0.34 11.9 4.03 0.33 12.3 9.61
± ± ± ± ± ± ±
RA
SS
2.03 0.11 4.2 2.01 0.11 4.2 3.7
13.83 0.72 18.4 1.62 2.18 0.38 0.29 0.43
± ± ± ± ± ± ± ±
4.04§ 0.25§ 3.4 0.35 0.65t 0.12* 0.11§ 0.03t
42.8 ± 2.4§ 124.9 ± 6.7§ 5.66 0.54 10.6 5.75 0.51 11.5 7.6
± 3.45t ± 0.23* ± 4.4t
± 3.38t ± O.24t ± 4.5 ± 3.2
9.01 0.52 14.1 1.56 2.09 0.23 0.18 0.43
± ± ± ± ± ± ± ±
SS
3.29 0.23 6.1 0.13 0.37 0.06 0.06 0.03
37.9 ± 2.4 115.0 ± 6.8 4.11 0.33 13.1 4.14 0.31 13.8 9.69
± 1.89 ± 0.12 ± 4.8
± 1.85 ± 0.12 ± 4.7 ± 4.5
16.83 0.90 18.7 1.15 1.87 0.49 0.40 0.45
± ± ± ± ±
4.23t 0.30t 3.9 0.29 0.42 ± 0.15§ ± 0.14* ± 0.02
44.3 ± 2.6* 127.1 ± 6.9§ 5.41 0.60 9.63 5.41 0.56 10.4 8.11
± 2.47t ± 0.23* ± 4.00*
± 2.48t ± 0.25* ± 4.0* ± 2.87
Definitionof abbreviations: AA ... room air control; SS ... steady state; VE • minute ventilation; VT ... tidal volume; f • frequency; TI ... inspiratory time; TE ... expiratory time; VTfTI ... mean inspiratory flow; VTiTE ... mean expiratory flow; TliTT ... inspiratory timeltotal respiratory cycle duration; Ppf ... resistive pressure measured at peak inspiratory flow; Flowpl ... inspiratory flow measured at peak inspiratory flow; ALpl • total pulmonary resistance measured at peak inspiratory flow; Ppp ... resistive pressure measured at peak negative esophageal pressure; Flowpp ... inspiratory flow measured at peak negative esophageal pressure; ALmax ... maximal total pulmonary resistance measured at peak negative esophageal pressure; ALexp ... expiratory resistance measured at peak expiratory flow. • Values are means ± SO.
t p < 0.05. *p
~ .iii
o ~1---+-+-+---4 40
1/ ,~ r~'\\ t~ /:'
2
Ql
a:: /
:
. ./"
'!
\
Subject MF +2 torr Hypercapnia
+2 torr Hypercapnia 0.45
+6 torr Hypercapnia
T
I
040t 035
~
20.:'
~./
10 10
'iii
\.:
;'
\ :
.
Ql
a::
o +--i1---+--+--+---4 o W ~ ~ Ml000
:
:
W
~
~
Inspiratory time, %
M100
f"\
u
" "
l>
.
Sd
o w
~
~
Ml000
W
~
~
Inspiratory time, %
1
0.30,
~ 025t
~
Ml00
C< 0.20
.9 u,
t
0.05
0.00
r
o
. f~
.-.~
I
~;~~t~;'
--+--t--'--t--I
2
4
6
8
10 12
a
2
4
6
Resistive Pressure, em Hp
8
10 12 Resistive Pressure, em H p
Fig. 7. A. The effect of hypercapnia on continuous within-breath RL, resistive pressure, and inspiratory flow in two nonsnorers with low RL. See figure 6 legend for explanation of symbols. B. Plots of inspiratory flow versus resistive pressure in the same two subjects.
values. Thus, measurement of RL at a who demonstrated a continuously rising fixed flow rate may not accurately reflect RL within inspiration. This corroborates previous studies demonstrating upper actual dynamic changes within the breath, airway collapsibility in snorers (9, 11,25, especially in snorers (25, 26). The anatomic and physiologic com- 26). Thus, RLmax was higher than RLpf plexity of the upper airway would argue in snorers but not in nonsnorers, suggesting dynamic narrowing of the upper for measurement of RL at more than one point within inspiration. To delineate the airway. dynamic changes of RL within a breath, Hypercapnia and Upper we measured RL at two points within inAirway Patency spiration in every trial, and at 10 equally spaced points in several representative Hypercapnic augmentation of ventilatosubjects. The first measurement of resis- ry drive was associated with decreased tance at peak inspiratory flow (RLpr) RLduring wakefulness and no significant reflects the highest attainable RLin a non- change during NREM sleep, suggesting collapsible tube with turbulent flow. preservation of upper airway patency However, measurement of RLpf does not during hypercapnic stimulation of ventake into account upper airway collapsi- tilatory drive. Our data corroborate those bility and the subsequent decrease in flow of previous studies that found a decrease and increase in RLafter peak inspiratory in RL in awake seated (27) or supine (28) flow. Therefore, we measured RLmax, . subjects, anesthetized dogs (29), anesthewhich is likely to reflect lowest upper air- tized cats (30), and sleeping infants (31). way dimensions within-breath (9, 11). The A more pronounced decrease in RL was ratio of RLmax/RLpf reflects dynamic noted in studies of anesthetized animals, narrowing of the upper airway. Finally, perhaps reflecting higher levels of hyperwe measured RL at 10 equally spaced capnia in these studies. An increase in points within inspiration in several rep- upper airway dimensions with increased resentative trials (15). This confirmed our ventilatory drive is not limited to hyperfindings from the two-point measure- capnic stimulation. Oliven and coworkment of RL,namely, the dynamic changes ers (32) found that ventilatory stimulain RLduring a breath, especially in snorers tion with salicylates was associated with
decreased RLin anesthetized dogs. Warner and colleagues (9) induced periodic breathing with sustained hypoxia. The combination of hypoxia and hypercapnia after a central apnea was associated with a high EMGinsp and low RL (9). Thus, high ventilatory drive is likely to decrease RL regardless of its mechanism. What is the mechanism of decreased RL in response to hypercapnic augmentation of ventilatory drive? A change in sleep state as a result of hypercapnia is an unlikely cause of decreased RL since sleep state remained constant by conventional criteria and spectral analysis of EEG. An increase in EELV is also an unlikely cause of decreased RL (21) since EELV remained constant. However, Van de Graaf (33) has shown that thoracic inspiratory activity provides caudal traction on the upper airway, which acts independent of upper airway dilating muscle activity to maintain upper airway patency despite increased negative intrathoracic pressures. This effect is proportional to the thoracic inspiratory activity. It is likely that this mechanism contributes to decreased RL, especially in subjects with high resistance and evidence of dynamic narrowing of the upper airway; the presence of a hypotonic
HYPERCAPNIA AND AIRWAY RESISTANCE DURING WAKEFULNESS AND NREM SLEEP
upper airway may "permit" a more effectivetransmission of traction. Our finding of a more pronounced decrease in RL in response to hypercapnia is entirely consistent with this explanation. Another likely mechanism is the augmentation of upper airway dilating muscles in response to hypercapnia. Our data do not address the role of individual muscles or the contribution of the genioglossus relative to other upper airway dilating muscles. The presence of phasic inspiratory EMG gg activity in subjects with high RLand flow limitation, combined with the lack of correlation between EMGgg augmentation and the change in RL, argue against the genioglossus as the primary upper airway dilating muscle. Conversely, the subject with the greatest decrease in RL did show the greatest recruitment of EMGgg • Thus, the augmentation of upper airway dilating muscle activity could also be responsible for decreased RL. It is likely that several muscle groups, including the genioglossus, participate in upper airway dilation. We conclude that inspiratory thoracic muscle activity and recruitment of upper airway dilating muscles contribute to decreased RL. The relative contribution of each potential mechanism is not addressed by our data. The hypercapnic effect on RLsuggests an increase in upper airway caliber during wakefulness and a smaller increase during NREM sleep. We considered several possibilities to explain the difference in hypercapnic effect on RL between wakefulness and NREM sleep. One possible explanation is the smaller upper airway diameter during NREM sleep. This is an unlikely explanation because an equal increase in diameter would represent a greater relative increase during NREM sleep, and hence lead to a greater decrease in RL. A higher threshold for activation of upper airway dilating muscle groups during NREM sleep is also unlikely; this would lead to increased RL at low levels of chemical stimuli, which did not occur. The most likely explanation is the lower tonic activity during NREM sleep (34), which renders upper airway dilating muscles less capable of transforming a stimulus into mechanical output, Le., dilatation, during NREM sleep. In other words, equal stimuli may result in greater upper airway dilatation during wakefulness. Thus, we view the difference in hypercapnic effect on RLbetween wakefulness and NREM sleep as a quantitative rather than a qualitative difference. This explanation is purely speculative in the absence of direct mea-
surements of neural output or quantitative EMG response studies. Our data corroborate previous studies, using EMG measurements, which suggested that the activation of upper airway muscles and inspiratory muscles is coordinated to maintain upper airway patency (6-8). However, our data do not address patterns of activation of upper airway inspiratory muscles versus the diaphragm since we did not obtain upper airway EMG data in every subject. Furthermore, measured activity of upper airway muscles does not necessarily reflect their shortening and/or dilating effect. Nevertheless, wedisagree with the notion that low levels of chemical stimuli would predispose the upper airway to collapse, attributed to a preferential activation of the inspiratory muscles relative to the upper airway dilating muscles (3-5). These studies were done in animals, and some were done under general anesthesia, which may account for part of the differences in findings (8). Most importantly, RL was not measured in any of these studies. Measurement of EMG activity is unlikely to be informative about airway patency, given the complex interaction between several upper airway dilating muscles. The upper airway has been shown to act as a Starling resistor (35), i.e., a passivelycollapsible segment if subjected to augmented negative pressures such as diaphragmatic pacing (36) or subatmospheric nasal pressure (37). Our data demonstrated the ability of the upper airway to maintain its patency despite augmented negative pressures. In this context, the upper airway departs from the classic Starling resistor model because it is under neural control. Teleologically, it is beneficial to possess common control mechanisms for both inspiratory muscles and upper airway muscles. Otherwise, upper airway would be subjected to collapse under conditions of high ventilatory drive. Our data also suggest that the recruitment of upper airway dilating muscles is sufficient in magnitude, and synchronized in timing, to prevent upper airway collapse in response to hypercapnic augmentation of ventilatory drive. Furthermore, our findings suggest that sleep per se does not alter the coordinated recruitment mechanisms for the two muscle groups in normal subjects and healthy snorers. However, this may not be the case in patients with obstructive sleep apnea (OSA). In a previous study during hypoxia-induced periodic breathing (9),
413
one patient with OSA developed obstructive apneas when breathing was initiated after central apnea. Nevertheless, our finding of a stabilizing role for hypercapnia in subjects with high RLmay explain, at least partially, the recent study that suggested a beneficial effect from CO 2 in patients with OSA (38). The mechanism(s) by which CO 2 may be beneficial in patients with OSA and the effects of other chemical stimuli such as hypoxia have not been studied. Finally, our data question the use of CO 2 response during sleep as an indicator of "chemosensitivity" for two reasons. First, sleep per se has a highly variable effect on airway resistance, which should by itself reduce the response to any stimulus. Second, the increase in RLis further modified during the administration of CO 2 • Thus, the ventilatory response to CO 2 depends not only on the levelof CO 2 but on baseline airway resistance and its change in response to hypercapnia. This complex interaction argues against the use of hypercapnic ventilatory response as an indicator of sleep effect on chemoresponsiveness. In summary, we have shown that hypercapnic augmentation of ventilation recruits inspiratory muscles and upper airway dilating muscles and maintains upper airway patency during wakefulness and NREM sleep. Acknowledgment The writers thank Carol Steinhart for excellent editorial assistance. References 1. Remmers JE, de Groot WJ, Sauerland EK, Anch AM. Pathogenesis of upper airway occlusion during sleep. J Appl Physiol1978; 44:931-8. 2. Brouillette RT, Thach BT. A neuromuscular mechanism maintaining extrathoracic airwaypatency. J Appl Physiol 1977; 46:772-9. 3. Weiner D, Mitra J, Salamone J, Cherniack NS. Effect of chemical stimuli on nerves supplying upper airway muscles. J Appl Physiol1982; 52:530-6. 4. Haxhiu MA, van Lunteren E, Mitra J, Cherniack NS. Comparison of the response of diaphragm and upper airway dilating muscle activity in sleeping cats. Respir Physiol1987; 70:183-93. 5. Parisi RA, Neubauer JA, Frank MM, Edelman N, Santiago TV. Correlation between genioglossal and diaphragmatic responses to hypercapnia during sleep. Am Rev Respir Dis 1987; 135:378-82. 6. Onal E, Lopata M, O'Connor TD. Diaphragmatic and genioglossal electromyogram responses to COl rebreathing in humans. J Appl Physio11981; 50:1052-5. 7. Patrick GB, Strohl KP, Rubin SB, Altose MD. Upper airway and diaphragm muscle responses to chemical stimulation and loading. J Appl Physiol 1982; 53:1133-7. 8. Hwang J-C, St John WM, Bartlett 0 Jr. Respiratory-related hypoglossal nerve activity: influence of anesthetics. J Appl Physiol 1983;
BADR, SKATRUD, SIMON, AND DEMPSEY
414 55:785-92. 9. Warner G, Skatrud JB, Dempsey JA. Effect of hypoxia-induced periodic breathing on upper airway obstruction during sleep. J Appl Physiol1987; 62:2201-11. 10. Sorkin B, Rapoport DM, Falk DB, Goldring RM. Canopy ventilation monitor for quantitative measurement of ventilation during sleep. J Appl Physiol 1980; 48:724-30. 11. Skatrud JB, Dempsey JA. Airway resistance and respiratory muscle function in snorers during NREM sleep. J Appl Physiol 1985; 59:328-35. 12. Konno K, Mead J. Measurement of the separate volume changes of rib cage and abdomen during breathing. J Appl Physiol 1967; 22:407-22. 13. Baydur A, Behrakis PK, Zin WA, Jaeger M, Milic-Emili J. A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 1982; 126:788-91. 14. von Neergaard J, Wirz K. Die Messung der StrOmungswiderstande in den Atemwegen des Menschen, insbesondere bei Asthma und Emphysem. Z Klin Med 1927; 105:51-82. 15. Henke KG, Dempsey JA, Kowitz JM, Skatrud JB. Effect of sleep-induced increases in upper airway resistance on ventilation. J Appl Physioll990; 69:617-24. 16. Sauerland EK, Harper RM. The human tongue during sleep: electromyographic activity of the genioglossus muscle. Exp Neuro11976; 51:160-70. 17. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques, and scoring system for sleep stages of human subjects. Bethesda: National Institutes of Health, 1988 (publication no. 204). 18. Perno GS. EEG monitoring. In: Tompkins WJ,
Webster JG, eds. Design of microcomputer based medical instrumentation. Englewood Cliffs: Prentice Hall, 1981; 148-56. 19. Colton T. Statistics in medicine. Boston: Little Brown and Co, 1974; 219-21. 20. Hudgel DW, Martin RJ, Johnson B, Hill P. Mechanics ofthe respiratory system and breathing pattern during sleep in normal humans. J Appl Physiol 1984; 56:133-7. 21. Begle RL, Badr S, Skatrud JB, Dempsey JA. Effect of lung inflation on pulmonary resistance during NREM sleep. Am Rev Respir Dis 1990; 141:854-60. 22. Waldron MA, Fisher JT. Differential effects of CO 2 and hypoxia on bronchomotor tone in the newborn dog. Respir Physiol 1988; 72:271-82. 23. Ingram RH Jr. Effects of airway versus arterial CO 2 changes on lung mechanics in dogs. J Appl Physiol 1975; 38:6703-7. 24. Hyatt RE, Wilcox RE. Extrathoracic airway resistance in man. J Appl Physiol1961; 16:326-30. 25. Hudgel DW, Hendricks C, Hamilton HB. Characteristics of the upper airway pressure-flow relationship during sleep. J Appl Physiol 1988; 64:1930-5. 26. Wiegand L, Zwillich CW, White DP. Collapsibility of the human upper airway during normal sleep. J Appl Physiol 1989; 55:1800-8. 27. Spann RW, Hyatt RE. Factors affecting upper airway resistance in conscious man. J Appl Physiol 1971; 31:708-12. 28. Series F, Cormier Y, Desmeules M, La Forge J. Influence of respiratory drive on upper airway resistance in normal men. J Appl Physiol 1989; 66:1242-9. 29. Oliven A, Odeh M, Gavriely N. Effect of
hypercapnia on upper airway resistance and collapsibility in anesthetized dogs. Respir Physio11989; 75:29-38. 30. Bartlett D Jr. Effects of hypercapnia and hypoxia on laryngeal resistance to airflow. Respir Physiol 1979; 37:293-302. 31. Miller MJ, DiFiore JM, Strohl KP, Carlo WA, Martin JR. Decreased supraglottic resistance in response to CO2 rebreathing in infants (abstract). Am Rev Respir Dis 1989; 139:AI82. 32. OlivenA, OdehM, GavrielyN. Effectofsalicylate on upper airway stability and pressure flow: relationship in anesthetized dogs. J Lab Clin Med 1989; 113:8-14. 33. Van de Graaf WB. Thoracic influence on upper airway patency. J Appl Physio11988;65:2124-33. 34. Sauerland EK, Orr WC, Harriston LE. EMG patterns of oropharyngeal muscles during respiration in wakefulness and sleep. Electromyogr Clin Neurophysiol 1981; 21:307-16. 35. Lambert RK, Wilson TA. Flow limitation in a collapsible tube. J Appl PhysioI1972; 33:150-3. 36. Hyland RH, Hutcheon MA, Perl A, et al. Upper airway occlusion induced by diaphragm pacing for primary alveolar hypoventilation: implications for the pathogenesis of obstructive sleep apnea. Am Rev Respir Dis 1981; 124:180-5. 37. Schwartz AR, Smith PL, Wise RA, Gold AR, Permutt S. Induction of upper airway occlusion in sleeping individuals with subatmospheric nasal pressure. J Appl Physiol 1988; 64:535-42. 38. Hudgel DW, Hendricks C, Dadley A. Alteration in obstructive apnea pattern induced by changes in oxygen- and carbon-dioxide-inspired concentrations. Am Rev Respir Dis 1988; 138:16-9.
