Why Don't All Heavy Snorers Have Obstructive Sleep Apnea?1,2
Olll J. POLO, MEHDI TAFTI, JEAN FRAGA, KIMMO V. K. PORKKA, YVES DEJEAN, and MICHEL BilliARD Introduction
Patients with obstructive sleep apnea (OSA) syndrome report a history of heavy snoring for years or decades before the appearance of daytime somnolence and respiratory pauses during sleep (l). Therefore snoring is an important risk factor for OSA. However, only a small fraction of habitual snorers appear to develop OSA. About 15070 of middle-aged and older men are habitual snorers, yet only about 1 to 4% of the population has OSA syndrome (2). The conditions under which strong inspiratory suction causes complete oropharyngeal occlusion are insufficiently understood. Also, the role of heavy snoring-induced vibration injury in the advancement of pharyngeal anatomic abnormalities observed in OSA (3-5) has not been clarified. According to the law of Bernoulli, local pressure conditions in a tube are determined by the velocity of the airflow (6). The velopharynx is the narrowest segment of the pharyngeal airways. Therefore, it is repetitively exposed to local increased airflow speed and decreased intraluminal pressure. The importance of the kinetic properties of the airflow on the patency of the upper airways is demonstrated by the discrepancy between the relatively low static occlusion pressures (3.5 to 12.0 cm H 2 0 ) observed in snorers (7), and their high tolerance of intrathoracic suction pressures (20 to 60 em H 2 0 ) (8) while air is flowing through. In view of this, we reasoned that snorers without apnea may have a reduced inspiratory suction effect locally in the oropharynx because of proportional differences in upper airway dimensions. Accordingly, we measured the upper airway cross-sectional areas at four levels in patients with OSA, heavy snorers without apnea, and control patients while all were awake. In an attempt to assess the breathing-related dynamic changes in the upper airway morphology we also measured the cross-sectional areas during deep inspiration. 1288
SUMMARY Patients with obstructive sleep apnea (OSA) and heavy snorers without apnea both show Intrathoracic suction pressures during sleep that exceed their static upper airway closing pressures. Complete airway occlusion, however, occurs only In the former patient group. We hypothesized that the kinetic properties of the airflow would be different In these two typas of patients because of differences In upper airway morphology. The pharyngeal computed tomography (CT) was used to measure the cross-sectional areas of the upper airways In 15 patients with OSA, 25 nonapnelc heavy snorers, and 14 control sublects while thay were awake. Nocturnal breathing was monitored with the static charge-sensitive bed (SeSB). The patients with OSA had a narrower airspace at the velopharyngeal (VP) level than the controls (p < 0.01); the nonapnelc snorers did not differ from the other groups. At the tongue base (TB) and the hyoid bone (HB) levels there was no difference between the OSA and the control groups, but the nonapneic snorers had narrower airways at both of these levels compared with control sublects (p < 0.01) and at the hyoid bone level compared with the OSA group (p < 0.05). The VP/HB ratio was the parameter that best distinguished the patients with OSA from the nonapnelc snorers (lower In the OSA group, p < 0.001). We suggest that airway collapse during sleep is favored by a narrow velopharynx associated with large hypopharynx. Some heavy snorers may not have an oropharyngeal collapse because the paak inspiratory suction pressure could already be damped down at the level of the relatively narrow AM REV RESPIR DIS 1991; 143:1288-1293 hypopharyngeal airways.
Methods Upper airway morphology was studied in three groups with clearly distinctive breathing patterns observed in sleep recordings. The OSA group consisted of 15 patients who presented episodes of OSA during at least 5% of the time in bed (TIB) but had no episodes of constant snoring with increased respiratory resistance. The increased respiratory resistance (IRR) group included 25 snorers who showed at least 5070 of increased respiratory load but had less than 5% of OSA. The control group consisted of 14 patients referred to the sleep clinic because of their snoring complaint but for whom the sleep recordings showed neither obstructive apnea nor increased respiratory resistance and in whom minor breathing abnormalities were present during less than 100/0 of TIB. Most of these subjects were not heard snoring during the recording night. Each patient was monitored all night with the static charge-sensitive bed ([SCSBj Biornatrs; Biorec, Thrku, Finland) and ear oximeter (Biox'" II; the BOC Group Inc., Louisville, CO). A pharyngeal computed tomographic (CT) study (Somatom DR3R; Siemens AG, Erlangen, Germany) was performed during the daytime. Static Charge-sensitive Bed The major problem in studying snoring is the lack of suitable techniques for differentiat-
ing heavy snoring with increased respiratory load from trivial snoring with noise as the only complaint. The standard polysomnography allows quantification of sleep apnea and hypopnea but provides no information about the severity of an eventual partial airwayobstruction. For the purpose of this study esophageal pressure measurements would have provided the most exact data. Because the esophageal balloon was not tolerated by all of our subjects, however, the SCSB was the method of choice. The SCSB is a sensitive movement sensor placed under a special foam plastic mattress, allowing noninvasive cardiorespiratory monitoring without electrodes or cables attached to the subjects (figure 1) (9-12). The sensor produces a three-channel output including
(Receivedin originalform November 27, 1989and in revised form November 15, 1990) 1 From the Sleep Disorders Unit, Gui de Chauliac Medical Center, the ENT Department, St. Charles Hospital, and the Department of Radiology,LapeyronieHospital, Montpellier, France,and the Department of Physiology, University of Turku, Turku, Finland. 2 Correspondence and requests for reprints should be addressed to Olli Polo, M.D., Department of Physiology, University of Turku, Kiinamyllynkatu 10, SF-20520 Turku, Finland.
SNORERS WITH AND WITHOUT SLEEP APNEA
Fig. 1. The setup of the SeSB recording used in this study. For recording of body movements (M), respiratory movements (A), and ballistocardiogram (B) no electrodes need be attached to the subject. After amplification and frequency filtration the movement signals are fed to a polygraph.
1289
~§:§~:~:~:~~§§jsi:~ : : : :~: ::~ §: : :~ :: ~: : :
11":&-:0' : t ' : .i ···.v
~ ..,..
FonH HnTTR ESS
:...': '. ".:-~'. ': ' . 1". ..:.:y ....;··." ··: s ·· :.l h H'g't EHE NT SENSO R
j
M R B
I
~
:\~~\~~ol\~~,\~,==~t1~ t
PREnHPLIFIER + CllnNNEL FILTERING
01'-2 I
Fig. 2. Simple periodic breathing (P-1). different patterns of obstructive periodic breathing (OP-1, OP-2, and OP-3), and the IAA pattern as they appear on a three-channel SeSB recording along with simultaneous measurements of the arterial oxygen saturation (Sa02). BeG = ballistocardiogram. For further details see text.
--..JJJ.~-.u. ~~1J)J.1"... -lollII
"~I
n
11111
m . __
.m
1ft, ,"__ .
S'1t
S9t
.
in
'UI
POLYGRnPIl
01'-1
P-l
. '
mtf
'I"
a'" 15r='
OP-J
.,
.........,....
..
y,~
""--
..........TlOH sec
-- T1lllR(:lOo)
s.o,
IRR WCMWDlI RESARATlOH
..., . '" '" -----n-...-----....-
en
S' -- -
sat .• ,
-
ICG
_T1lllR(:lOo)
.
. s.o,
---o(.----t.....- - - - - - - - - - - -toAOVEMENT ~
1
Fig. 3. A simultaneous SeSB and esophageal pressure (Pe) recording during two episodes of OSA followed by a period of the IAR pattern. The increased intrathoracic pressure variation is associated with increased respiratory variation of the ballistocardiogram (spiking BeG).
body movements, respiratory movements, and the ballistocardiogram (BCG). The BCG reflects the mechanical activity of the heart. The systolic waves are produced by heart contractions and displacement of blood from the heart into the large arteries (13). Normal breathing induces a slight respiratory oscillation in the systolic wave amplitude. This oscillation gives the characteristic aspect of spiking BCG when the intrathoracic pressure fluctuations are elevated in connection with OSA or increased respiratory resistance. Partial airway 0 bstruction can therefore be detected by visualizing the secondary hemodynamic changes with the BCG during increased respiratory drive (14, 15). The validity of the SCSB in detecting OSA has previously been demonstrated by several authors (11, 16-19).
Principles for Scoring the SCSB Recordings The three-channel SCSB recordings were
various types of periodic breathing, distinguishing those with a normal or an increased breathing resistance and those with an early movement arousal, respectively. The OP-3pattern is associated with more severe arterial oxygen desaturation than the other periodic breathing events (17). The conventional apnea index and the percentage of the OP-3 pattern are not directly comparable. Presuming there are on average three or four obstructive apneas in a 2-min epoch, 5070 of TIB would equal 4.5 or 6 apneas per hour, respectively. The IRR pattern is associated with increased intrathoracic pressure variation (figure 3). The subject is snoring heavily or, after uvulopalatopharyngoplasty, breathing with a loud wheezing sound (20). A moderate oxygen desaturation is often observed, although desaturation also may be absent. The episode is terminated either by a movement arousal (figure 2) or, more rarely, by a spontaneous waning of the breathing amplitude.
min
BeG
visually analyzed according to simple body movement, respiratory movement, and BCG criteria (10, 20). The frequencies of various periodic breathing patterns and the continuous IRR pattern (figure 2) were determined as percentage of TIB. A 2-min epoch with symmetrically waxing and waning periodic breathing was scored P-1 (Periodic 1) if the breathing episodes were hot accompanied with spiking BCG. If the spikes were present the epoch was scored OP-l (Obstructive Periodic 1). OP-2 was scored if a body movement occurred immediately after the resumption ofthe breathing efforts, and OP-3if the movement arousal occurred only after a period of increasing breathing efforts with associated BCG spiking. The IRR pattern was defined by the presenceof waxing respiratory amplitude associated with spiking BCG, continuing without waning for longer than 1 min. The OP-3 pattern (figure 2) essentially corresponds to OSA (19), whereas the other breathing events (P-l, OP-l, and OP-2) imply
Pharyngeal CT Study For the pharyngeal CT study the subjects were awake, lying supine with the neck placed in a neutral position. Scans were performed at 8-mm intervals during quiet breathing from the posterior part of the hard palate to the hyoid bone (HB). The exposure time was 3 s. The cross-sectional areas were measured with the cursor incorporated in the scanner software at the nasopharynx ([NP] from the scan visualizing the lowest extent of the hard palate), the velopharynx ([VP] the smallest crosssectional area observed in the three to five scans visualizing the soft palate), the tongue base level ([TB] the smallest area of two to three scans) and the HB level. The scanning was repeated at the velopharyngeal (VPi) and the tongue base (TBi) levels after deep inspiration while holding breath at total lung capacity throughout the 3-s exposure time. Statistical Analyses Comparison of variances and testing for normality of the study groups were performed before statistical analyses. Differences in the anthropometric variables between the three . groups were assessed using one-way analysis of variance. The effect of sex was controlled by two-way analysis of variance. Because of unequal groups in terms of body weight and body mass index (BMI), one-way analysis of covariance (ANCOVA) employing both of these variables as covariates was used to obtain the adjusted effects ofthe patient groups. Bonferroni's t test was used as the multiple comparison procedure. The frequency distribution of the three types of patients (OSA, IRR, and control) in the strata with different VP/HB ratios was assessed by the Pearson chi-square test. All the calculations wereperformed with the BMDP software library for microcomputers (21). Results
About 80% of the patients were men (table 1). Women were equally represented
1290
POLO, TAFTI, FRAGA, PORKKA, DEJEAN, AND BILLIARD
TABLE 1
p=.013
ANTHROPOMETRIC DATA
N Women Age, yr Height, m Weight, kg BMI, kg/m'
OSA ± SEM
IRR ± SEM
Control ± SEM
15 3 (20%) 56.2 ± 3.2 1.71 ± 0.02 88.1 ± 3.0' 30.1 ± 1.1'
25 4 (16%) 48.1 ± 1.9 1.71 ± 0.02 82.3 ± 2.9 28.1 ± 1.0
14 2 (14%) 50.6 ± 3.1 1.69 ± 0.02 73.9 ± 2.4' 25.8 ± 0.74'
• Values differ significantly at the level p
< 0.05 (Bcntsrronl t test.
1.0
0.4 0.2
FREQUENCY OF THE VARIOUS BREATHING PATTERNS (PERCENTAGE OF TIME IN BED) OSA ± SEM
IRR ± SEM
± ± ± ± ± 0 81.4 ±
512 3.8 13.0 14.1 0.6 15.1 46.7
TIB, min P-1, % OP-1, % OP-2, % OP-3, % IRR,% All patterns, %
521 3.2 10.8 7.8 59.6
16.8 1.1 3.4 3.3 8.4 4.9
**
± ± ± ± ± ± ±
12.7 2.3 2.3 2.6 0.17 2.3 3.3
Control ± SEM
± ± ± ± 0 0 3.5 ±
530 1.1 0.9 1.5
15.9 0.40 0.37 0.51
0.0
~ ~ ~ V. ~ ~ ~ ~
VP/TB 4.0
** *** n
II
~
TABLE 2
NS
II
0.8 0.6
Olll J. POLO, MEHDI TAFTI, JEAN FRAGA, KIMMO V. K. PORKKA, YVES DEJEAN, and MICHEL BilliARD Introduction
Patients with obstructive sleep apnea (OSA) syndrome report a history of heavy snoring for years or decades before the appearance of daytime somnolence and respiratory pauses during sleep (l). Therefore snoring is an important risk factor for OSA. However, only a small fraction of habitual snorers appear to develop OSA. About 15070 of middle-aged and older men are habitual snorers, yet only about 1 to 4% of the population has OSA syndrome (2). The conditions under which strong inspiratory suction causes complete oropharyngeal occlusion are insufficiently understood. Also, the role of heavy snoring-induced vibration injury in the advancement of pharyngeal anatomic abnormalities observed in OSA (3-5) has not been clarified. According to the law of Bernoulli, local pressure conditions in a tube are determined by the velocity of the airflow (6). The velopharynx is the narrowest segment of the pharyngeal airways. Therefore, it is repetitively exposed to local increased airflow speed and decreased intraluminal pressure. The importance of the kinetic properties of the airflow on the patency of the upper airways is demonstrated by the discrepancy between the relatively low static occlusion pressures (3.5 to 12.0 cm H 2 0 ) observed in snorers (7), and their high tolerance of intrathoracic suction pressures (20 to 60 em H 2 0 ) (8) while air is flowing through. In view of this, we reasoned that snorers without apnea may have a reduced inspiratory suction effect locally in the oropharynx because of proportional differences in upper airway dimensions. Accordingly, we measured the upper airway cross-sectional areas at four levels in patients with OSA, heavy snorers without apnea, and control patients while all were awake. In an attempt to assess the breathing-related dynamic changes in the upper airway morphology we also measured the cross-sectional areas during deep inspiration. 1288
SUMMARY Patients with obstructive sleep apnea (OSA) and heavy snorers without apnea both show Intrathoracic suction pressures during sleep that exceed their static upper airway closing pressures. Complete airway occlusion, however, occurs only In the former patient group. We hypothesized that the kinetic properties of the airflow would be different In these two typas of patients because of differences In upper airway morphology. The pharyngeal computed tomography (CT) was used to measure the cross-sectional areas of the upper airways In 15 patients with OSA, 25 nonapnelc heavy snorers, and 14 control sublects while thay were awake. Nocturnal breathing was monitored with the static charge-sensitive bed (SeSB). The patients with OSA had a narrower airspace at the velopharyngeal (VP) level than the controls (p < 0.01); the nonapnelc snorers did not differ from the other groups. At the tongue base (TB) and the hyoid bone (HB) levels there was no difference between the OSA and the control groups, but the nonapneic snorers had narrower airways at both of these levels compared with control sublects (p < 0.01) and at the hyoid bone level compared with the OSA group (p < 0.05). The VP/HB ratio was the parameter that best distinguished the patients with OSA from the nonapnelc snorers (lower In the OSA group, p < 0.001). We suggest that airway collapse during sleep is favored by a narrow velopharynx associated with large hypopharynx. Some heavy snorers may not have an oropharyngeal collapse because the paak inspiratory suction pressure could already be damped down at the level of the relatively narrow AM REV RESPIR DIS 1991; 143:1288-1293 hypopharyngeal airways.
Methods Upper airway morphology was studied in three groups with clearly distinctive breathing patterns observed in sleep recordings. The OSA group consisted of 15 patients who presented episodes of OSA during at least 5% of the time in bed (TIB) but had no episodes of constant snoring with increased respiratory resistance. The increased respiratory resistance (IRR) group included 25 snorers who showed at least 5070 of increased respiratory load but had less than 5% of OSA. The control group consisted of 14 patients referred to the sleep clinic because of their snoring complaint but for whom the sleep recordings showed neither obstructive apnea nor increased respiratory resistance and in whom minor breathing abnormalities were present during less than 100/0 of TIB. Most of these subjects were not heard snoring during the recording night. Each patient was monitored all night with the static charge-sensitive bed ([SCSBj Biornatrs; Biorec, Thrku, Finland) and ear oximeter (Biox'" II; the BOC Group Inc., Louisville, CO). A pharyngeal computed tomographic (CT) study (Somatom DR3R; Siemens AG, Erlangen, Germany) was performed during the daytime. Static Charge-sensitive Bed The major problem in studying snoring is the lack of suitable techniques for differentiat-
ing heavy snoring with increased respiratory load from trivial snoring with noise as the only complaint. The standard polysomnography allows quantification of sleep apnea and hypopnea but provides no information about the severity of an eventual partial airwayobstruction. For the purpose of this study esophageal pressure measurements would have provided the most exact data. Because the esophageal balloon was not tolerated by all of our subjects, however, the SCSB was the method of choice. The SCSB is a sensitive movement sensor placed under a special foam plastic mattress, allowing noninvasive cardiorespiratory monitoring without electrodes or cables attached to the subjects (figure 1) (9-12). The sensor produces a three-channel output including
(Receivedin originalform November 27, 1989and in revised form November 15, 1990) 1 From the Sleep Disorders Unit, Gui de Chauliac Medical Center, the ENT Department, St. Charles Hospital, and the Department of Radiology,LapeyronieHospital, Montpellier, France,and the Department of Physiology, University of Turku, Turku, Finland. 2 Correspondence and requests for reprints should be addressed to Olli Polo, M.D., Department of Physiology, University of Turku, Kiinamyllynkatu 10, SF-20520 Turku, Finland.
SNORERS WITH AND WITHOUT SLEEP APNEA
Fig. 1. The setup of the SeSB recording used in this study. For recording of body movements (M), respiratory movements (A), and ballistocardiogram (B) no electrodes need be attached to the subject. After amplification and frequency filtration the movement signals are fed to a polygraph.
1289
~§:§~:~:~:~~§§jsi:~ : : : :~: ::~ §: : :~ :: ~: : :
11":&-:0' : t ' : .i ···.v
~ ..,..
FonH HnTTR ESS
:...': '. ".:-~'. ': ' . 1". ..:.:y ....;··." ··: s ·· :.l h H'g't EHE NT SENSO R
j
M R B
I
~
:\~~\~~ol\~~,\~,==~t1~ t
PREnHPLIFIER + CllnNNEL FILTERING
01'-2 I
Fig. 2. Simple periodic breathing (P-1). different patterns of obstructive periodic breathing (OP-1, OP-2, and OP-3), and the IAA pattern as they appear on a three-channel SeSB recording along with simultaneous measurements of the arterial oxygen saturation (Sa02). BeG = ballistocardiogram. For further details see text.
--..JJJ.~-.u. ~~1J)J.1"... -lollII
"~I
n
11111
m . __
.m
1ft, ,"__ .
S'1t
S9t
.
in
'UI
POLYGRnPIl
01'-1
P-l
. '
mtf
'I"
a'" 15r='
OP-J
.,
.........,....
..
y,~
""--
..........TlOH sec
-- T1lllR(:lOo)
s.o,
IRR WCMWDlI RESARATlOH
..., . '" '" -----n-...-----....-
en
S' -- -
sat .• ,
-
ICG
_T1lllR(:lOo)
.
. s.o,
---o(.----t.....- - - - - - - - - - - -toAOVEMENT ~
1
Fig. 3. A simultaneous SeSB and esophageal pressure (Pe) recording during two episodes of OSA followed by a period of the IAR pattern. The increased intrathoracic pressure variation is associated with increased respiratory variation of the ballistocardiogram (spiking BeG).
body movements, respiratory movements, and the ballistocardiogram (BCG). The BCG reflects the mechanical activity of the heart. The systolic waves are produced by heart contractions and displacement of blood from the heart into the large arteries (13). Normal breathing induces a slight respiratory oscillation in the systolic wave amplitude. This oscillation gives the characteristic aspect of spiking BCG when the intrathoracic pressure fluctuations are elevated in connection with OSA or increased respiratory resistance. Partial airway 0 bstruction can therefore be detected by visualizing the secondary hemodynamic changes with the BCG during increased respiratory drive (14, 15). The validity of the SCSB in detecting OSA has previously been demonstrated by several authors (11, 16-19).
Principles for Scoring the SCSB Recordings The three-channel SCSB recordings were
various types of periodic breathing, distinguishing those with a normal or an increased breathing resistance and those with an early movement arousal, respectively. The OP-3pattern is associated with more severe arterial oxygen desaturation than the other periodic breathing events (17). The conventional apnea index and the percentage of the OP-3 pattern are not directly comparable. Presuming there are on average three or four obstructive apneas in a 2-min epoch, 5070 of TIB would equal 4.5 or 6 apneas per hour, respectively. The IRR pattern is associated with increased intrathoracic pressure variation (figure 3). The subject is snoring heavily or, after uvulopalatopharyngoplasty, breathing with a loud wheezing sound (20). A moderate oxygen desaturation is often observed, although desaturation also may be absent. The episode is terminated either by a movement arousal (figure 2) or, more rarely, by a spontaneous waning of the breathing amplitude.
min
BeG
visually analyzed according to simple body movement, respiratory movement, and BCG criteria (10, 20). The frequencies of various periodic breathing patterns and the continuous IRR pattern (figure 2) were determined as percentage of TIB. A 2-min epoch with symmetrically waxing and waning periodic breathing was scored P-1 (Periodic 1) if the breathing episodes were hot accompanied with spiking BCG. If the spikes were present the epoch was scored OP-l (Obstructive Periodic 1). OP-2 was scored if a body movement occurred immediately after the resumption ofthe breathing efforts, and OP-3if the movement arousal occurred only after a period of increasing breathing efforts with associated BCG spiking. The IRR pattern was defined by the presenceof waxing respiratory amplitude associated with spiking BCG, continuing without waning for longer than 1 min. The OP-3 pattern (figure 2) essentially corresponds to OSA (19), whereas the other breathing events (P-l, OP-l, and OP-2) imply
Pharyngeal CT Study For the pharyngeal CT study the subjects were awake, lying supine with the neck placed in a neutral position. Scans were performed at 8-mm intervals during quiet breathing from the posterior part of the hard palate to the hyoid bone (HB). The exposure time was 3 s. The cross-sectional areas were measured with the cursor incorporated in the scanner software at the nasopharynx ([NP] from the scan visualizing the lowest extent of the hard palate), the velopharynx ([VP] the smallest crosssectional area observed in the three to five scans visualizing the soft palate), the tongue base level ([TB] the smallest area of two to three scans) and the HB level. The scanning was repeated at the velopharyngeal (VPi) and the tongue base (TBi) levels after deep inspiration while holding breath at total lung capacity throughout the 3-s exposure time. Statistical Analyses Comparison of variances and testing for normality of the study groups were performed before statistical analyses. Differences in the anthropometric variables between the three . groups were assessed using one-way analysis of variance. The effect of sex was controlled by two-way analysis of variance. Because of unequal groups in terms of body weight and body mass index (BMI), one-way analysis of covariance (ANCOVA) employing both of these variables as covariates was used to obtain the adjusted effects ofthe patient groups. Bonferroni's t test was used as the multiple comparison procedure. The frequency distribution of the three types of patients (OSA, IRR, and control) in the strata with different VP/HB ratios was assessed by the Pearson chi-square test. All the calculations wereperformed with the BMDP software library for microcomputers (21). Results
About 80% of the patients were men (table 1). Women were equally represented
1290
POLO, TAFTI, FRAGA, PORKKA, DEJEAN, AND BILLIARD
TABLE 1
p=.013
ANTHROPOMETRIC DATA
N Women Age, yr Height, m Weight, kg BMI, kg/m'
OSA ± SEM
IRR ± SEM
Control ± SEM
15 3 (20%) 56.2 ± 3.2 1.71 ± 0.02 88.1 ± 3.0' 30.1 ± 1.1'
25 4 (16%) 48.1 ± 1.9 1.71 ± 0.02 82.3 ± 2.9 28.1 ± 1.0
14 2 (14%) 50.6 ± 3.1 1.69 ± 0.02 73.9 ± 2.4' 25.8 ± 0.74'
• Values differ significantly at the level p
< 0.05 (Bcntsrronl t test.
1.0
0.4 0.2
FREQUENCY OF THE VARIOUS BREATHING PATTERNS (PERCENTAGE OF TIME IN BED) OSA ± SEM
IRR ± SEM
± ± ± ± ± 0 81.4 ±
512 3.8 13.0 14.1 0.6 15.1 46.7
TIB, min P-1, % OP-1, % OP-2, % OP-3, % IRR,% All patterns, %
521 3.2 10.8 7.8 59.6
16.8 1.1 3.4 3.3 8.4 4.9
**
± ± ± ± ± ± ±
12.7 2.3 2.3 2.6 0.17 2.3 3.3
Control ± SEM
± ± ± ± 0 0 3.5 ±
530 1.1 0.9 1.5
15.9 0.40 0.37 0.51
0.0
~ ~ ~ V. ~ ~ ~ ~
VP/TB 4.0
** *** n
II
~
TABLE 2
NS
II
0.8 0.6
