Effect of Ethanol on the Arousal Response to Airway Occlusion during Sleep in Normal Subjects1- 3

RICHARD B. BERRY, MICHAEL H. BONNET, and RICHARD W. LIGHT Introduction

Ethanol ingestion worsens obstructive sleep apnea by increasing the frequency and duration of apneas (1-3). The prolongation of apnea results in an increase in the severity of arterial oxygen desaturation. It has been hypothesized that the prolongation of apneas is due to an increase in the "arousal threshold" (1, 3), although the nature of this threshold has not been determined. The termination of obstructive apnea is believed to depend heavily on arousal (4, 5). Arousal can be defined as an abrupt increase in the frequency of electroencephalogram (BEG) activity and a simultaneous increase in the amplitude of the chin electromyographic (BMG) activity (6). Evidenceof arousal usually precedes or coincides with the preferential increasein upper airwaytone that restores airway patency and terminates the obstructive apnea (4). The precise events leading to arousal and apnea termination are still not well understood. Central nervous systemstimulation from a decrease in P02 , an increase in Pe02 , and sensations of inspiring against an occluded airway appear to be additive in producing arousal and apnea termination (5,7-13). The relative importance of these factors is still not clear. Certainly hypoxemia, in the absenceof hypercapnia or airway occlusion, is a poor stimulus for arousal (7). Hyperoxic hypercapnia is a more reliable arousal stimulus, however, normal subjects may require increases of 10 to 15 mm Hg in the end tidal Pe0 2 above the waking level to elicit arousal (8, 9). Alternatively, recent evidence suggests that sensations associated with the degree of inspiratory effort may be important in producing arousal from sleep during hypoxemia, hypercapnia, or airwayocclusion (10-12). Gleeson and coworkers (10) have recently shown in normal subjects that the maximum negative esophageal pressure deflection preceding arousal from sleep in a given subject is similar whether arousal follows induced hypercapnia,

SUMMARY The effect of ethanol on the arouaal response to airway occlusion during non-rapid

eye movement sleep was studied In normal male SUbjects by testing the responaa to the occlusion of a face mask covering the nose and mouth on a control (C) and an ethanol (E) Ingestion (2 ml/kg of 50% vodka) night In random order. In part A, five SUbjects breathed room air while In part B another five SUbjects bresthed a mixture of air and oxygen adjusted to maintain a baseline sleeping sao, of 98%. In both parts, the time to arouaal (TTA) was significantly longer on E nights. The TTA (mean ± SEM) in part A on C versus E nights was 14.6 ± 1.9 versus 20.6 ± 1.4 s In stage 2 and 19.9 ± 1.9 versus 29.2 ± 1.81n stage 3/4 (p < 0.01).The corresponding values In part B were 22.2 ± 3.6 versus 39.9 ± 8.4 In stage 2 and 32.1 ± 4.1 versus 63.7 ± 9.6 In stage 3/4 (p < 0.01). In part B, the maximum deflections In airway prassure were measured at a supraglottic locstlon during airway occlusion to reflect the degree of Inspiratory effort. The maximum airway suction pressure preceding arousal was significantly higher on E nights. Conversely, the rate of Increase In Insplrstory effort (maximum pressure) during occlusion was decreased by E. We conclude that moderate ethanol Ingestion prolongs the time to arousal following airway occlusion by Increasing the threshold of Inspiratory effort associated with arouaal and by decreasing the rate of Increase in the magnitude of Inspiratory efforts. AM REV RESPIR DIS 1992; 145:445-452

hypoxia, or an added resistive inspiratory load. This occurs eventhough the ventilatory responses to these stimuli are very different. These investigators hypothesized that it is the sensations associated with the inspiratory effort that trigger arousal and that the degree of effort can be quantitated by the maximum esophageal pressure deflection before arousal. Thus, arousal occurs when inspiratory effort reaches a certain level (pressure arousal threshold). Studies in patients with obstructive sleep apnea have also suggested that inspiratory effort may be the main stimulus for arousal. Vincken and colleagues (11) proposed that arousal and airway opening in the obstructive sleep apnea syndrome occur when the pressure time index (PTI) of the diaphragm reaches the fatiguing range. However, Wilcox and coworkers (12) found that the PTI at arousal in such patients was not always in the fatiguing range. Although the intersubject range of the PTI at arousal was wider than that found by Vincken and colleagues, each patient tended to arouse at a characteristic PTI. These studies suggest that arousal from sleep occurs when the occluded inspirations reach a certain level of intensity (which may vary between subjects).

Airway (mask) occlusion during sleep in normal subjects results in intermittent inspiratory efforts (negative airway and pleural pressure swings)that progressively increase in amplitude until arousal occurs (13). Based on the above findings, we propose that the time to arousal following airway occlusion could be prolonged either because of an increase in the magnitude of the airway suction pressure required to trigger arousal (arousal pressure threshold), a decrease in the rate of increase of the amplitude of suction pressure (respiratory response to occlusion), or a decrease in the degree of effort on the first breath (figure 1).A combination of factors could also be responsible.

(Received in original form January 11, 1991 and in revised form May 20, 1991) I From the Pulmonary Section, Long Beach VA Medical Center, Long Beach, and the University of California-Irvine, Irvine, California. 2 Presented in part at the American College of Chest Physicians meeting, October 1990, Toronto, Ontario, Canada. 3 Correspondence and requests for reprints should be addressed to Richard B. Berry, M.D., Pulmonary Section lllP, Long Beach VAMedical Center, 5901 East 7th Street, Long Beach, CA 90822.

445

446

BERRY, BONNET, AND LIGHT

arousal

NORMI\L

~ '~,~"'-""

normal initial effort

normal pr ee eure

'"

Ihreshotd

normal ,lope

A

INCREI\SED TIME TO I\ROUSI\L

' !

I •

~

increased pressure threshold

•

B

I ~ ;

.

!

i

t

'''''''''-'':

:

,

c

i

d ecr•••• d Initial effort

""""''''''

decreased

""""

slope

~

r:

Fig. 1. A schematic representation of the possible changes in pressure generation during occluded inspiratory efforts before arousal is depicted. A normal pattern shows arousal once the suction pressure reaches the arousal pressure threshold. The slope of the maximum pressure deflections is I1PmaxlT. Three factors prolonging the time to arousal are (A) an increase in the arousal pressure threshold, (B) a decrease in the slope, or (e) a decrease in the suction pressure on the first breath (initial effort). The three illustrated possibilities assume that only one of the three factors at a time is abnormal. A combination of changes in these factors is also possible.

pressure

The effects of ethanol on the time to arousal (TTA), rate of increase in inspiratory effort, or level of inspiratory effort preceding arousal during airway occlusion have not been studied in normal subjects. Such information could help explain the mechanisms by which ethanol ingestion increases apnea length. We hypothesized that ethanol ingestion should increase the time to arousal following airway occlusion. Furthermore, we hypothesized that either the inspiratory effort preceding arousal should be

greater or the rate of increase in inspiratory effort should be decreased on ethanol nights. To test these hypotheses, we studied the effect of ethanol ingestion on the respiratory and arousal responses to mask occlusion in normal subjects. Methods Ten normal male subjects were studied. The subjects were of normal body weight and did not have a history of snoring, daytime sleepiness, or chronic ethanol ingestion. Subjects were studied on both a control and an etha-

Occlusion Valve

Pneumotachome ter

Mask pressure, End-Tidal

PC02

leak Detector

Fig. 2. The mask occlusion apparatus used in part A.

nol night. The order was randomized. Written informed consent was obtained from all subjectsbeforetheir participation in the study. The project was approved by the Institution Review Board of our hospital. On the ethanol night, subjects ingested 2 ml/kg of 100 proof vodka (50070) mixed in juice over 0.5 h beginning 1 h before bedtime. The ethanol levelwas measured twice at bedtime using an SD-2 alcohol breath analyzer (CMI, Owensboro, KY) and the results were averaged. In part A (n = 5) subjects were studied breathing room air and without an upper airway catheter and topical airway anesthesia. In part B (n = 5) a low flow of oxygen was added to the inspired room air so that the sleeping arterial oxygen saturation was 98%. In addition, a pressuretransducer tipped catheter (Millar, Houston, TX) was inserted via the nose into the upper airway to measure changes in airway pressure. This was necessary as upper airway collapse (14),manifested by a plateau in mask pressures despite a progressiveincrease in inspiratory effort, was noted in three of the five subjects in part A. This made measurement ofthe maximum suction pressure before arousal impossible. In both parts A and B, the presence and stage of sleep was monitored using two pairs of EEG leads (C4-Al, 02-Al), two pairs of electrooculographic leads, the chin EMG leads using standard methods (15). An electrocardiogram (EKG) lead was monitored as well.Arterial oxygen saturation was continuously measured using pulse oximetry (Ohmeda 3700; Ohmeda, Boulder,CO). End tidal Pco, wasrecorded using a capnograph (model 223; Puritan-Bennett, Los Angeles, CA). The mask occlusion apparatus used in parts A and B is depicted in figures 2 and 3, respectively. A face mask covering the nose and mouth was "glued" to the subjects using a medical grade silicone elastomer (Factor II, Lakeside, AZ) and held in place by head straps. A leak detector consisted of a length of plastic tubing containing multiple small holes that surrounded the mask at the maskface interface. This tubing was connected to another CO, monitor (Marquette 7500, Milwaukee, WI). Small mask leaks could be detected as a non-zero reading on the CO, monitor. The face mask contained separateinspiratory and expiratory ports. The expiratory port was fitted with a one-way valve. In part A (figure 2), the inspiratory port also contained a one-way.valve and was connected to a quiet pneumatic valve that could occlude this port. Apneumotachograph was connected to the valve to'. detect inspiratory flow. Mask pressure was measured using a pressure transducer (Validyne Corp., Northridge, CA) connected to a pressure tap on the mask. A catheter to measure end tidal Pco, was also attaehed to the mask. In part B (figure 3), the.inspiratory port did not contain a one-wayvalveand was connected to six feet of low compliance respiratory tubing that passed through the wall of the subject's room into the monitoring room, The tubing was connected to an occlusion

447

ETHANOL AND AROUSAL FROM SLEEP

that time of night would not influence the relative effects of stage 2 and stage 3/4 sleep.

Fig. 3. The mask occlusion apparatus used in part B. The bias flow was turned off before mask occlusion (balloon inflation).

valve system (Hans Rudolph, Kansas City, MO) allowinginspiration of room air through a one-way valve unless a balloon was inflated. Even though the occlusion valve was separated from the mask by tubing, during occlusion no inspiration was possible because of the low compliance of the tubing. A pneumotachograph was positioned external to the balloon and was used to measure inspiratory flow. It was calibrated using a rotameter. A small amount of bias flowconsisting of a mixture of medical grade air and oxygen (1 to 3 Lzmin) was also infused into the inspiratory tubing at a site 30 em from the mask. Because of the one-way valve in the occlusion apparatus,this flow was directed out through the mask expiratory valve. The bias flow was turned off just before the last nonoccluded inspiration before balloon inflation (mask occlusion) . The amount of oxygen was slowly increased until the sleeping arterial oxygen saturation was 98070. Thus, the length of inspiratory tubing served as a reservoir of a mixture of room air and oxygen and the bias flow prevented rebreathing of CO,. The subjects' exhaled air was sampled at the nostrils using a small nasal cannula to determine the end tidal Pco.. Before the mask was attached to each subject's face, a 5 French catheter with a pressure transducer tip wasinserted through one nostril af ter minimal topical anesthesia with 1% lidocaine was administered to one nasal passage and the upper airway. The catheter tip was placei:l16to 18cm from the nares and in this location it was in a supraglottic position. The catheter was calibrated before insertion into the subjects by applying a negative pressure to a small chamber in which the catheter was inserted. The chamber was connected to a water manometer. Under condi-

B ial Flow OJ/ A ir

tions of no flow (mask occlusion), changes in the supraglottic pressure equal changes in pleural (and esophageal) pressure. Mask pressure was measured as in part A. In parts A and B, all variables were recorded on a 12-channel Grass 78D polygraph (Grass, Quincy, MA) using a paper speed of 10 mm/s. A near infrared camera and video monitor system allowed continuous visual inspection of the subjects.

Airway Occlusion Procedure In both parts A and B, the inspiratory airway was closed during expiration in stable stage 2 and 3/4 non-rapid eyemovement (NREM) sleep. Subjects were required to have been in a stable stage of sleep without arousals for at least 3 min before mask occlusion was initiated. The breathing pattern was required to be regular and the end tidal Pea, stable. The occlusion was quickly terminated when there was evidence of arousal from theEEG and EMG tracings. An arousal was defined as an increase in the EEG frequency with appearance of alpha waves and a simultaneous increase in the EMG amplitude (6). Such arousals were invariably associated with an abrupt change in the mask or supraglottic pressure tracings. If arousal occurred on the first inspiratory effort, the data were not included in the analysis. From four to eight occlusions lasting longer than one inspiratory effort were recorded in stage 2 sleep and in stage 3/4 sleep on ethanol and control nights. Occlusions were performed within the first 3 h after lights out and subjects were then allowed to sleep undisturbed in the laboratory for the remainder of the night. On any given night, mask occlusions in stage 2 sleepwere mixed with occlusions in stage 3/4 sleep so

Data Analysis Part A . The TTA following mask occlusion, the nadir in Sao, following mask occlusion , and the Sao, and end-tidal Pea, preceding occlusion on ethanol andnon-ethanol night were compared using the analysis of variance for repeated measures (16). In this analysis, the ethanol versus control and stage 2 versus stage 3/4 conditions were considered within subject variables . Statistical computations were performed using the BMDP program P2V (BMDP Statistical Software, Los Angeles, CA). A value of p < 0.05 was considered to represent statistical significance. Part B The time to arousal and the end tidal Pea, preceding mask occlusion were compared as in part A. As the preocclusion Sao, was set at 98% by adjusting the supplemental oxygen flow, this was not compared. Arterial oxygen desaturation did not occur following airway occlusion; therefore the degree of desaturation was not analyzed . The airway pressure (supraglottic pressure) during occlusion was analyzed in the following manner. The maximum negative pressure deflection of the initial inspiratory effort (PI) and the final complete effort preceding arousal (PF) weremeasured. A complete effort was one in which arousal did not occur before a defmite maximum pressurehad been attained. The difference in the initial and final maximum pressures (~Pmax = PF - PI) was calculated . A mean slope of maximum pressure deflection (~Pmax/T) was determined by dividing ~Pmax by the elapsed time between the maximum pressure deflections on the initial and final efforts (T). The values of PI, PF, and ~Pmax/T were compared using the analysis of variance for repeated measures as described in part A. The ~Pmax/T was further analyzed by factoring this variable into the product of (~Pmax/#) and (#/T), where # is the number of occluded inspiratory efforts following the initial effort before arousal. The (~Pmax/#) is the mean increase in the maximum suction pressure per breath and the (ti/T) is the rate of occluded efforts (num. ber/second). In this way one can see that ~Pmax/T could vary as a result of changes in the mean progressive increase in suction pressure with each succeeding inspiratory effort or changes in the frequency of occluded efforts. Both (APmax/#) and (#/T) wereanalyzed using the same methods as ~Pmax/T. The duration of respiratory efforts during mask occlusion was analyzed as follows. The inspiratory time (11) was defined as the' time from the start of the supraglottic pressure deflection to the maximum negative deflection. The inspiratory time of the initial and final pressure deflections (111 and l1F) were compared using the analysis of variance for repeated measures with sleep stage (2 versus 3/4), drug (control versus ethanol), and time (I versus F) as within subjects factors. The mean rate of pressure generation on the first

448

Subject Part A 1 2 3 4 5 Mean SO Part B 6 7 8 9 10 Mean SO

BERRY, BONNET, AND LIGHT

TABLE 1

EEG

SUBJECT CHARACTERISTICS

(C4A1 )

(inches)

Weight (Ib)

Ethanol Level (mgld/)

21 20.8 0.8

72 68 68 70 73 69.5 2.4

165 142 145 165 225 154.3 20.8

97 126 99 100 135 105.5 12.2

24 20 40 21 19 22.8 6.2

72 71 70 76 74 71.4 2.5

155 170 155 195 140 166.0 26.4

112 118 89 94 107 107.7 14.9

Age (yr)

Height

21 20 20

22

EEG (02 A1 )

Chin

EMG

EKG Vsum

_ _ _ _ _ _ _ _ _ _ _ _......

~_t_ I!

-~=:J.d~

Insp. Airflow (L/sac) Mask Pressure (C mH20 )

The SUbject characteristics and bedtime

Tidal Volume

Sa02

-500cc

1~~r-

__

m

--------

-------

-

Fig. 4. A typical polygraph tracing shOWingan occlusion in part A. The mask pressure increased progressively with each inspiratory effort until arousal occurred.

ethanol levelsare listedin table 1for parts A and B.The ethanol levelsmay not have represented the peak ethanol levels because of variability in the rate of ingestion and absorption.

Part A The time to arousal in part A was significantly (p < 0.01) greater on ethanol nights (table 2). The time to arousal was also greater from stage 3/4 than from stage 2 sleep on both the control and ethanol nights (p < 0.02). Although the arterial oxygen saturation and end tidal

TABLE 2 OCCLUSION RESULTS FROM PART A * Control

%:l:

Pre Pco,. mm Hg§ MIN Sao,.

%11

Ethanol

Stage 2

Stage 3/4

Stage 2

Stage 3/4

14.6 (1.9)

19.9 (1.9)

20.6 (1.4)

29.2 (1.8)

96.0 (0.54)

96.0 (0.45)

94.6 (0.70)

93.9 (0.66)

43.5 (2.1)

43.5 (1.9)

45.2 (1.2)

47.3 (1.3)

93.3 (0.86)

92.6 (1.11)

87.1 (1.8)

84.3 (2.6)

DeHnitionof IJbbrfly/al/on$: pre. preocclusion; MIN Sao, = nadir in Sao, alter occlusion; TTA • lime 10arousal; NS • nol significant • Values are means wKh SEM in parentheses. t Control versus ethanol, p < 0.01; slage 2 versus stage 314, p < 0.02. Not significent, conlrol versus ethanol lIIld stage 2 versus stage 314. § NOI significant. control versus ethanol and stage 2 versus stage 3/4. MControl versus ethanol, p < 0.03; not significant, stage 2 versus stage 314.

*

• • .•

PC02

Results

Pre Sao".

t • • I' "It

End tidal

and final respiratory efforts was computed as (PI/D!) and (PF/nF) and compared using the analysis of variancefor repeated measures. The proportion of occlusion trials resulting in arousal following the first inspiratory effort (stages 2 to 4 combined) was compared between control and ethanol nights using the paired t test. The upper airway resistance at peak flow on the breath before mask occlusion was obtained by dividing the difference between the supraglottic pressure and the mask pressure by the peak flowrate. The bias flow wasturned off just before this breath. When upper airway collapse was noted during the airway occlusions, the pressure at which the mask pressure and supraglottic pressure diverged was noted (upper airway closing pressure). The measurements for stages 2 to 4 were pooled and compared on ethanol and control nights using the paired t test.

ITA.st

~

Pe02 preceding mask occlusion were slightlylower and higher, respectively, on the ethanol nights, these differences did not reach statistical significance. The nadir of Sao, following mask occlusion was significantly lower on ethanol nights. In two of the five subjects, mask occlusion resulted in a progressive rise in the maximum inspiratory pressure until arousal occurred (figure 4). In the other three, the mask pressure frequently reached a plateau early in the mask occlusion especially on ethanol nights. This pattern was noted by Issa and Sullivan to be consistent with upper airway collapse (13, 14). In the two subjects in whom the maximum airwaysuction pressure deflection could be determined before arousal, the values were higher on ethanol nights and in stage 3/4 sleep.The averages of the mean maximum suction pressure before arousal (PF) for the two subjects on the control and ethanol nights, respectively. were 15.4 and 22.4 em H 20 in stage 2 sleep aad 19.1 ane 25.7 em H 20 in stage 3/4 sleep.

Part B The time to arousal was greater on ethanol nights in both stage 2 and stage 3/4 sleep (figure 5). Ethanol ingestion nearly doubled the mean ITA in both stage 2 and stage 3/4 sleep. The effect of ethanol was more pronounced in stage 3/4 sleep (interaction of drug versus sleep

449

ETHANOL AND AROUSAL FROM SLEEP

120

TABLE 3

100

MAXIMUM PRESSURES AND RATES OF CHANGE DURING OCCLUSION·

~

"'

o~-• «

..: ~

o ~

PI (em H2O)

PF (em H2O)

aPmaxlT (em H2O/s)

aP max/# (em H2O/#)

(#/s)

Stage 2 control

12.1 (2.4)

17.4 (3.0)

0.36 (0.04)

1.52 (0.29)

0.25 (0.03)

Stage 2 ethanol

14.8 (1.4)

23.7t (2.0)

O.29t (0.04)

1.10:1: (0.20)

0.27:1: (0.03)

Stage 3/4 control

13.8 (0.89)

23.8§ (1.8)

0.43 (0.08)

1.82 (0.46)

0.26 (0.03)

Stage 314 ethanol

18.1 (1.9)

33.4t§ (3.3)

0.31t (0.05)

1.19:1: (0.25)

0.28:1: (0.02)

80

~

:

60

0

w....

40

~

~

20

Control

Ethanol

STAGE 2

Control

EthanOl

STAGE 3,4

Fig. 5. The mean time to arousal during mask occlusion for each subject on control and ethanol nights in part B. Each open symbol and closed triangle refers to a given subject. The values arelistedseparately for stage 2 and stage 314 sleep. The solid circles are the group means. The group mean (± SEM) values on ethanol nights exceeded those on control nights in stage 2 sleep (39.9 ± 8.4 versus 22.2 ± 3.6) and stage 314sleep (63.7 ± 9.6 vs 32.1 ± 4.1 s) (p < 0.01). The values in stage 314 sleep exceeded those in stage 2 sleep on control (p < 0.05) and ethanol nights (p < 0.01).

stage was significant, p = 0.04). Analysis of the simple effects revealed that the TTA was also significantly longer during stage 3/4 sleep as compared with stage 2 sleep on control (p < 0.05) and ethanol (p < 0.01) nights. The end tidal Pe02 (mean ± SEM) preceding mask occlusion was very similar on control (stage 2,40.2 ± 0.6 mm Hg; stage 3/4, 40.0 ± 0.9) and ethanol (stage 2, 40.6 ± 0.7; stage 3/4, 40.9 ± 1.0) nights. None of these values were significantly different. These preocclusion Pe02 values are slightly lower than those in part A and this may Ravebeen due to the bias flow, which reduced the effective dead space of the

C4 -A, °2- A,

• Values are means with SEM in parentheses.

t Control versus ethanol, p < 0.01.

*Control versus ethanol, p < 0.05. §

Stage 2 versus stage314, p < 0.01.

breathing circuit. The preocclusion Sa02 were all 98 % by experimental design. A typical polygraph tracing for part B is shown in figure 6. There was a progressive rise in the maximum supraglottic pressure swings until arousal occurred even though the mask pressure reached a plateau value. This divergence in mask and supraglottic pressure tracings represents airway collapse above the catheter tip. The maximum supraglottic pressure deflections on the initial occluded breath (PI) were higher on ethanol nights in both stage 2 sleep and stage 3/4 sleep (table 3), although the differences did not quite reach statistical significance (P = 0.06). The values in stage 3/4 sleep were also higher than in stage 2 sleep (p = 0.06).

."I'- PI) was due to an increase in the rate of pressure generation rather than an increase in the duration Of inspiratory effort. The rate of pressure generation on the initial and final inspiratory efforts was greater on ethanol nights and in stage 3/4 sleep. Neither TIl nor TIF differed between control and ethanol nights or between stage2 and stage 3/4 sleep. The resistance at peak airflow across the upper aiway on the breath beforeocelusion (mean ± SEM) was higher on ethanol than control nights in stage 2 sleep (9.7 ± 5.1 versus 6.0 ± 2.3, p =

PIITII (em H2 O/s)

Til

PFITIF (em H2O/s)

• Valuesare meanswith SEM in parentheses.

t PllTlI versus PFITIF, p < 0.01. l Controlversus ethanol, p < 0.02. Stage2 versus stage 314, p < 0.05. IIStage 2 versusStage 314, p < 0.01.

§

0.07) and stage 3/4 sleep (11.0 ± 6.2 versus 7.5 ± 6.0 em H 20/L/s, p = 0.07).

These calculations were not controlled for body position. We compared the resistances previous to the arousal measurements without regard to body position. The upper airway closing pressure (mean ± SEM) on control and ethanol nights (stages 2 and 3/4 combined) was 9.7 ± 1.9 on the ethanol night and 11.8 ± 0.96 em H 20 on the control nights (p = 0.01). Thus, the upper airway had a slightly increased tendency to collapse on ethanol nights. Discussion

This study shows that ethanol ingestion' impairs the arousal response to airway occlusion in normal subjects such that the time to arousal following airway occlusion is prolonged during NREM sleep. In part A, we made no attempt to control the preocclusion Sao 2 • However, despite slightly lower Sao, and higher Pco, values before occlusion, our subjects had longer TTA on ethanol nights. The nadir in oxygen saturation following mask occlusion was greater on ethanol nights, showing that the prolongation in the time to arousal was clinically significant. In part B, the effects of hypoxemia were eliminated with supplemental oxygen and the preocclusion Pe0 2 values were essentially the same on control and ethanol nights. As in part A, the time to arousal was prolonged on ethanol nights. In parts A and B, the time to arousal was prolonged in slow-wavesleep (stage 3/4) compared with stage 2 sleep. In part B, ethanol ingestion increased the maximum suction pressure attained before arousal occurred and decreased the rate of buildup of maximum suction

pressure with time (decreased aPmax/T). Thus, according to the scheme in figure 1, these changes would increase the time to arousal. The maximum suction pressure preceding arousal was also higher in stage 3/4 than in stage 2 sleep. The latter finding is consistent with increased arousal thresholds from many stimuli in slow-wave sleep compared with stage 2 sleep (17). The reduction in aPmax/T induced by ethanol was not due to a decrease in the frequency of inspiratory efforts, but rather a decrease in the increment in suction pressure with each effort. The time needed to reach a given PF with a given aPmax/T also depends on the maximum pressure of the initial occluded effort (see figure 1). The fact that the PI was slightly but not significantly higher on ethanol nights should have decreased rather than increased the time to arousal. Therefore, changes in PIon ethanol nights did not contribute to the increase in the TTA. The higher PIon ethanol nights may have been due to an increase in ventilatory . drive caused by higher upper airway resistance (18, 19). For example, ventilatory drive as estimated by the occlusion pressure (20) during sleep is higher in snorers (upper airway narrowing) than in non-snorers (21). However, although ethanol ingestion may'have slightly increased the baseline ventilatory drive because of resistive loading, it decreased the rate of increase in maximum suction pressure during airway occlusion (aPmax/T). Ethanol ingestion did not appear to significantly affect the duration of occluded respiratory efforts (table 4). Like Issa and Sullivan (13), we found that the mean rate of inspiratory pressure generation during the final occluded effort

451

ETHANOL AND AROUSAL FROM SLEEP

(PF/TIF) was greater than that on the initial occluded breath (PI/TIl). As TIl and TIF did not differ, increases in the maximum suction pressure (PF >PI) weredue to a higher mean rate of inspiratory pressure generation during the final inspiratory effort. The fact that PF/TIF was greater on ethanol nights and in stage 3/4 sleep was undoubtedly due, at least in part, to the longer time of occlusion in these conditions. The rate of pressure generation was also higher on the initial occluded breath (PI/TIl) on ethanol nights and in stage 3/4 sleep. This may have occurred for the same reasons that PI increased on ethanol nights as discussed previously. It is possible that ethanol could have direct effects on ilPmax/T and the arousal threshold as well as indirect effects on these factors via changes in the baseline sleeping P0 2or Pe02. Although the effects of acute increases in Pe02 and decreases in P0 2 on the ilPmax/T and arousal threshold are not known, they should either have no effect or tend to increase the ilPmax/T and decrease the arousal threshold. In our study, determination of the Po 2 and Pe02 at the time of arousal was not possible because of our noninvasive methods. Subjects invariably inspired deeply the moment the mask occlusion was released, altering the end tidal Pe02. However, despite the fact that the baseline oxygen saturation was slightly lower and end tidal Pe02 slightly higher on ethanol nights in part A, the time to arousal was longer. In part B, the effects of hypoxemia on arousal were eliminated with supplemental oxygen. The Pe02 at arousal on ethanol nights should have been higher as the preocclusion end tidal Pe02 was essentially the same on control nights but the occlusion time was longer. This assumes that the rate of increase in Pe0 2with time did not differ on control and ethanol nights. Awake CO 2 production (22) is not changed by ethanol ingestion, so that it is unlikely that sleeping production would differ. In any case, a higher Pe02 would be expected to increase the ilPmax/T and decrease the arousal pressure threshold. The fact that the converse was true is evidence that direct effects of ethanol on the arousal process predominated over possible effects from changes in Pe02. The TTA values in part B were longer than part A. There are several possible explanations. First, the subjects in part A and B were different persons. However, the two groups were quite similar

in age and bedtime ethanol levels. Second, hypoxemia was not a stimulus for arousal in part B. Oxygen therapy is known to prolong apneas in some patients with the obstructive sleep apnea syndrome (23,24). Another explanation for the longer TTA values in part B, is that upper airway anesthesia was used in this portion of the study. Upper airway sensation is important for maintaining airway patency and detecting airway occlusion (25, 26). For example, Basner and coworkers showed that the time to arousal after mask occlusion increased from a mean of 11 to 25 s after subjects had the nose and oropharynx anesthetized with 4070 lidocaine (26). Although upper airway anesthesia could have produced the longer TTA in part B, we do not believe the use of lidocaine affected our conclusions regarding the effect of ethanol on arousal. First, similar effects of ethanol on the time to arousal were noted as in part A, in which no anesthesia was used. Second, occlusions were performed at least 1 h after anesthesia was applied (topical lidocaine has a duration of action of 1to 2 h). Third, our application of anesthesia was quite selective and one nasal passage was completely free from anesthesia and the bulk of lidocaine was delivered to the area where the tip of the supraglottic catheter was located. Also, a much smaller amount of topical anesthesia was used (5 ml of 1070 lidocaine) than in the study mentioned previously. Fourth and most important, the same method of anesthesia was used on both ethanol and control nights. Thus, upper airway anesthesia could have prolonged the time to arousal on both control and ethanol nights, but it is unlikely to have differentially affected the TTA on ethanol and control nights. In our study, we noted a small percentage of arousals on the first inspiratory effort. These occurred almost entirely in stage 2 sleep. The frequency of these events was not different on control and ethanol nights. The maximum pressure on the inspiratory effort of these short latency arousals on either the ethanol or control nights was much lower than the mean PF of the prolonged occlusions. Such short latency arousals may be due to different stimuli such as sound (closing valves) or other sensations (sudden absence of bias flow). In any case, when we reanalyzed our data for TTA and PF including these short latency arousals, none of our conclusions were altered. Upper airway collapse was noted in our

normal subjects, even though these subjects did not snore or develop apnea on either the control or ethanol nights. Issa and Sullivan have reported upper airway collapse in snorers during airway occlusion (14)but not in normal subjects (13). Alternatively, Schwartz and coworkers (27) induced upper airway obstruction in normal subjects with a mean negative nasal mask pressure of 13.3em H 20 . This value is very close to the mean closing pressure of 11.8 em H 20 found in part B of our study on the control night. The most likely explanation for the presence of airways collapse in the present study is that our method of airway occlusion differed from the one used by Issa and Sullivan (13). They used a valveless system with a high bias flow rate and occluded the airway by simultaneous inflation of balloons on both the inspiratory and expiratory sides of the nose mask at or above functional residual capacity. Because it is difficult to time such occlusions precisely at end expiration, the airwaywas probably occluded slightly above functional residual capacity. It is also possible that a small amount of continuous positive airway pressure was present in their system. These factors should have increased upper airway volume at the time of occlusion and thus increased the resistance of the airway to collapse (28). In our study, ethanol ingestion did increase the tendency of the airway to collapse (decreased upper airway closing pressure). This finding has been reported in snorers (14). This raises the possibility that the effect of ethanol on arousal was due to an increased tendency of upper airway to collapse and the resulting lack of transmission of negative inspiratory pressure to sensory receptors above the site of collapse. However, the closing pressure on both the control and ethanol nights was less than PF especially in stage 3/4 sleep. Thus, on both control and ethanol nights, airway collapse began to occur during occluded inspiration wellbefore the one triggering arousal (before PF was reached). Yet,the differences in the time to arousal and PF between the ethanol and control conditions were most pronounced in stage 3/4 sleep. Therefore, the effect of ethanol on arousal is unlikely to be explained by changes in the tendency in the upper airway to collapse. In summary, our findings document that ethanol ingestion impairs the respiratory and arousal responses to airway occlusion so that the time to arousal following airway occlusion is prolonged. In

452

BERRY, BONNET, AND LIGHT

addition, this study provides data consistent with a model of arousal from airwayocclusion in which the time to arousal may be delayed by an increase in the inspiratory pressure arousal threshold and a decrease in the rate of increment in inspiratory effort with time. References 1. Guilleminault C, Rosekind M. The arousal

threshold: sleep deprivation, sleep fragmentation, and obstructive sleep apnea syndrome. Bull Eur Physiopathol Respir 1981; 17:341-9. 2. Scrima L, Broudy M, Nay KN, Cohn MA.lncreased severityof obstructive sleep apnea after bedtime alcohol ingestion. Sleep 1982; 5:318-28. 3. Issa FG, Sullivan CEo Alcohol, snoring, and sleep apnea. J Neurol Neurosurg Psychiatry 1982; 5:353-9. 4. Remmers JE, Degroot WJ, SauerlandEK, Anch AM. Pathogenesis of upper airway occlusion during sleep. J Appl Physiol 1978; 44:931-8. S. Phillipson EA, Sullivan CEo Arousal: the forgotten response to respiratory stimuli. Am Rev Respir Dis 1978; 1I8:807-9. 6. West P, Kryger MH. Sleep and respiration: terminology and methodology. In: Kryger MH, ed. Clinics in chest medicine: sleep disorders. Philadelphia: W. B. Saunders, 1985; 692. 7. Phillipson EA, Sullivan CEo Ventilatory and arousal responses to hypoxia in sleeping humans. Am Rev Respir Dis 1982; 125:632-9. 8. Berthon-Jones M, Sullivan CEo Ventilation and arousal responses to hypercapnia in normal sleeping humans. J Appl Physiol 1984; 57:59-67.

9. Douglas NJ, White DP, Weil JV, Pickett CK, Zwillich CWo Hypercapnic ventilatory response in sleeping adults. Am Rev Respir Dis 1982; 126: 758-62. 10. Gleeson K, Zwillich CW, White DP. Arousal from sleep in response to ventilatory stimuli occurs at a similar degree of ventilatory effort irrespective of the stimulus. Am Rev Respir Dis 1989; 142:295-300. 11. Vincken W, Guilleminault C, Silvestri L, Cosio M, Grassino A. Inspiratory muscle activity as a trigger causing the airways to open in obstructive sleep apnea. Am Rev Respir Dis 1987; 135:372-7. 12. WilcoxPG, ParePD, Road JD, F1eetham JA. Respiratory muscle function during obstructive sleep apnea. Am Rev Respir Dis 1990; 142:533-9. 13. Issa FG, Sullivan CEoArousal and breathing responses to airway occlusion in healthy sleeping adults. J Appl Physiol 1983; 55:1ll3-9. 14. Issa FG, Sullivan CEo Upper airway closing pressures in snorers. J Appl Physio11984;57:528-35. 15. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques, and scoring system for sleep stages in human adults. Los Angeles: Brain Information Service/Brain Research Institute, University of California, 1968. 16. Keppel G, Design and analysis. Englewood Cliffs, NJ: Prentice Hall, 1982. 17. Bonnet MH. Performance during sleep. In: Webb WB, ed: Biological rhythms, sleep, and performance. New York: John Wiley & Sons, 1982; 205-37. 18. Gugger M, Molloy J, Gould GA, et al. Ventilatory and arousal responses to added inspiratory resistance during sleep. Am Rev Respir Dis 1989; 140:1301-7. 19. Robinson RW, White DP, Zwillich cs: Moderate alcohol ingestion increases upper airway

resistance in normal subjects. Am Rev Respir Dis 1985; 132:1238-41. 20. Whitlaw WA, Derenne JP, Milic-Emili J. Occlusion pressure as a measure of respiratory center output in conscious man. Respir Physiol 1975; 23:181-99. 21. White DP. Occlusion pressure and ventilation during sleep in normal humans. J Appl Physiol 1986; 61:1279-87. 22. Johnstone RE, Reier CEo Acute respiratory effects of ethanol in man. Clin Pharmacol Ther 1983; 14:501-8. 23. Motta J, Guilleminault C. Effects of oxygen administration in sleep induced apnea. In: Guilleminault C, Dement WC, eds. Sleep apnea syndromes. New York: Alan R. Liss, 1978. 24. Martin RJ, Sanders MH, Gray BA, Pennock BE. Acute and long-term ventilatory effects of hyperoxia in the adult sleep apnea syndrome. Am Rev Respir Dis 1982; 125:175-80. 25. McNicholas WT, Coffey M, McDonnell T, O'Reagan R, Fitzgerald MX. Upper airwayobstruction during sleep in normal subjects after selective topical oropharyngeal anesthesia. Am Rev Respir Dis 1987; 135:1316-9. 26. Basner R, Ringler J, Garpestad E, Schwartzstein R, Weinberger S, Weiss JW. Upper airway anesthesia prolongs the time to arousal from airway occlusion induced during NREM sleep (abstract). Am Rev Respir Dis 1990; 141:AI27. 27. Schwartz AR, Smith PL, Wise RA, Gold AR, Permut S. Induction of upper airway occlusion in sleeping individuals with subatmospheric nasal pressure. J Appl Physiol 1988; 64:535-42. 28. Olson LG, Ulmer LG, Saunders NA. Pressurevolume properties of the upper airway of rabbits. J Appl Physiol 1989; 66:759-73.