Vol. 89, Issue 1, 120-130, July 2000
Characterization of pharyngeal resistance during sleep in a
spectrum of sleep-disordered breathing
R.
Tamisier1,2,
J. L.
Pepin2,
B.
Wuyam2,
R.
Smith2,
J.
Argod2, and
P.
Levy2
1 Department of Respiratory Medicine, University Hospital,
Nice, and 2 Sleep Laboratory, PRETA-TIMC, UMR CNRS 5525,
Grenoble, France
 |
ABSTRACT |
Aims of the study were 1)
to compare Hudgel's hyperbolic with Rohrer's polynomial model in
describing the pressure-flow relationship, 2) to use this
pressure-flow relationship to describe these resistances and to
evaluate the effects of sleep stages on pharyngeal resistances, and
3) to compare these resistances to the pressure-to-flow
ratio (
P/
). We studied 12 patients: three with upper airway
resistance syndrome (UARS), four with obstructive sleep hypopnea
syndrome (OSHS), three with obstructive sleep apnea syndrome (OSAS),
and two with simple snoring (SS). Transpharyngeal pressures were
calculated between choanae and epiglottis. Flow was measured by use of
a pneumotachometer. The pressure-flow relationship was established by
using nonlinear regression and was appreciated by the Pearson's square
(r2). Mean resistance at peak pressure (Rmax)
was calculated according to the hyperbolic model during stable
respiration. In 78% of the cases, the value of
r2 was greater when the hyperbolic model was
used. We demonstrated that Rmax was in excellent agreement with
P/. UARS patients exhibited higher awake mean Rmax than normal
subjects and other subgroups and a larger increase from wakefulness to
slow-wave sleep than subjects with OSAS, OSHS, and SS. Analysis of
breath-by-breath changes in Rmax was also a sensitive method to detect
episodes of high resistance during sleep.
sleep-disordered breathing; upper airway; model of pressure-flow
relationship
| |
INTRODUCTION |
THE RELATIONSHIP BETWEEN
AIRFLOW and pressure in the upper airway has been well
characterized during wakefulness and can be adequately predicted by the
Rohrer polynomial equation (20), which assumes that the
caliber of the conduit through which airflow is passing is constant.
However, during sleep, loss of muscle tone results in variable
narrowing of the upper airway during inspiration, with consequent flow
limitation. As the negative inspiratory pressure increases in
magnitude, so too does the degree of airway narrowing, and hence the
pressure-flow relationship becomes curvilinear. For normal subjects
during sleep, Hudgel et al. (9) have proposed a
mathematical model incorporating a hyperbolic equation, which better
describes this pressure-flow relationship. This model worked well in a
study on normal, nonobese, nonsnoring subjects. However, the airflow
limitation during sleep in these subjects was relatively small, and the
intrathoracic pressure generated to overcome this was also small (in
the range 0 to 
5 cmH2O). The degree of flow limitation
does not allow anticipation of the absolute level of inspiratory
efforts found in different subgroups of subjects (i.e., normal and
apneic). Moreover, for a given flow, upper airway narrowing has to
intensify as the downstream pressure increases. The upper
airway shape and cross-sectional area are modified in apneic subjects.
It is well known that upper airway collapsibility is much higher in
apneic subjects than in normal subjects. These anatomic and functional
differences may result in different pressure-to-flow ratios
(P/) both before, e.g., the variable part of the linear
relationship between flow and pressure, and after the inspiratory flow
limitation has been established because downstream pressure is much
greater during sleep in OSAS than in normal subjects.
Accordingly, it has been clearly demonstrated that a visual aspect of
flow-limited breath may correspond to various P/
relationships (3). Thus these specific conditions in
apneic subjects could have modified P/ compared with that of
normal subjects. It is not known whether Hudgel's model also applies
to episodes of more severe upper airway flow limitation that occur in
patients with obstructive sleep apnea or with the upper airway
resistance syndrome (UARS), in which the intrathoracic pressure swings
may be as large as 60 cmH2O (5). Nor is it
known whether this model is flexible enough to predict the
pressure-flow relationship in these patients, because the upper airway
resistance changes with different sleep stages.
In contrast to the situation for normal subjects (9,
28), there are currently few data in the published
literature that describe changes in upper airway resistance during
sleep in individuals with different sleep-related respiratory
disorders. The aims of this study were, therefore, 1) to
compare Hudgel's hyperbolic model with Rohrer's polynomial model in
describing the pressure-flow relationship in patients with abnormal
sleep-related upper airway resistance; 2) to use this
pressure-flow relationship to describe upper airway resistance in a
spectrum of patients with sleep-disordered breathing; 3) to
confirm that the measurements of resistance derived from the hyperbolic
model and from the P/ provided a comparable value; and
4) to evaluate the effects of different sleep stages on
upper airway resistance.
In this study, obstructive apnea is disordered breathing characterized
by a complete cessation of breathing during 10 s or longer, and
obstructive hypopnea is characterized by a clear decrease from baseline
in the amplitude of breathing, as measured with a pneumotachometer,
50% with a duration of 10 s. High resistance (HR) is an event in
which effort progressively increases (pharyngeal pressure becomes more
negative), terminated by a sudden change in pressure to a less negative
level, during 10 s or longer. Patients with obstructive sleep
apnea syndrome (OSAS) exhibit an apnea-hypopnea index (AHI) >10/h with
predominantly apneas, those with obstructive sleep hypopnea syndrome
(OSHS) have an AHI > 10/hr with predominantly hypopneas, and
those with UARS have a high resistance index (HRI) > 10/h and
AHI < 10. Individuals classified as exhibiting simple snoring
(SS) have combined AHI and HRI < 10/h.
| |
METHODS |
Twelve patients (9 men) with suspected sleep-disordered
breathing attending the sleep laboratory for investigation were
studied. Their age was 51.3 ± 11.0 yr, and their body mass index
was 25 ± 2.3 kg/m2 (means ± SD).
Protocol
Polysomnography.
All patients had full nocturnal polysomnography (Respisomnograph
system, Nellcor Puritan Bennett, Minneapolis, MN) analyzed manually
with the KnightScan software package (Nellcor Puritan Bennett). The
physiological signals, recorded at 128 Hz, included two
electroencephalogram channels (CZ-O1 and C3-A2), submental electromyogram, and electrooculogram. Sleep stages were analyzed by use
of the standard criteria of Rechtschaffen and Kales (19). Microarousals were scored by using ASDA criteria (1).
Chest wall and abdominal movements were assessed by noncalibrated
inductive plethysmography and oxygen saturation by pulse oximetry (Biox 3740 Ohmeda, Louisville, CO). The pressure and flow measurements are
detailed in the following paragraph. Respiratory events were scored
according to usual criteria, enabling stratification of patients into
different types of sleep-disordered breathing. Body position was
monitored continuously.
Upper airway pressure and flow measurements.
Airflow was measured via a pneumotachometer (Kontron Instruments,
Quentin, France) installed on a full-face mask. This system had a dead
space of <100 ml and produced a linear reading of airflow within the
range 0-1.5 l/s, with an estimated error of ± 0.01 l/s.
Before each study, the pneumotachometer was reset to zero and
calibrated with a fixed-flow generator. Leaks were detected when a
drift in the baseline of the signal occurred. A pressure transducer was
connected to the mask by a noncollapsible tube to measure the mask
pressure. A solid-state multitransducer catheter was used to measure
pressures at three different levels in the pharynx simultaneously
(Gaeltec, Isle of Skye, Scotland, UK). The frequency
response for each microtransducer was 10 Hz at 10°. For a pressure
that cycled between +4 and 4 cmH2O, the latency of
response was 2.5 ms. Three pressure transducers were integrated in the
Gaeltec. They were positioned 80, 120, and 180 mm from the nostril, in
the nasopharynx, velopharynx, and oropharynx, respectively (Fig.
1). Care was taken to ensure that the
membrane of each transducer was positioned parallel to the direction of airflow. Direct inspection of the pharynx and prerecording
visualization of the pressure curves permitted control of the Gaeltec
position. Correct positioning of each transducer was subsequently
confirmed with cephalometry at the end of the night. The neck position
was standardized at the start of the night, all patients being in a
comparable supine position. In 10 of the 12 patients, a cephalometric X-ray was done to evaluate the upper airway morphology and allowed assessment of the pharyngeal catheter's positioning. These X-rays were
made on the morning after the polysomnography, with the Gaeltec still
in place. This allowed us to control for the position of each
transducer, although this was less critical regarding the naso- and
hypopharyngeal sensors. In four patients, all the transducers were in
an adequate position, including those at the tip of the soft palate. In
six patients, the naso- and hypopharyngeal transducers were correctly
positioned, but, in five of these six patients, the velopharyngeal
transducer was 10, 12, 19, 21, and 27 mm above the uvula, and one
patient had the transducer close to the uvula. However, this limitation
affects any study addressing this issue in sleep breathing disorders.
In this condition, there are marked changes in upper airway anatomy
(i.e., soft palate length) that are much more pronounced than in normal
subjects. Thus, when the sensor was adjusted above the hard palate and
at the basis of the tongue level, because of the variable length of the
soft palate and the uvula, an adequate location of the velopharyngeal
sensor was not systematically obtained. This explains why, although we ensured the positioning of the two sensors that provided our
transpharyngeal measurements very cautiously, we were unable to provide
a velopharyngeal resistance measurement.
Calibration of the Gaeltec catheter was performed before each study
using a humidified pressure chamber heated to 37°C. The signal was
recorded at 16 Hz, and the maximum error of the system was estimated to
be 0.6 cmH2O. The pressure detected with this type of
catheter is restricted to static pressure because the membrane of the
sensor was parallel to the flow. In the present study, the dynamic
pressures were considered to be negligible (see APPENDIX).
Data were stored and analyzed on Excel software (Microsoft, Redmond,
WA), and calculation of regression curves and correlation coefficients
was performed by Kaleidagraph software (Synergy Software, Reading, PA).
Analysis
Comparison between the two mathematical models of pressure-flow
relationship.
Two different equations that have been used to describe the
pressure-flow relationship in the upper airway were compared (Fig. 2).
View larger version (10K):
[in this window]
[in a new window]
|
Fig. 2.
Hyperbolic and polynomial pressure-flow
relationship. Dashed line, Rohrer equation; solid line, Hudgel
equation (see text).  , Asymptote for peak flow;  , pressure at
50% of peak flow.
|
|
First is the Rohrer equation (20)
where P is the pressure (cmH2O), is flow
(l/s), and the constants K1 and
K2 represent the y-intercept and
slope, respectively, of the linear form of the Rohrer equation. The
second is Hudgel's hyperbolic equation (9)
where is the asymptote for peak flow (i.e., the flow at
which pressure reaches infinity), and is the pressure at 50% of
this flow.
Two hundred fifteen respiratory cycles were randomly selected. They
were observed in different sleep stages [stage 1, stage 2, stage 3/4,
and rapid eye movement (REM) sleep] and during wakefulness. One
non-flow-limited cycle was chosen during stable respiration and two
flow-limited cycles at the beginning and the end of an episode of
increased upper airway resistance (i.e., HR). Flow-limited cycles were
selected on the visual aspect of flow. Approximately the same number of
respiratory cycles was chosen for each patient in each sleep stage.
Each respiratory cycle was fitted by the polynomial and hyperbolic
equations. Nonlinear regression by the Marquardt method, which combines
the method of steepest descent and the Gauss-Newton method
(17), was used to correlate the data with each equation
and thus to determine which model best describes the pressure-flow
relationship in this group of patients. Pearson's correlation test
(r2) was used to compare the polynomial and
hyperbolic regression curves for each sleep stage. This test was valid
because the two equations had the same number of parameters.
Upper airway resistance measurements in a spectrum of
sleep-disordered breathing.
Recordings of 5 to 10 stable (in terms of pressure and flow aspects)
consecutive respiratory cycles were made during each sleep stage and
during wakefulness for each subject, representing 600 respiratory
cycles. The breaths were selected during stable respiration in
undisturbed sleep periods. In the 20 s preceding the period of
analysis, no respiratory events should be present and no microarousal
should be detectable. This prevented analysis of periods including
either respiratory events or hyperventilation breaths just after apneas
or hypopneas. The only exclusion criterion applied for selecting the
period of analysis in OSAS, OSHS, UARS, and SS was the occurrence of an
event (e.g., apnea, hypopnea, HR episode). In HR episodes, there is a
progressive increase in inspiratory flow limitation that accompanies
the increase in respiratory effort. However, stable inspiratory flow
limitation could be included in the analysis.
For each respiratory cycle, the resistance was estimated from the
values of peak flow () and pressure (P). There were actually three different situations: 1) If there was no flow
limitation, the flow used to calculate resistance was clearly the peak
of flow, which coincided with the peak of pressure. 2) The
cycle was flow limited, but there was a plateau without further
decrease in flow. The plateau value of flow was taken together with the value of peak pressure, which again coincided. 3) The cycle
was flow limited, but there was a decrease in flow after the plateau. The reduction in flow preceded the occurrence of the peak pressure. The
last part of the breath was not taken in account; the last value of the
flow plateau and the corresponding pressure value were used to
calculate resistance. Regarding the third situation, we chose the last
flow value of the plateau so as not to include periods of rapid
reduction in sections of the pharynx, as suggested by a reduction in
flow concomitant with a further increase in pressure. Resistance (R)
was calculated by using the equation
derived from the hyperbolic Hudgel equation
and also directly from the pressure-flow relationship
R
P/ = P/. Mean values and SD of
resistance for each period were calculated from the two methods and
subsequently compared. All interpretation of resistance variation in
different sleep-disordered breathing, during wake and sleep stages, was
done with resistance calculated by the equation derived from the
hyperbolic model.
| |
RESULTS |
The demographic data and polysomnographic characteristics of the
patients are presented in Tables 1 and
2 and Fig.
3. The patients demonstrated a spectrum
of respiratory sleep disorders. Two individuals were considered as
exhibiting SS, three as having UARS, four as having OSHS, and three as
having OSAS.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 3.
Difference in apnea-hypopnea index (AHI) and high
resistance index (HRI) in various subgroups of patients with sleep
breathing disorders. OSAS, obstructive sleep apnea syndrome; OSHS,
obstructive sleep hypopnea syndrome; UARS, upper airway resistance
syndrome; SS, simple snoring. There is a progressive decrease in AHI
and increase in HRI, which may correspond to different underlying
mechanisms (see text).
|
|
Comparison Between the Two Mathematical Models of the Pressure-Flow
Relationship
The two equations (hyperbolic and polynomial) were applied to 215 respiratory cycles, both flow limited and non-flow limited (~18
cycles studied in each of the 12 patients). A Pearson's square correlation was done for each respiratory cycle to assess the validity
of the two models. Median (range) r2 values were
0.88 (0.31-0.99) and 0.92 (0.53-0.99) in the polynomial and
hyperbolic models, respectively. In 78% of the cases, the value of
Pearson's square was greater when the hyperbolic model was used. The
difference in correlation for the flow-limited respiratory cycles was
even more pronounced, the hyperbolic model being better in 86% of the
cases. In this condition, the median (range) r2
was 0.83 (0.31-0.99) in the polynomial vs. 0.91 (0.65-0.99)
in the hyperbolic model. Using the hyperbolic model, we calculated resistance at the peak of pressure, aiming to demonstrate that cycle-by-cycle measurement of resistance perfectly reflected the progressive occurrence of a HR episode (Fig.
4).
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 4.
Variation in pharyngeal resistance estimated by the
hyperbolic model (Rhyperbolic,pharyngeal)
during a high resistance (HR) episode in stage 2 sleep. Upper airway
resistance was estimated cycle by cycle. In this HR episode, there
is a clear crescendo in resistance and a return to baseline at the end
of the event. The fifth cycle, which does not exhibit any flow
limitation, shows a decrease in resistance.
|
|
Comparison of Rhyperbolic and P/
Resistances were calculated by using P/ at the peak
pressure in 600 respiratory cycles during stable breathing at different sleep stages. The obtained values correlated very closely with those
derived directly from the calculation using the hyperbolic model. This
was demonstrated both by the linear correlation test (r2 = 0.98, P < 0.0001)
and the Bland and Altman plot (Fig. 5).
Thus Rhyperbolic gives no additional information than
P/ when upper airway resistance is measured during different
sleep stages. Furthermore, Fig. 4 demonstrates that cycle-by-cycle
measurement of resistance at the peak pressure perfectly reflected the
progressive occurrence of a HR episode.
View larger version (11K):
[in this window]
[in a new window]
|
Fig. 5.
Bland-Altman plot comparing values of resistance,
determined at peak pressure, derived from either the hyperbolic model
(Rhyperbolic) or a pressure-to-flow ratio
(RP/). Mean of the differences (bias) in
resistance was 0.3 cmH2O · s · l 1. The
difference between results with the two measures of resistance will
fall within 9.1 cmH2O · s · l1
of the mean difference 95% of the time.
|
|
Nasal Resistance Changes from Wakefulness to Sleep
In all subjects, awake nasal resistance was higher than in the
"normal" range. There was no clear trend to increase from
wakefulness to deep sleep (stage 3/4), except in the UARS group. See
Fig. 6 and Table
3.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 6.
Evolution of nasal resistance
(Rhyperbolic,nasal) from wakefulness to deep sleep in the
various subgroups of sleep-disordered breathing.
|
|
Pharyngeal Resistance Changes from Wakefulness to Sleep
Awake upper airway resistance in OSAS, OSHS, and SS were in the
"normal" range. Conversely, UARS demonstrated a threefold increase
in upper airway resistance compared with normal subjects (11.9 ± 6.3 vs. 4.6 ± 0.8 cmH2O · s · l1). There was a general trend toward an increase in
pharyngeal resistance from wakefulness to deep sleep (stages 3 and 4).
The magnitude of this increase varied considerably depending on the subgroup of patients. The progression of increase in mean resistance (Rmax) during stable respiration between wakefulness and stage 3/4 was
frank in UARS but lesser in SS, OSHS, and OSAS. See Fig. 7 and Table
4.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 7.
Increase in
Rhyperbolic,pharyngeal from wakefulness to
deep sleep in the various subgroups of sleep-disordered breathing. When
the increase in resistance during sleep was considered, there was a
frank rise for UARS compared with OSAS and OSHS.
|
|
UARS thus exhibited a high awake upper airway resistance and
demonstrated a further increase in deep sleep, with a mean value of
161.53 ± 75.24 cmH2O · s ·
l1. Huge variations in respiratory drive, leading to
pressure and flow irregularities, were observed during phasic episodes
in REM sleep. Accordingly, it was difficult to identify periods of
stable respiration (consecutive respiratory cycles) suitable to
calculate upper airway Rmax (see below) during REM sleep in all
patients. As a consequence, during REM sleep, upper airway resistances
exhibited considerable cycle-to-cycle variation (see Fig.
8 and large SD in Table 4), and mean
values during this period tended to be quite different from those
obtained during other sleep stages.
View larger version (40K):
[in this window]
[in a new window]
|
Fig. 8.
Rhyperbolic,pharyngeal
variation during stable respiration in REM sleep in a subject with
OSAS. During REM sleep, resistance exhibited large variations in
consecutive respiratory cycles. The fluctuation in driving pressure is
principally responsible for this variation.
|
|
| |
DISCUSSION |
Our study provides the first evaluation of upper airway resistance
in different types of sleep-disordered breathing, ranging from SS to
OSAS. In the present study, OSAS, OSHS, and SS subjects had normal
awake resistance, whereas UARS patients exhibited awake resistance
threefold higher than normal subjects (28). Moreover, upper airway resistance was further greatly increased (fourfold) during
slow-wave sleep in UARS. In all subjects, the hyperbolic Hudgel
equation was a better mathematical model to describe the pressure-flow
relationship during sleep. This was particularly true for flow-limited
cycles. This model was robust enough to accurately quantify the changes
in upper airway resistance encountered during the different sleep
stages. Finally, the resistance derived from the hyperbolic model at
the peak of pressure and flow was comparable to the value provided by
the P/ equation.
Limitations in Pharyngeal Pressure Measurements
The presence of a pharyngeal catheter could potentially alter the
dynamics of the upper airway. It could also disturb the structure and
quality of sleep (2). Acceptance by the patient is also an
issue (we experienced at least 10% refusal; see Ref. 12).
Finally, positioning is critical because the sensor may be in contact
with pharyngeal structures, making the measurement highly sensitive to
any movement of the catheter, particularly during increased respiratory efforts.
The second transducer was theoretically at the tip of the soft
palate and thus could have been used to estimate the velopharyngeal resistance. However, this was not possible in the present study because
this transducer was not always at the tip of the soft palate on account
of the variability of the length and shape of the uvula in
sleep-disordered-breathing patients. Taking in account these
limitations, we determined that estimation of the velopharyngeal resistance was not possible. Apart from these general limitations, the
type of catheter used to detect pharyngeal pressures may be of
importance, depending on its ability to detect dynamic pressure. These
physical characteristics are detailed in the APPENDIX.
Because our sensor was poorly sensitive to dynamic pressure, this
component should be negligible in our experiment; thus this supports
the validity of our pressure measurements.
Flow Measurements
The use of a facial mask had the advantage of adequately measuring
airflow but excluded the possibility of separately quantifying nasal
and oral breathing. The equipment probably favored nasal respiration
because mouth opening is more difficult in that condition. Different
routes of breathing have been shown to be associated with different
levels of collapsibility (14). However, during our
measurements, we never detected any abrupt change in transpharyngeal resistance that could have resulted from a shift from nasal to oral
breathing or the converse.
Impact of Body Position
We carefully looked to the position in which the measurements were
performed. For 7 of the 12 patients, body position was identical for
all measurements (3 supine, 1 in left lateral, and 3 in right lateral
position). There was no clear difference in resistance values according
to body position. Four patients were in lateral position when awake and
supine when asleep. Although this might have led to slight
underestimation of the awake resistance value, there was no systematic
trend to support that hypothesis. In the last patient, the awake and
stage 1 measurements were obtained when the patient was supine, whereas
the other measurements were performed in right lateral position. In
that case, the increase in resistance in stage 2, stage 3/4, and REM
sleep is so high that, again, no effect of body position was detectable.
Regarding neck position, this was standardized in all the patients
at the start of the night, but patients were not maintained in a fixed
position throughout the night. A fixed position would have been
difficult to tolerate in addition to the catheter and the facial mask
and probably would have strongly interfered with sleep quality.
Validation of Pressure-Flow Relationship
Hudgel et al. (9) demonstrated in 1988 that the
hyperbolic model was the best fit of pressure-flow curves in normal
men. We have also demonstrated that, in a spectrum of patients with sleep-disordered breathing, the hyperbolic model was the more accurate
fit of pressure-flow relationship, whatever the sleep stage and whether
or not inspiratory flow limitation occurred. Flow limitation represents
a stagnation of flow despite an increase in driving pressure, and, in
this condition, the pressure-flow curve must be asymptotic. The
polynomial equation does not lead to any asymptote. Thus, as expected,
particularly for flow-limited cycles, the hyperbolic model was better.
Validation of the Methods Used to Estimate Pharyngeal Resistance
We compared the values of resistance, determined at peak pressure
in 600 inspiratory cycles, derived by either the hyperbolic model or a
pressure-to-flow ratio. At high resistance values (>80 cmH2O · s · l1), any method is
questionable because the flow is very low. Although we did not find any
statistical difference between Rhyperbolic and
RP/, this is also evidenced by the dispersion of
the Bland-Altman plot above 80 cmH2O · s · l1 (see Fig. 5).
However, these differences may not be clinically relevant. This finding
was actually not expected because the only situation in which the
resistance should be fully described by P/ is when each point
corresponds to a constant section and a laminar flow of the
fluid. None of these conditions is met when resistance is
measured at the pharyngeal level. First, cross-sectional area is highly
variable throughout the different levels of upper airway. Moreover,
variations of cross-sectional area occur within breaths. These
conditions determine the regime of the flow, which is likely to be
highly turbulent. Second, the pharyngeal walls are mobilized by a
suction effect related to the velocity of the fluid. These two factors
result in a loss of energy by friction and pharyngeal wall
displacement. Thus the concordance between these two values actually
reflects the adequacy of the model to fit with the experimental values,
which indicates the strength of the correlation. However, this
is no longer true when flow decreases with increasing inspiratory
efforts, as mentioned under METHODS. In this case, the
resistance determined by the hyperbolic fit underestimates the true
resistance, which is better estimated by P/. We thus did not
include the last part of the breaths exhibiting a diminished flow after
the plateau of inspiratory flow limitation. In this case, the
resistance value is affected by several uncontrolled factors, such as
reduction in cross-sectional area and increase in the dynamic pressure value.
Upper Airway Resistance in Different Sleep Stages in a Spectrum of
Sleep-Disordered Breathing
We chose to calculate resistance during stable respiration periods
to compare the different subgroups of patients. The comparison was
valid because the same criteria were used for the breath selection. Stable inspiratory flow limitation could be included in the analysis. In that case, flow limitation occurred when the resistance was high but
stable, as it frequently occurred in stage 3/4 or in deep stage 2. We
observed this phenomenon in all the UARS patients but also in some of
the OSAS and OSHS patients. However, this inspiratory flow limitation
was clearly distinct from a HR episode. In HR, resistance increased
progressively breath by breath (see Fig. 4). Moreover, the estimation
of the resistance was made in 5 to 10 consecutive breaths, and the
small standard deviation for each resistance measurement supports the
fact that the upper airway was actually stable. Thus we believe that
the selection criteria that we used made the resistance values obtained
in the different groups of patients comparable. Sleep stages are known to influence resistance of the upper airway in normal individuals (28). With the onset of sleep, muscle activity is
progressively reduced in both tonic and phasic components of upper
airway dilating muscle activity (26, 30).
Upper airway narrowing occurs during sleep and varies significantly,
depending on a number of factors such as pharyngeal anatomy, posture,
sleep stage, and level of muscle tone (16,
21). The reduction in pharyngeal size associated with the
loss of muscle tone modifies the properties of the pharyngeal wall and
diminishes the upper airway's ability to resist the negative intraluminal pressures generated during inspiration (7).
Thus, even in normal individuals, marked changes in upper airway
resistance have been described during sleep (28) and tend
to become more pronounced during slow-wave sleep (stage 3/4),
especially in men (27). This increase in resistance has
been described as being at least partly responsible for the fall in
minute ventilation at the onset of sleep (10,
30). Our study is the first to describe the effects of
sleep stage on pharyngeal resistance in a group of individuals
representing a spectrum of sleep-disordered breathing. As in normal
subjects, there was a general trend for patients to progressively
increase their pharyngeal resistance as they entered deeper sleep
stages. However, the magnitude of increase from light to deep slow-wave
sleep was higher than in normal subjects (25).
Upper Airway Resistance and Pathophysiology of Upper Airway
Collapse
Although the subgroups of patients were small, there were
differences in the progressive increase of upper airway resistance from
wakefulness to sleep according to the predominant type of sleep
breathing disorders (Fig. 6). This suggests different underlying mechanisms in upper airway control (7). Pharyngeal patency is dependent on both the size of the upper airway and the stiffness of
the pharyngeal walls. A substantial reduction in upper airway size
while awake has been demonstrated in OSAS patients compared with
snorers (6, 18). To counteract small upper
airway size, the level of neural activation of upper airway dilator
muscles is abnormally elevated in patients with obstructive apneas or hypopneas (15). Schwab et al. (22), using
ultrafast CT scan, have studied upper airway cross-sectional area
during awake tidal breathing. They found larger variations in OSAS
patients, suggesting increased airway compliance. This was particularly
true during expiration, i.e., when the dilator muscles become less
active. In patients with simple snoring, and probably in UARS, upper
airway size is less reduced than in apneic subjects (13),
and wall stiffness is generally considered to be higher compared with
OSAS patients. As expected, pharyngeal collapsibility, as assessed by
the measurement of critical pressure (4), is progressively increased from snorers to hypopneic and, finally, apneic patients. When
collapsibility is high and the lumen is already narrowed, as in OSAS, a
small increase in resistance and negative inspiratory pressure is able
to produce complete airway collapse. Thus apneic subjects can
demonstrate normal upper airway resistance during awake periods as a
result of pharyngeal dilator compensatory hyperactivity. They can also
demonstrate infinite upper airway resistances during apneas, but their
upper airway are not stable enough to sustain intermediate levels of
upper airway resistance. On the contrary, UARS patients who exhibit
more stiffening of pharyngeal walls may generate very high levels of
upper airway resistance and marked negative inspiratory pressure during
sleep without apnea or frank hypopnea. Our data support this view,
because awake upper airway resistance in UARS patients was threefold
higher than in normal subjects and was further augmented during sleep
(being fourfold higher in slow-wave sleep compared with wakefulness).
As expected, the patients with predominant hypopnea (i.e., with an
intermediate collapsibility between SS and OSAS) exhibited levels of
resistance in stage 3/4 comparable to those of the apneic subjects in
stage 2. The mechanism underlying obstructive hypopneas is less clear. It is likely that many hypopneas are incomplete obstructive apneas, i.e., there is partial collapse of a highly compliant airway wall in
response to a relatively smaller increase in resistance and negative
intrathoracic pressure than during an apnea. Alternatively, an
obstructive hypopnea may be a variant of a HR episode during which high
inspiratory resistance is applied to an upper airway that is not stiff
enough to maintain normal airflow. These two mechanisms are distinct
but are not necessarily distinguishable. We looked at the difference
between hypopnea and HR episodes in another experiment with OSHS
patients (IHA = 30 ± 25/h with IA = 3 ± 5/h),
using a facial mask, a pneumotachometer, and esophageal pressure
measured by a balloon catheter (data still unpublished). There was no
significant difference in maximal esophageal pressure between hypopnea
and HR episode (22 ± 6 vs. 19 ± 6 cmH2O),
whereas the difference in flow reduction was, by definition, very
significant (58 ± 11 vs. 21 ± 5%, P < 0.001). This supports the fact that collapsibility is increased during
hypopneas compared with HR episodes. The data of the present study,
however, apply to resistance during stable respiration. In the OSHS
group, the resistance values seem to be lower than in OSAS. Conversely,
in the UARS group, in which both hypopneas and HR episodes coexist, the
resistance observed during stable respiration is much higher than in
patients exhibiting mainly hypopneas. The length of events is
different, with the HR episodes being longer than hypopneas. We can
then hypothesize that 1) HR during stable respiration is
probably not possible for hypopneic patients because of their
pharyngeal collapsibility, 2) the resistance obtained during
hypopneas is higher (this is supported by the results of esophageal
pressure provided above, because the "transpharyngeal" resistance
is actually higher during hypopneas), and 3) the slope of
increase in resistance in HR episodes is lower compared with hypopnea
but is not contradictory with a higher resistance during stable
respiration. Both are in favor of a more stable pharynx. Whether the
difference in resistance between apneic and UARS in REM and stage 3/4
is actually related to different levels of collapsibility, however,
remains to be further established by critical pressure measurements
during sleep in these different subgroups of patients.
The physiology of the upper airway during REM sleep is still
unclear. Some authors have found that upper airway collapsibility or
compliance is increased (24). This seems logical given
that the reduction in muscle tone is at a maximum during this state of
vigilance (29). This is supported by the fact that apneic events are more frequent and longer during REM sleep (23).
Conversely, cross-sectional area during both tonic and phasic REM,
although reduced, seems to be stable within the breath
(16), suggesting a limited collapsibility. In our study,
in REM sleep, the cycle-by-cycle fluctuation in resistance was
important. This was at least partly due to the large cycle-by-cycle
variability in driving pressure explained by the changes in respiratory
drive during phasic REM (11). During REM sleep (Fig. 8),
resistance at peak pressure is highly variable compared with that in
non-REM sleep, in which upper airway resistance showed very little
variation, as attested by the limited standard deviation of
the mean value of upper airway resistance (see Table 4).
The nonlinearity in the pressure-flow relationship during inspiration
is commonly caused by narrowing of an hypotonic upper airway in
response to the negative intrathoracic pressure developed during
inspiration (3, 8). It is generally accepted
that the recognition of inspiratory flow limitation is an adequate tool
to identify elevated upper airway resistance both in normal subjects
and in patients with sleep-disordered breathing (3, 8). Hosselet et al. (8) found that, in each
sleep stage, including REM sleep, the resistance of an abnormal-contour
breath (i.e., flow limited) was higher than the resistance of a breath with normal contour. Clark et al. (3) were less
categorical and found that, during non-REM sleep, flow-rate shape was
effective in differentiating severe types of inspiratory flow
limitation but was not able to consistently detect low levels of
increased resistance. We have shown that, during HR episodes in non-REM sleep (Fig. 4), an increase in resistance progressively occurs in
accordance with an aggravation of inspiratory flow limitation. During
REM sleep, effort-related pressure swings are often low and erratic. In
these conditions, flattened flow contour or esophageal pressure, if
considered alone, are probably unable to describe the fluctuation in
upper airway resistance accurately.
In conclusion, upper airway resistance measurements were performed in a
spectrum of patients with sleep-disordered breathing. The more accurate
fit of the pressure-flow relationship was the hyperbolic model.
P/ at peak pressure is a reliable method to estimate upper
airway resistance. We found, as in normal subjects, a general trend for
patients to progressively increase their pharyngeal resistance as they
entered deeper sleep stages. The slope of increase in upper airway
resistance from wakefulness to sleep was substantially different in
UARS patients compared with snorers or apneic subjects. This suggests
different underlying mechanisms in upper airway control in these
different subgroups of patients.
| |
APPENDIX |
There are three physical components for the pressure in a tube:
viscoelastic pressure, which depends on flow and viscosity of the
fluid; dynamic pressure, which depends on flow and the section of the
tube; and static pressure, which depends on the volume of the tube and
the quantity of fluid. The measurement of pharyngeal pressure performed
via balloon catheter [such as that used by White et al.
(28)] includes static pressure and a variable proportion
of dynamic pressure. When this type of catheter is used in a narrowed
upper airway, a further reduction of upper airway size is induced by
the inflated balloon, and thus the relative proportion of dynamic
pressure is likely to increase. By contrast, we used a catheter that
mainly detected static pressure. The dynamic component, which could be
present, was undetected by this type of catheter, and, therefore, the
inspiratory change in pressure could potentially be
overestimated. In the present study, the pressure sensors were
at the two pharyngeal levels (nasopharynx and hypopharynx), at which
the changes in cross-sectional area are rather limited through the
respiratory cycle. In this context, a theoretical calculation of
dynamic pressure suggested that the dynamic component was negligible at
both the nasopharyngeal and hypopharyngeal levels. Static and dynamic
pressures are the two principal components of the driving pressure.
|
|
|
|
Dynamic pressure is correlated with the square of fluid
velocity. The speed of the fluid is determined by the section of the
tube where the measure is made and by the flow crossing the tube. With
increase of flow, the more dynamic pressure increases, the more the
static pressure is underestimated. In sleep-disordered breathing, the
respective sections of nasopharynx and hypopharynx are ~100 and 200 mm2 (7). During sleep, flow is in a range of
0-0.5 l/s. In this condition, the dynamic pressure calculated at
0.5 l/s was 0.14 cmH2O in the nasopharynx and 0.025 in the hypopharynx.
| |
ACKNOWLEDGEMENTS |
This study was supported by a grant from PHRC 1997, French
Ministry of Health.
| |
FOOTNOTES |
Address for reprint requests and other correspondence: P. Levy,
Sleep and Respiration Unit, EFCR, Hôpital Michalon, BP 217 38 043 Grenoble, Cedex 9, France (E-mail: Patrick.levy{at}imag.fr or
PLevy{at}chu-grenoble.fr).
The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement"
in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Received 1 April 1999; accepted in final form 8 March 2000.
| |
REFERENCES |
1.
American Sleep Disorders Association.
EEG arousals. Scoring rules and examples. A preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association.
Sleep
15:
173-184,
1992[Medline].
2.
Chervin, RD,
and
Aldrich MS.
Effect of esophageal pressure monitoring on sleep architecture.
Am J Respir Crit Care Med
156:
881-885,
1997[Abstract/Free Full Text].
3.
Clark, SA,
Wilson CR,
Satoh M,
Pegelow D,
and
Dempsey JA.
Assessment of inspiratory flow limitation invasively and noninvasively during sleep.
Am J Respir Crit Care Med
158:
713-722,
1998[Abstract/Free Full Text].
4.
Gold, AR,
and
Schwartz AR.
The pharyngeal critical pressure: the whys and hows of using nasal continuous positive airway pressure diagnostically.
Chest
110:
1077-1088,
1996[Free Full Text].
5.
Guilleminault, C,
Stoohs R,
Clerk A,
Cetel M,
and
Maistro P.
A cause of excessive daytime sleepiness: the upper airway resistance syndrome.
Chest
104:
781-787,
1993[Abstract/Free Full Text].
6.
Haponik, EF,
Bohlman PL,
Allen RP,
Goldman SM,
and
Bleecker ER.
Computerized tomography in obstructive sleep apnea.
Am Rev Respir Dis
127:
221-226,
1983[ISI][Medline].
7.
Horner, RL.
Motor control of the pharyngeal musculature and implication for the pathogenesis of obstructive sleep apnea.
Sleep
19:
827-853,
1996[ISI][Medline].
8.
Hosselet, J,
Norman GR,
Ayappa I,
and
Rapoport DM.
Detection of flow limitation with a nasal cannula/pressure transducer system.
Am J Respir Crit Care Med
157:
1461-1467,
1998[Abstract/Free Full Text].
9.
Hudgel, DW,
Hendricks C,
and
Hamilton HB.
Characteristics of the upper airway pressure-flow relationship during sleep.
J Appl Physiol
64:
1930-1935,
1988[Abstract/Free Full Text].
10.
Kay, A,
Trinder J,
and
Kim Y.
Individual differences in relationship between upper airway resistance and ventilation during sleep onset.
J Appl Physiol
79:
411-419,
1995[Abstract/Free Full Text].
11.
Kline, LR,
Hendricks JC,
Davies RO,
and
Pack AI.
Control of the activity of the diaphragm in rapid-eye-movement sleep.
J Appl Physiol
61:
1293-1300,
1986[Abstract/Free Full Text].
12.
Loiodice, C,
Selek M,
Loire M,
Levy P,
and
Pepin JL.
Mesure de l'effort respiratoire (pression oesophagienne) et estimation des résistances des voies aériennes supérieures au cours du sommeil: indication, limites, resultats.
Neurophysiol Clin
28:
507-520,
1998[Medline].
13.
Mayer, P,
Pepin JL,
Bettega G,
Veale D,
Ferretti G,
Deschaux C,
and
Levy P.
Relationships between body mass index, age and upper airway measurements in snorers and sleep apnea patients.
Eur Respir J
9:
1801-1809,
1996[Abstract].
14.
Meurice, JC,
Marc I,
Carrier G,
and
Series F.
Effect of mouth opening on upper airway collapsibility in normal sleeping subjects.
Am J Respir Crit Care Med
153:
255-259,
1996[Abstract].
15.
Mezzanotte, WS,
Tangel DJ,
and
White D.
Waking electromyogram in sleep apnea versus normal controls (a neuromuscular compensatory mechanism).
J Clin Invest
89:
1571-1579,
1992.
16.
Morrell, MJ,
and
Badr MS.
Effect of NREM sleep on dynamic within-breath changes in upper airway patency in humans.
J Appl Physiol
84:
190-199,
1998[Abstract/Free Full Text].
17.
Motulsky, HJ,
and
Ransnas LA.
Fitting curves to data using nonlinear regression: a practical and nonmathematical review.
FASEB J
1:
365-374,
1987[Abstract].
18.
Pepin, JL,
Veale D,
Ferretti G,
Mayer P,
and
Levy P.
Obstructive sleep apnea syndrome: hooked appearance of the soft palate in awake patients-cephalometric and CT findings.
Radiology
210:
163-170,
1999[Abstract/Free Full Text].
19.
Rechtschaffen, A,
and
Kales A
(Editors).
A Manual of Standardized Terminology, Techniques, and Scoring System for Sleep Stages of Human Subjects. Bethesda, MD: National Institute of Neurological Diseases and Blindness, 1971. (NIH Publ. No. 204).
20.
Rohrer, F.
The resistance in the human airway and the influence of branching of bronchial systems on frequency of breathing at different lung volumes.
Pflügers Arch
162:
255-299,
1915.
21.
Rowley, JA,
Zahn BR,
Babcock MA,
and
Badr MS.
The effect of rapid eye movement (REM) sleep on upper airway mechanics in normal human subjects.
J Physiol (Lond)
510:
936-976,
1998.
22.
Schwab, RJ,
Gefter WB,
Hoffman EA,
Gupta KB,
and
Pack AI.
Dynamic upper airway imaging during awake respiration in normal subjects and patients with sleep disordered breathing.
Am J Respir Crit Care Med
148:
1385-1400,
1990.
23.
Series, F,
Cormier Y,
and
La Forge J.
Influence of apnea type and sleep stage on nocturnal postapneic desaturation.
Am Rev Respir Dis
141:
1522-1526,
1990[ISI][Medline].
24.
Shea, SA,
Edwards JK,
and
White DP.
Effects of sleep-wake transitions and REM sleep on genioglossal response to upper airway negative pressure (Abstract).
Am J Respir Crit Care Med
157:
A653,
1998.
25.
Skatrud, JB,
and
Dempsey JA.
Airway resistance and respiratory muscle function in snorers during sleep.
J Appl Physiol
59:
328-335,
1985[Abstract/Free Full Text].
26.
Tangel, DJ,
Mezzanotte WS,
and
White DP.
Influence of sleep on tensor palatini EMG and upper airway resistance in normal men.
J Appl Physiol
70:
2574-2581,
1991[Abstract/Free Full Text].
27.
Trinder, J,
Kay A,
Kleinman J,
and
Dunai J.
Gender differences in airway resistance during sleep.
J Appl Physiol
83:
1986-1997,
1997[Abstract/Free Full Text].
28.
White, DP,
Lombard RM,
Cadieux RJ,
and
Zwillich CW.
Pharyngeal resistance in normal humans: influence of gender, age, and obesity.
J Appl Physiol
58:
365-371,
1985[Abstract/Free Full Text].
29.
Wiegand, L,
Zwillich CW,
Wiegand D,
and
White DP.
Change in upper airway muscle activation and ventilation during phasic REM sleep in normal men.
J Appl Physiol
71:
488-497,
1991[Abstract/Free Full Text].
30.
Worsnop, C,
Kay A,
Pierce R,
Kim Y,
and
Trinder J.
Activity of respiratory pump and upper airway muscles during sleep onset.
J Appl Physiol
85:
908-920,
1998[Abstract/Free Full Text].
J APPL PHYSIOL 89(1):120-130
8570-7587/00 $5.00
Copyright © 2000 the American Physiological Society
TOP
ABSTRACT
INTRODUCTION
