| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Royal Prince Alfred Hospital (Ms. Milross; Drs. Piper, Grunstein, and Bye; Mr. Norman, and Mr. Willson), Camperdown,Sydney; and Faculty of Medicine (Dr. Sullivan), University of Sydney, Sydney, Australia.
Correspondence to: Peter T. P. Bye, MD, FCCP, Department of Respiratory Medicine, Royal Prince Alfred Hospital, Missenden Rd, Camperdown NSW 2050, Australia; e-mail: peterb{at}mail.med.usyd.edu.au
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Design: Cross-sectional analysis of sleep studies, lung function, respiratory muscle strength, and evening and morning arterial blood gas measurements in patients with stable CF. A questionnaire addressing sleep quality was administered. Forward stepwise regression analysis was used to identify the parameters that best predict sleep-related desaturation, hypercapnia, and respiratory disturbance.
Setting: Sleep investigation unit and lung function laboratory.
Patients: Thirty-two patients with CF and FEV1 < 65% predicted, in stable clinical condition. Patients were aged 27 ± 8 years (mean ± 1 SD) with FEV1 of 36 ± 10% predicted, evening PaO2 of 68 ± 8 mm Hg, and PaCO2 of 43 ± 5 mm Hg.
Results: Evening PaO2 (p < 0.0001) and morning PaCO2 (p < 0.01) were predictive of the average minimum oxyhemoglobin saturation per 30-s epoch of sleep (r2 = 0.74; p < 0.0001). Evening PaO2 (p < 0.001) was predictive of the rise in transcutaneous carbon dioxide (TcCO2) seen from non-rapid eye movement (NREM) to rapid eye movement (REM) sleep (r2 = 0.37; p < 0.001). In addition, there was some relationship between expiratory respiratory muscle strength and the REM respiratory disturbance index (r2 = 0.22; p < 0.01).
Conclusion: Evening PaO2 was found to contribute significantly to the ability to predict both sleep-related desaturation and the rise in TcCO2 from NREM sleep to REM sleep in this subgroup of patients with CF.
Key Words: cystic fibrosis • hypoxemia • sleep-disordered breathing
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Awake resting SpO2 < 94% and an FEV1 < 65% predicted have been suggested5 6 as the variables with the most significant correlation with sleep-related desaturation in patients with CF. There is limited information on the predictive value of daytime PaCO2 in sleep-disordered breathing, and no studies have assessed predictors of sleep-related changes in carbon dioxide levels or indexes of nocturnal respiratory disturbance in patients with CF. The aim of this study was to compare sleep variables of oxygenation, changes in carbon dioxide, and respiratory disturbance to daytime measurements of lung function, respiratory muscle strength, arterial blood gas (ABG) measurements, and subjective sleep quality in patients with CF. Furthermore, we aimed to examine for the ability of daytime variables to have significant predictive value for sleep-disordered breathing.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This study was conducted at Royal Prince Alfred Hospital in Sydney, Australia, and approved by the ethics committee of our institution (protocol No. X97–0204). Written, informed consent was obtained from all patients.
Anthropometric and Lung Function Measurements
Anthropometric and lung function data were obtained on the day
of the diagnostic sleep study. Measurements of spirometry were
performed using a mass flow sensor (Sensormedics Vmax 20; Sensormedics
Corporation; Yorba Linda, CA), which was calibrated before each study
and compared with normal predicted values of Quanjer et
al.8
Lung volumes were determined by body plethysmography
(Gould 2800; Gould Electronics; Dayton, OH). Results were compared with
normal predicted values of Goldman and Becklake.9
Inspiratory muscle pressure at residual volume (PImax) and
expiratory muscle pressure at total lung capacity (PEmax)
were recorded using a hand-held pressure gauge, and the results were
compared with normal predicted values of Wilson et al.10
A
gas chromatograph (1085 D Series PF/Dx; Medical Graphics Corporation;
St. Paul, MN) was used to measure carbon monoxide transferred per liter
of lung volume (KCO). The normal predicted value for
KCO was 6.9 mL CO/mm Hg/min (standard temperature and
pressure, dry), a value based on mean laboratory values for
normal nonsmoking healthy adults obtained in our laboratory.
ABG Tensions
Initial awake ABG tensions were obtained with the patient seated
and breathing room air, usually in the late afternoon prior to the
sleep study. Morning ABG measurements were taken immediately on
awakening, so that any delay from sleep to wakefulness was minimized.
Sleep Study Recordings
During polysomnography, continuous recordings were made on a
computerized system (Sleepwatch; Compumedics; Melbourne, Australia).
Sleep stage was determined from two channels of EEG (C3/A2, O2/A1), two
channels of electro-oculogram (right outer canthus/A1, left outer
canthus/A2), and from the submental electromyogram (EMG). Respiratory
variables were monitored using abdominal and thoracic impedance bands
for chest wall movement and diaphragm EMG electrodes to reflect
respiratory effort. Nasal airflow was measured using nasal prongs
attached to a flow sensor (AutoSet; ResMed; Sydney, Australia).
SpO2 was measured with a finger probe
(model 3700e; Ohmeda; Boulder, CO). The resting awake value of
SpO2 was noted, and
SpO2 recording continued overnight.
Transcutaneous carbon dioxide (TcCO2)
was also measured continuously overnight (TCM3; Radiometer; Copenhagen,
Denmark). TcCO2 and
SpO2 were recorded simultaneously on
both the Sleepwatch system as well as a personal computer.
Sleep stages were scored in 30-s epochs according to the standard
criteria of Rechtschaffen and Kales.11
An EEG arousal was
defined as an abrupt increase in EEG frequency for 
3 s that in REM
sleep was accompanied by an increase in submental EMG amplitude. Sleep
efficiency was defined as the total sleep time (TST) as a percentage of
the time available for sleep. Apnea was defined as cessation of airflow
for 10 s, or a cessation of airflow for < 10 s with an oxygen
desaturation of 3% or an arousal. Hypopnea was defined as a
reduction in amplitude of airflow, or thoracoabdominal wall movement of
> 50% for 10 s, or a reduction in airflow or
thoracoabdominal wall movement of > 50% for < 10 s if it was
accompanied by an oxygen desaturation of 3% or an arousal. The
number of apneas and hypopneas per hour of non-rapid eye movement
(NREM), REM, and TST were calculated and reported as the respiratory
disturbance index (RDI).
The absolute minimum sleep SpO2 was
documented for the entire night but also for REM and NREM sleep. The
percentage of TST, REM, and NREM sleep time with
SpO2 
90% was calculated. The
minimum average SpO2
(SpO2min·av) was calculated as the
mean of the minimum value for SpO2 in
each 30-s epoch of sleep.
SpO2min·av was calculated for TST,
NREM sleep time, and REM sleep time.
Respiratory events leading to desaturation and increases in carbon
dioxide occur predominantly in REM sleep in the CF population. These
changes in carbon dioxide can be represented by the change in
TcCO2 from NREM to REM sleep, or the
maximum TcCO2 measured during the
night compared with a baseline TcCO2
reading. The change in TcCO2 from
NREM to REM sleep in our study was calculated for each REM period. Due
to the drift often seen in the TcCO2
trace, a line of best fit was drawn between four points on the
TcCO2 trace: one at the start and end
of each REM period, taking into account the approximately 3-min time
delay for the device, and a point 5 min prior to and following each REM
period as long as the TcCO2 reading
was stable. A perpendicular line was then drawn to the peak
TcCO2 reading for that REM period
being measured (Fig 1
). A weighted average was then obtained for the
TcCO2 for each subject.
TcCO2 awake was the baseline value
recorded at least 7 min after probe placement while the patient was
breathing spontaneously.
|
Data Analysis and Statistics
Linear correlation analyses were performed between awake and
sleep measurements, but due to the large number of correlations being
performed, we chose to analyze data by fitting the variables into a
forward stepwise regression model to determine the most important
predictor variables. Daytime variables used as potential predictor
variables included lung function variables (FEV1
percent predicted, FVC percent predicted, total lung capacity [TLC]
percent predicted, residual volume [RV] percent predicted, RV/TLC
percent predicted, functional residual capacity percent predicted,
KCO percent predicted); respiratory muscle strength
measurements (PImax percent predicted and PEmax
percent predicted); anthropometric values (age, sex, and body mass
index [BMI]); evening and morning ABG measurements (pH,
PaO2,
PaCO2); and the PSQI score. In order
to reduce the possibility of a type-1 error, only three sleep variables
were chosen as outcome variables:
SpO2min · av,
TcCO2 from NREM to REM sleep, and
REM RDI.
As awake resting SpO2 and evening PaO2 are both measurements of the level of arterial blood oxygenation, it is logical that they are correlated highly with one another. Therefore, when fitting them into a model of forward stepwise regression, once one is selected the other does not provide any additional information. Hence, in our forward stepwise regression calculations, evening PaO2 was included as a potential predictor variable and awake resting SpO2 was not. Data are reported as mean ± 1 SD. Results were considered statistically significant at the p < 0.05 level.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
94% had an
SpO2min·av < 90%, while 12 of 16
subjects with a resting SpO2 < 94%
had an SpO2min·av 90%. Thirty
of 32 patients had an absolute minimum sleep
SpO2 of 90% (Table 2
). Subjects had a mean awake TcCO2 of
47 ± 7 mm Hg, ranging from 37 to 70 mm Hg. The maximum sleep
TcCO2 was 55 ± 10 mm Hg, ranging
from 41 to 82 mm Hg. The rise in
TcCO2 from NREM to REM sleep was
2.1 ± 1.6 mm Hg, ranging from an increase of zero to 5.6 mm Hg
(Table 2)
. The mean values for REM, NREM, and TST RDI are shown in
Table 3
. Subjects had a mean NREM RDI of 0.7 ± 1.3 events per hour, REM RDI
of 12.2 ± 11.4 events per hour, and TST RDI of 3.2 ± 3.4 events
per hour.
|
|
90% (r = 0.79; p < 0.0001). Therefore,
SpO2min · av was chosen as the
parameter to represent sleep-related oxygenation for the calculations
of the relationship with lung function parameters, respiratory muscle
strength, BMI, awake resting
SpO2 percentage, evening
and morning ABG measurements, and the PSQI. Awake SpO2 (r = 0.77; p < 0.0001), evening PaO2 (r = 0.78; p < 0.0001), morning PaO2 (r = 0.70; p < 0.0001), evening PaCO2 (r = -0.59; p < 0.01), morning PaCO2 (r = - 0.66; p < 0.0001), and morning pH (r = 0.50; p < 0.01) were significantly correlated with SpO2av percentage. The only lung function parameter to show a significant correlation with the SpO2av was FEV1 percent predicted (r = 0.45; p < 0.01). PImax was also significantly correlated to SpO2av (r = 0.44; p < 0.05). The PSQI did not correlate significantly with SpO2av.
Forward stepwise regression revealed the combination of evening
PaO2 (p < 0.0001) and morning
PaCO2 (p < 0.01) to most strongly
predict sleep-related oxygenation as represented by
SpO2av
(r2 = 0.74; p < 0.0001; Fig 2
, A and B). The remaining variables did not add
significantly to the ability of the evening
PaO2 and morning
PaCO2 to account for the
variability in TST SpO2av.
With simple regression, the evening
PaO2 accounted for 61%
(r2 = 0.61; p < 0.0001) of the
variability in SpO2av,
while in combination with the morning
PaCO2, using forward
stepwise regression, 74% of the variability in
SpO2av could be explained.
The equation to calculate the predicted
SpO2av percentage would be
as follows:
 |
|
TcCO2) sleep are correlated
with one another (r = 0.51; p < 0.01), and since the
size of the TcCO2
reflects the severity of the REM-related hypoventilation, we examined
the relationship of the
TcCO2 with daytime
measurements of lung function, respiratory muscle strength, awake
resting SpO2, evening and
morning ABG measurements, and the PSQI.
Awake SpO2
(r = - 0.66; p < 0.0001), evening
PaO2
(r = - 0.61; p < 0.01), morning
PaO2 (r
=- 0.53; p < 0.01), evening
PaCO2
(r = 0.45; p < 0.05), and morning
PaCO2
(r = 0.50; p < 0.01) were significantly correlated with
TcCO2. Lung
function parameters that showed a significant correlation with
TcCO2 were
FEV1 percent predicted
(r = - 0.47; p < 0.01), RV percent predicted
(r = 0.43; p < 0.05), and RV/TLC percent predicted
(r = 0.39; p < 0.05). PEmax
percent predicted was also significantly correlated with
TcCO2
(r = - 0.47; p < 0.05). PSQI did not correlate
significantly with
TcCO2.
Evening PaO2 (p < 0.001) was the
only variable identified in a forward stepwise regression model as
being predictive of the change in carbon dioxide from NREM to REM sleep
as represented by TcCO2
(r2 = 0.37; p < 0.001; Fig 3
). Other measured variables did not add significantly to the ability of
evening PaO2 to account for
the variability seen in
TcCO2. The equation to
calculate the predicted change in
TcCO2 from NREM to REM
sleep is as follows:
 |
|
PEmax percent predicted was the only variable to show any
correlation with REM RDI (r = - 0.47; p < 0.01).
Similarly, in a forward stepwise regression,
PEmax percent predicted was found to be the only
variable with predictive power for REM RDI
(r2 = 0.22; p < 0.01). The
equation to calculate the predicted REM RDI is as follows:
 |
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Previously, Versteegh and colleagues5
showed that an awake
resting SpO2 of 93.8% was the most
discriminatory for predicting sleep-related
SpO2, defined as one hourly mean
SpO2 of 90%, with a positive
predictive value of 50%, and that the addition of other variables,
including exercise parameters and lung function, did not add
significantly to the discriminatory power.5
ABG tensions,
measures of carbon dioxide levels during sleep, or polysomnography
allowing for sleep staging and respiratory event scoring were not
measured in that study. Our data have demonstrated the additional value
of measuring morning PaCO2 along with
evening PaO2 for predicting
sleep-related SpO2, and provides the
clinician with a predictive equation that describes a continuum rather
than a probability (50%) at a single point (93.8%). In addition, the
current study is to our knowledge the first to examine the ability to
predict changes in carbon dioxide and indexes of respiratory
disturbance during sleep in this patient group.
An interesting observation from our work was that 3 of 16 patients,
despite an awake resting SpO2 of
94%, had an SpO2av of < 90%.
Sleep-related hypoxemia would have been missed if sleep studies were
not performed in these patients. Frangolias and
colleagues15
recently reported sleep studies in a group of
patients with CF, with a wide spectrum of lung disease. This study was
designed to assess for the ability to predict nocturnal
SpO2 using daytime pulse oximetry,
exercise testing, and spirometry. There was no assessment of
respiratory disturbance, sleep-related carbon dioxide changes,
measurements of ABG tensions, nor full polysomnography. All patients
with an awake resting SpO2 of
< 93% desaturated nocturnally, but there was a heterogenous response
with regard to nocturnal desaturation with values of awake resting
SpO2 of > 93%.15
Nocturnal desaturation was found to be uncommon in patients with milder
lung disease (FEV1 > 65% predicted); however,
when using FEV1 and awake resting
SpO2 alone, they were only able to
correctly predict 26% of all patients with clinically significant
nocturnal desaturation, which they defined as
SpO2 < 90% for > 5% of the
night.15
Further studies in patients with CF and mild lung
disease, using full polysomnography and a comprehensive battery of
tests including complete pulmonary function testing and ABG tension
measurements, may allow us to better determine if nonsleep study
measurements can predict when patients with CF first present with
sleep-disordered breathing.
Previous studies have used various methods of describing desaturation
during sleep. However, a definitive measure of sleep
SpO2 has yet to be determined.
Variables described have included the greatest percentage fall in
SpO2,3
16
the percentage
of time spent with SpO2 > 90% and
< 90%,6
minimum sleep
SpO2,1
2
3
mean sleep
SpO2,14
lowest hourly
mean,5
and mean minimum
SpO2.17
In this study,
we reported the relationship between the minimum sleep
SpO2,
SpO2av, and percentages of time with
SpO2 90%. In order to reduce the
possibility of type-1 error, only one measure of oxygenation during
sleep could be chosen as an outcome variable in this study. As the
absolute minimum sleep SpO2 may not
appropriately reflect the subject’s oxygenation for the entire night,
we reasoned that the SpO2av more
accurately quantifies sleep oxygenation due to the high sampling rate,
and it was therefore chosen as the variable to reflect oxygenation
during sleep in this study.
Although Versteegh and colleagues5 found that sleep-related desaturation occurred only in subjects with an FEV1 < 65% predicted in their study, spirometry did not add significantly to the discriminatory power of resting SpO2 to predict sleep-related desaturation. Our study did not include subjects with an FEV1 > 65% of predicted, and thus the range in spirometry was limited. Our results in patients with moderate and severe lung disease showed no additional predictive information to be gained by lung function parameters when performed in conjunction with ABG tension measurement. Similarly, in a study of overnight pulse oximetry in patients with CF, FEV1 was shown to be poorly correlated with SpO2 during sleep, defined as percentage of TST with SpO2 < 90%.6 Other studies that have examined for relationships between lung function and overnight oximetry have reported significant correlations with measures of severity of lung disease, including measures of lung hyperinflation3 13 and airflow obstruction.3 6 14 No significant correlations were shown between respiratory muscle strength or nutritional status in a previous study13 of sleep-related desaturation in patients with CF. Despite the individual correlations seen between measurements of lung function and sleep-related desaturation in the above-mentioned studies as well as this present study, we conclude that for patients with moderate-to-severe lung disease, lung function measurements when performed in conjunction with ABG measurements do not add to our ability to predict desaturation during sleep.
A novel feature of this study was the assessment of the ability to predict changes in carbon dioxide levels during sleep and indexes of nocturnal respiratory disturbance. In the forward stepwise regression model, the evening PaO2 was predictive of the TcCO2 from NREM to REM sleep, but it accounted for a lesser degree of the variability than the variables that were found to be predictive of SpO2av.
PEmax percent predicted was found to correlate significantly with and be predictive of REM RDI, although it accounted for only 22% of the variability in the REM RDI. PEmax may reflect the strength of cough. Cough has been described as contributing significantly to sleep disruption in patients with CF,16 although the contribution of cough to sleep disruption in this study was not quantified.
In this study, we have shown that evening PaO2 and morning PaCO2, rather than measurements of lung function, were predictive of nocturnal oxygenation as represented by TST SpO2av in patients with CF and moderate-to-severe lung disease. Evening PaO2 was found to contribute significantly to the ability to predict both nocturnal desaturation and the rise in TcCO2 from NREM to REM sleep.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
This research was performed at Royal Prince Alfred Hospital, Camperdown, Sydney, Australia.
These studies were supported by the National Health and Medical Research Council of Australia.
Received for publication November 13, 2000. Accepted for publication April 24, 2001.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  D S Urquhart, H Montgomery, and A Jaffe Assessment of hypoxia in children with cystic fibrosis Arch. Dis. Child., November 1, 2005; 90(11): 1138 - 1143. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. J. Dobbin, D. Bartlett, K. Melehan, R. R. Grunstein, and P. T. P. Bye The Effect of Infective Exacerbations on Sleep and Neurobehavioral Function in Cystic Fibrosis Am. J. Respir. Crit. Care Med., July 1, 2005; 172(1): 99 - 104. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Jankelowitz, K. J. Reid, L. Wolfe, J. Cullina, P. C. Zee, and M. Jain Cystic Fibrosis Patients Have Poor Sleep Quality Despite Normal Sleep Latency and Efficiency Chest, May 1, 2005; 127(5): 1593 - 1599. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
