Responses of Diaphragm and External Oblique Muscles to Flow-resistive Loads During Sleep1-3

DOUGLAS A. HUTT, RICHARD A. PARISI, NORMAN H. EDELMAN, and TEODORO V. SANTIAGO

Introduction

During wakefulness, rapid neuromuscular compensation for an added inspiratory resistance has been shown to occur by severalmechanisms that serveto maintain tidal volume. Inspiratory muscle activity may be augmented directly through increased neural activation as reflected by amplitude of the electromyogram (EMG) or by prolongation of inspiratory time (1, 2). Inspiratory force generation also may be enhanced through increased diaphragm operating length at end expiration by recruitment of expiratory muscles (3). During slow-wave sleep (SWS) and rapid eye movement (REM) sleep, compensatory increases in respiratory output have generally been found to be diminished or absent (4-6). However, output of the respiratory system has usually been measured as tidal volume (VT) or early inspiratory occlusion pressure (PO.!), which may not have been sensitive to each of the individual mechanisms potentially involved in this response. Furthermore, it is difficult in some cases to differentiate mechanoreflex responses to a load from subsequent responses to stimulation of central and peripheral chemoreceptors. Reflex load compensation may be especially important during the various sleep states in which decreased activity of upper airway dilator musclesleads to increased pharyngeal resistance and collapsibility (7, 8). The present study was designed to specifically examine the immediate (mechanoreflex) response to an inspiratory flow-resistive load during sleep byassessing individual components of the response, including changes in inspiratory and expiratory EMG activity, and respiratory timing. Methods Preparation Six male goats were surgically prepared under general anesthesia in two stages. Two weeks before the studies, electrodes for monitoring sleep-wake state, including electroencephalogram (EEG), electrooculogram

SUMMARY Although it Is generally agreed that rapid respiratory compensation for externally applied Inspiratory loads Is Impaired or absent during sleep, the individual components of the "loadcompensating reflex" may not be inhibited by sleep to the same degree. We studied the effect of inspiratory flow-resistive loading (18 em H20/L/s) for two consecutive breaths on Inspiratory (diaphragm) and expiratory (external oblique) muscle activity, and respiratory timing, In six awake and sleeping goats. During the first loaded breath In the awakestate, peak Integrated diaphragmatic electromyogram activity (EMGdl) Increased 16.7 ± 3.9% (p < 0.01), peak Integrated external oblique EMGactivity (EMGeo) Increased 21.0 ± 7.5% (p < 0.001), and electrical inspiratory time (TI) Increased 18.1 ± 2.1% (p < 0.01). In contrast, loading did not significantly change peak EMGdl or EMGeo on the first or second breaths In any sleep state. However, TI was significantly Increased during loading In all sleep states (p < 0.01) to a similar degree seen during wakefulness. loading did not significantly alter electrical expiratory time. No significant differences were noted between the first and second loaded breaths. We conclude that the reflex Increases in peak EMGof both Inspiratory and expiratory muscles In response to Inspiratory flow-resistive loading during the awakestate are absent during all stages of sleep; however, one aspect of load compensation, prolongation of TI, Is preserved during sleep and aids in maintaining tidal volume. AM REV RESPIR DIS 1991; 144:1107-1111

(EOG), and EMG were implanted by a previously described technique (9). The electrodes for the EEG and EOG consisted of stainless steel, self-tapping screws and were placed in the roof of the bony orbit (EOG) and the floor of the medial frontal sinus and in the parietal and interparietal bones (EEG); paired stainless steel electrodes for tonic, nonrespiratory EMG monitoring were implanted in the cleido-occipital muscle. A heat-sealed polyethylene catheter (PE190;Clay Adams, Parsippany, NJ) was passed into the aorta via the renal artery 1 wk later, as previously described (9). The catheter was secured to the skin by a cutaneous exteriorization button. The diaphragm was exposed through a midline abdominal incision, and paired gold wire electrodes 5 mm apart were sutured to the costal diaphragm near the central tendon. Similar electrodes also were sutured into the external oblique muscle through a separate incision in the lateral abdominal wall.

Conduct of the Studies Goats were kept on a reverse sleep-wake cycle and the experiments were performed during the day when the animals normally slept. They were not permitted to nap during the 6 h before the study. During the studies, the goats were restrained by the horns in the sternalrecumbent position with the angle of neck extension fixed at approximately 120°. A snugfitting mask was placed over the snout; this mask was previouslyfound not to significantly

alter baseline diaphragmatic EMG activity (10).Ventilation was measured by a pneumotachygraph (Fleisch no. 2; Dynasciences, Blue Bell, PAl positioned in the inspiratory side of a Rudolph valve and connected to a differential pressure transducer (Validyne BC158;ValidyneEngineering Corp., Northridge, CAl. Volume calibrations were made with a l-L syringe. Expired COl was continuously measured with an infrared COl analyzer (Godart Capnograph, Bilthoven, Holland). Arterial blood samples obtained from the catheter were analyzed for pH, Pco., and Pol using an automated blood gas analyzer (ABL30; Radiometer, Copenhagen, Denmark). Sleep-wakestate, ventilation, end-tidal Pco., blood pressure, and diaphragmatic and

(Received in originalform October 2, 1990and in revised form March 21, 1991) 1 From the Division of Pulmonary and Critical Care Medicine, Department of Medicine, Robert Wood Johnson Medical School, Universityof Medicine and Dentistry 0 f New Jersey, New Brunswick, New Jersey. 2 Supported by Research Grants No. HL-01549 and HL-16022, and Training Grant No. HL-07467 from the National Heart, Lung, and Blood Institute. 3 Correspondence and requests for reprints should be addressed to Richard A. Parisi, M.D., Division of Pulmonary and Critical Care Medicine, UMDNJ-Robert Wood Johnson Medical School, One Robert Wood Johnson Place-CN 19, New Brunswick, NJ 08903-0019.

1107

1108

HUTT, PARISI, EDELMAN, AND SANTIAGO

TABLE 1

external oblique EMG activities werecontinuously monitored.

Sleep Staging Wakefulness,SWS,and REM sleepwereidentified by standard criteria (9). REM sleep was further characterized as ''phasic REM" (pREM; presence of rapid eye movements) or "tonic REM" (TREM; absence of eye movements). All measurements were obtained when a particular sleep stage was unequivocally ascertained. Respiratory Muscle EMG Analysis The diaphragmatic and external oblique EMG signals were amplified and band-pass filtered from 20 to 500 Hz, full-wave rectified, and processed by a Paynter filter with a time constant of 200 ms to generate a moving average (Charles Ward Enterprises, Ardmore, PA). The diaphragmatic EMG (EMGdi) signal was quantified as the peak inspiratory moving average and as the rate of rise of inspiratory activity (peak/Tr). The external oblique EMG (EMGeo) signal was quantified as the peak expiratory moving average. "Electrical" inspiratory time (11) was measured as the period from the onset of incrementing EMGdi activity until the peak of the moving average. Expiratory time (Ts) was measured from the EMGdi peak to the beginning of the next inspiration. "Mechanical" inspiratory and expiratory times were taken from the inspiratory flow recording. Because baseline EMGeo varied with small changes in body position, only periods with stable active expiratory EMGeo activity were used. Responses to Loading The flow-resistive load placed in the inspiratory line consisted of a series of fine wiremesh disks and measured approximately 18 em H 2 0 / L / s at a flow rate of 1 Lis. When a particular sleep state could be clearly defined, the load was intermittently applied for two consecutive breaths. There were at least five consecutive unloaded breaths between load applications. Between five and 20 observations were obtained in each sleep-wake state for each animal. Arousals or state changes caused by application of the load were infrequent. Each measurement was expressed as a ratio of the value obtained on a particular loaded breath to the average value obtained from the three prior nonloaded breaths (baseline). Statistical Analysis All data were analyzed by using either a generallinear model for unbalanced analysis of variance (ANOVA) or a one-way ANOVAand the Scheffe, Duncan, and student's ttests (11).

Results Baseline VT, electrical 11, Ts, and arterial blood gases for all six animals are shown in table 1;these values are consistent with previous reports in this species (9, 10). VT decreased progressively from wake-

VENTILATORY MEASUREMENTS DURING SLEEP Awake

336 1.13 1.89 7.45 32.1 99.7

VT, ml TI, s TE, s pHa, mm Hg Pac0 2 , mm Hg Pa02, mm Hg

± ± ± ± ± ±

SWS

51 0.14 0.13 0.02 0.8 2.3

290 1.15 1.96 7.43 33.5 100.5

± ± ± ± ± ±

48' 0.14 0.11 0.02 1.0 2.5

TREM

PREM

± ± ± ± ± ±

240 ± 45,ft 1.09 ± 0.12 1.93 ± 0.21 7.40 ± 0.03' 36.9 ± 1.3,t 93.1 ± 3.3,t

275 1.11 1.95 7.40 36.3 94.5

41,t 0.12 0.18 0.03' 1.3,t 3.1,t

Definition of abbreviations: VT = tidal volume; TI = electrical inspiratory time; TE = electrical expiratory time; pHa = arterial pH; Paco 2 = arterial partial pressure of CO 2; Pao2 = arterial partial pressure of 02; SWS = slow-wave sleep; TREM = tonic rapid eye movement sleep; PREM '" phasic rapid eye movement sleep. * p < 0.05 versus awake. t p < 0.05 versus SWS. :j: P < 0.05 versus TREM.

fulness to SWS to TREM to PREM (p < 0.05). Pao2 was significantly reduced (p < 0.05), whereas Paco2 was increased (P < 0.05) during both TREM and PREM compared with SWS and wakefulness. Figure 1 shows representative recordings from one animal illustrating typical, consistently observed EMGdi and EMGeo responses to an inspiratory flowresistive load. Figure lA represents the awake state during which imposition of the load produced a reduction in VT and an increase in peak inspiratory EMGdi and expiratory EMGeo on both the first and second loaded breaths. This increased peak EMGdi and EMGeo activity in response to loading was not observed during SWS (figure IB) or TREM

(figure l'C). A consistent finding, shown in figure 2, was the absence of expiratory EMGeo activity during PREM sleep periods with or without the load. In figure 3, the effects of loading on peak moving average inspiratory and expiratory EMG activity, and electrical 11 and Ts are shown, represented as the ratio of loaded to unloaded values. Application of the load during wakefulness increased peak EMGdi (figure 3A) 16.7 ± 3.9010 (p < 0.01) and 15.3 ± 5.4% (p < 0.01) for the first and second loaded breaths, respectively. This was accompanied by a significant increase in 11of 18.1 ± 2.1010 (p < 0.01) and 22.0 ± 2.3% (p < 0.01) on the first and second loaded breaths, respectively (figure 3B). In con-

c

B

A EEG

*

It,. Ii hi' I

1

'tw' • Ll ., II; aVril

EOG

E~O

-5see_

J ~

t

Fig. 1. Representative recordings demonstrating diaphragmatic (EMGdi) and external oblique (EMGeo) responses to inspiratory flow-resistive loading. The traces shown are the electroencephalogram (EEG), cervical electromyogram (EMG), electrooculogram (EOG), EMGeo and EMGdi moving averages, and tidal volume (VT). During wakefulness (panel A), application of an 18 cm H20fUs flow-resistive load (arrow) caused an increase in peak EMGdi and EMGeo for both loaded breaths. During SWS (panel B) and TREM (panel C) sleep, the increase in peak EMGdi and EMGeo during loading was not present.

INSPIRATORY AND EXPIRATORY MUSCLE RESPONSES TO RESISTIVE LOADS DURING SLEEP

1109 1.1 "0

III

"80 §

Er1G

l

o

0 0

1.0

0.9

0.8

w ~

EOG

:3

a > .....J -c D

-''''-

i=

0.7 0.6

0.5

AWAKE

SWS

TREM

PREM

SLEEP STATE

Fig. 4. Effect of loading on tidal volume (VT). Imposition of the load caused a decrease in VT in all sleep states for both loaded breaths (p < 0.001). Although VT decreased more during sleep (especially PREM), these differences were not significant. Open bars = first loaded breath; hatched bars = second loaded breath.

Fig. 2. Representative recordings during PREM sleep. The traces shown are the same as in figure 1. The EOG reveals rapid eye movements in a typical burst pattern surrounded by periods of less frequent rapid eye movements, showing an episode of phasic REM (PREM) between adjacent periods oftonic REM (TREM). During PREM, there was an abrupt loss of EMGeo phasic activity which promptly resumed with the return of TREM.

trast to the awake state, loading produced no significant peak EMGdi response during SWS, TREM, or PREM. However, 11was prolonged with loading during all sleep states to a similar degree as during the awake state. There were no significant differences between the first and second loaded breaths for any state. Results for peak EMGeo activity (figure 3C) were similar to peak EMGdi activity. While

Fig. 3. Effect of loading on peak diaphragmatic EMG activity (EMGdi), electrical inspiratory time (TI), peak external oblique EMG activity (EMGeo), and electrical expiratory time (Te) during wakefulness, SWS, TREM, and PREM sleep. Open bars represent the first loaded breath; hatched bars denote the second loaded breath. In panel A, loading significantly increased peak EMGdi during wakefulness (p < 0.01) for both loaded breaths without a significant difference between the loaded breaths. TI (panel B) was found to be significantly increased in each sleep-wake state (p < 0.05 for each state) by loading but not significantly different between states. Similar to EMGdi, loading signif· icantly increased peak EMGeo during wakefulness (p < 0.005) but not during sleep (panel C). In contrastto TI, Te(panel D) was not significantly altered by loading. Open bars = first loaded second loaded breath; hatched bars breath. All data are mean ± SE.

awake, peak EMGeo increased 21 ± 7.50/0 (p < 0.005) and 27.3 ± 6.90/0 (p < 0.005) for the first and second loaded breaths, respectively. Again, no response was observed during any sleep state and there was no difference between the first and second loaded breaths. Ts was not significantly affected by load imposition during any sleep-wake state (figure 3D). VT, shown in figure 4, was significantly

A.'·4

8. '.4

1.3

]1.3 "0 C

o

c:

-Z,.2 "0

ell

-g

0'.'

.;:;.

0.8

0.11

AWAKE

SWS

lllEll

0.11

PRDI

AWAKE

SWS

TIlEll

PREll

NIIAKE

SW5

lllEll

PREll

D.,.4

C.1.4

1.3

=

0.11

AWAKE

SWS

lllEll

PREll

0.11

SLEEP STATE

reduced in all sleep states for both loaded breaths (p < 0.001). However, there were no significant state-related differences in this reduction. Because electrical and mechanical 11 differ, and because loading may have a differential effect on these measurements, we also measured the effect of loading on the relationship between electrical and mechanical 11 (figure 5). We found that mechanical 11 was always greater than electrical 11,and loading did not systematically alter their relationship. Discussion Based on previous studies, it is generally agreed that certain mechanisms responsible for rapid adjustments in respiratory output in response to inspiratory flowresistive loading in the awake state are impaired or absent during sleep (4-6). This study documents that a heretofore unstudied and potentially important load compensating mechanism, augmentation of expiratory abdominal muscle activity, is abolished during SWS and TREM. In addition, we document that the expiratory abdominal muscles cannot contribute to load compensation during PREM because they are atonic in that state. However, a major finding of this study is that the immediate load-induced prolongation of 11remains intact during SWS and REM sleep. We found that flow-resistive loading resulted in an immediate, i.e., mechanoreflex-mediated increase in peak EMGdi activity during wakefulness but not during SWS. This is in agreement with a previous report by Hudgel and coworkers (8). We have extended these observations by demonstrating that flow-resistive loading did not cause an increase in peak EMGdi during tonic or phasic REM

1110

HUTT, PARISI, EDELMAN, AND SANTIAGO

1.5

1.4

1.3

MECHANICAL Ti (sec!

1.2

1.1

I 1.3

1.2

I 1.4

i

1.5

ELECTRICAL Ti (secl Fig. 5. Relationship between mechanical and electrical inspiratory time (TI). The line of identity is shown. Open symbols represent control measurements; solid symbols represent loaded breaths in each state. Mechanical TI was greater than electrical TI during all sleep states. Note that both mechanical TI and electrical TI increased during loading to a similar degree. Data are mean ± SE. Circles = PREM; triangles = SWS; squares = TREM; diamonds = awake.

sleep. This finding supports the concept that the mechanism responsible for the increase in the EMGdi amplitude response to a load is unique to the awake state. We also found in this study that a mechanically important expiratory muscle (external oblique) responded in a similar fashion to the diaphragm in that augmentation during loading was seen only in the awake state. Weevaluated abdominal expiratory as wellas inspiratory muscle activity during load imposition because abdominal muscle activity may enhance inspiratory diaphragmatic force output by altering diaphragmatic operating length (12). This phenomenon was previously observed by Lopata and coworkers (13), who reported that inspiratory flow-resistive loading increased expiratory gastric pressures and decreased abdominal end-expiratory anteroposterior dimensions when compared with the unloaded state in awake human subjects. The effectiveness of this reflex in maintaining VT for two breaths following abrupt load imposition was not clearly apparent in our study, as VT did not differ between the first and second loaded awake breaths despite enhancement of EMGeo after the first loaded breath. Furthermore, there was no significantly greater reduction of Vr during sleepwhen this reflex was not active. A change in end-expiratory length could not affect the first loaded inspiration because the load could not have been sensed during the previous expiration. It is likely that variability in the VT results partly explains

our findings, as a trend toward greater decreases produced by loading during sleep relative to wakefulness was actually present, The other notable finding of this study was that the immediate prolongation of inspiratory duration (11) by flow-resistive loading, which is seen during wakefulness, is preserved during all sleep states. This is in contrast to the findings of other investigators that sleep abolishes or substantially diminishes the increase in 11during loading (6, 8, 14). The reasons for these differences are unclear, but a major issue involves the magnitude of the increase in 11 in the awake state. Using a resistiveload of similar magnitude, Iber and coworkers (6) found that during the awake state the immediate augmentation in 11 was on the order of 40 to 50070, whereas during non-REM sleep the increase in 11was only 5 to 10070. Weigand and coworkers (14)also reported a small (relativeto awake)but significant increase in 11 during sleep. Thus, the principal difference between their results and the present study is the magnitude of 11 prolongation by loading in the awake state because the increases we observed during sleep weresimilar. Axen and Haas (15) have demonstrated that the awake first breath ventilatory responses to flowresistive loading in naive human subjects are extremely variable and the changes observed in 11 form a continuum from marked increases to marked decreases. Thus, a significant conscious influence may account for the variability in this first breath response during wakefulness,

leading to some of the reported differences in the prolongation of 11. It is possible that our goats wereless likely to use varying conscious strategies than previously studied humans. Alternatively, it may be that the methodology used in measuring 11 may have contributed to these differences. In the present study, TI was measured from the moving average EMGdi. Other studies have measured mechanical TI from airflow recordings. It is known that mechanical TI exceeds electrical TI, particularly during loading due to an increased respiratory system time constant (16).Although loading did not systematically alter the relationship between mechanical and electrical Tt in our experiments, this may have occurred during previous studies in which EMGdi was not recorded. Finally, many of the earlier studies were performed in humans, and species difference may be relevant. Humans have a relatively weak Hering-Breuer reflex (17);thus, our findings regarding the prolongation of TIduring loading may have been related to the strength of the Hering-Breuer reflex in the goat. We observed that phasic expiratory activity of the external oblique muscle was suppressed during PREM sleep. Similar inhibition of EMG activity has been reported for other muscles with phasic respiratory activity, including the genioglossus (10) and the external intercostals (18)during PREM sleep. Thus, it appears that during PREM sleep not only is another load-compensating mechanism (enhancement of expiratory muscle activity) lost but, perhaps more important, these muscles are atonic during this state. This implies that during PREM sleep the ability of the expiratory muscles to adjust end-expiratory lung volume is lost. This mechanism may be partly responsible for the significant reduction in VT seen during REM sleep (9, 19), at least under conditions of increased airway resistance when expiratory muscles are most active. The major difference between the EMGdi responses to loading during the awake and sleep states is that the rate of I rise of the EM Gdi (peak/Tt) was preserved during wakefulness but reduced during all sleep states. This is further illustrated in figure 6. Decreased rate of rise of EMGdi activity during loaded breaths while asleep may be partly explained by a study by Pack and coworkers (20), who showed that in anesthetized animals the rate of rise of the phrenic neurogram decreased when airflow rates delivered by the ventilator were reduced,

INSPIRATORY AND EXPIRATORY MUSCLE RESPONSES TO RESISTIVE LOADS DURING SLEEP

~

1.1

~

"0

s

~

1.0