Download Oxygen Supplementation and Cardiac

Oxygen Supplementation and
Cardiac-Autonomic Modulation
in COPD*
Matthew N. Bartels, MD, MPH; John M. Gonzalez, BS; Woojin Kim, BS; and
Ronald E. De Meersman, PhD
Study objectives: Patients with COPD have an increased sympathetic modulation and reduced
baroreflex sensitivity (BRS). Therefore, we studied the effects of breathing 31% supplemental
oxygen (SuppO2) on autonomic modulation in a group of COPD patients.
Design: We measured autonomic modulation before and during the administration of SuppO2 on
51 patients with COPD using time-frequency analysis of R-R intervals and BP before and after
intervention. This was done via a counterbalanced crossover design. The BRS index was
determined using the sequence method.
Results: Significant differences were seen in oxygen saturation levels following breathing with
SuppO2 ([mean ⴞ SD] 96.4 ⴞ 1.5%) when compared to those seen after breathing with compressed air (CA) (92.8 ⴞ 2.9%; p < 0.0001). Significant increases were seen in the natural
log-transformed high-frequency modulation (HFln) (SuppO2, 10.8 ⴞ 1.3 natural logarithm [ln]
ms2/Hz; CA, 10.6 ⴞ 1.3 ln ms2/Hz; p < 0.028) and BRS (SuppO2, 3.3 ⴞ 2.2 ms/mm Hg; CA,
2.8 ⴞ 1.8 ms/mm Hg) following the supplemental oxygen treatment (p < 0.015). The lowfrequency/high-frequency ratio of heart rate variability revealed significant differences between
the two treatments (SuppO2, 2.7 ⴞ 1.2; CA, 3.1 ⴞ 1.3; p < 0.008). The analysis of BP variability
data revealed significant decreases in the HFln (CA, 6.9 ⴞ 1.0 mm Hg2/Hz; SuppO2, 6.5 ⴞ 1.2 mm
Hg2/Hz; p < 0.0001). Hemodynamic data also revealed a decrease in mean heart rate after
breathing SuppO2 compared with that after breathing CA (CA, 87.3 ⴞ 13.3 beats/min; SuppO2,
85.0 ⴞ 12.4 beats/min; p < 0.0004). The arterial pulse pressure significantly decreased when
breathing SuppO2 compared with that when breathing CA (CA, 57.2 ⴞ 13.5 mm Hg; SuppO2,
53.3 ⴞ 13.0 mm Hg; p < 0.0023).
Conclusion: Oxygen supplementation in COPD patients significantly and favorably alters autonomic modulation.
(CHEST 2000; 118:691– 696)
Key words: autonomic modulation; baroreflex sensitivity; chronic obstructive lung disease; COPD; oxygen supplementation
Abbreviations: BRS ⫽ baroreceptor sensitivity; CA ⫽ compressed air; HFln ⫽ log-transformed high-frequency modulation; ln ⫽ natural logarithm; SatO2 ⫽ oxygen saturation; SBP ⫽ systolic BP; SuppO2 ⫽ 31% supplemental oxygen;
TF ⫽ time-frequency
and hypercapnia have been shown to
H ypoxia
increase sympathetic nerve activity. Saito et al
1
demonstrated that the degree of hypoxia correlated
with the degree of muscle nerve activity. Other
*From the Human Performance Laboratory, Presbyterian Hospital of the New-York Presbyterian Hospital, New York, NY; and
the Department of Rehabilitation Medicine, College of Physicians and Surgeons of Columbia University, New York, NY.
This research was supported in part through the Rehabilitation
Medicine Scientist Development Program Fellowship, NIH-1K12-HD01097– 01A1, 1998 –2000.
Manuscript received November 16, 1999; revision accepted April
11, 2000.
Correspondence to: Matthew N. Bartels, MD, MPH, 180 Fort
Washington Ave, Harkness Pavilion, Room 184, Suite 199, New
York, NY 10032; e-mail: [email protected]
studies have suggested that hypercapnia also leads to
increases in muscle sympathetic traffic and that
combined hypercapnia and hypoxia synergistically
increase sympathetic activity2,3 through impaired
baroreceptor-cardiac reflex control.4 Disruption of
autonomic reflexes with increased sympathetic modulation, loss of vagal activity, and altered baroreceptor sensitivity (BRS) have been shown to be major
risk factors for cardiac morbidity and mortality.5–7
Since patients with COPD experience hypoxia and
hypercapnia and since prior preliminary research has
demonstrated reduced BRS in COPD patients,8 we
studied a group of patients with severe COPD to
evaluate their autonomic modulation in response to
breathing 31% supplemental oxygen (SuppO2).
CHEST / 118 / 3 / SEPTEMBER, 2000
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21952/ on 05/15/2017
691
Table 1—Patient Characteristics*
Characteristics
Values (n ⫽ 51)
Age, yr
Height, cm
Weight, kg
BMI, kg/m2
63.18 ⫾ 7.92
168.01 ⫾ 9.80
65.79 ⫾ 16.91
23.84 ⫾ 4.14
*Values given as a mean ⫾ SD. BMI ⫽ body mass index.
Materials and Methods
Subjects
Fifty-one patients with COPD were enrolled in the study. See
Tables 1 for the subject characteristics and Table 2 for pulmonary
profile data. All studies were carried out in the Human Performance Laboratory of the College of Physicians and Surgeons of
Columbia University. All procedures were approved by the
Human Subject Committee of Columbia University. Following
the signing of the consent form, subjects were prepared for the
data collection protocol. Subjects were recruited sequentially
from the patients referred for exercise testing. All enrolled
subjects had received established medical diagnoses of COPD
with no other active respiratory disease. All subjects were
instructed to maintain their normal daily routine and medications
on the day of data collection but to refrain from taking any
short-acting bronchodilators on the morning of data collection.
Subjects were randomized to the order of the treatments.
Design
A counterbalanced crossover design after randomization was
used.
Table 2—Pulmonary Profile Data*
Characteristics
MVV
L/min
% Normal
FEV1
L
% Normal
FVC
L
% Normal
DLCO
mL/min/mm Hg
% Normal
MIP
CM H2O
% Normal
MEP
CM H2O
% Normal MEP
Arterial pH
Paco2, mm Hg
Pao2, mm Hg
Values (n ⫽ 51)
30.55 ⫾ 15.26
30.54 ⫾ 14.53
0.73 ⫾ 0.35
28.94 ⫾ 11.68
1.90 ⫾ 0.79
54.55 ⫾ 18.03
6.99 ⫾ 3.45
30.22 ⫾ 12.42
60.76 ⫾ 16.94
82.20 ⫾ 92.70
117.08 ⫾ 35.35
75.00 ⫾ 26.40
7.42 ⫾ 0.03
42.24 ⫾ 8.64
65.14 ⫾ 11.87
*Values given as a mean ⫾ SD. MVV ⫽ maximum voluntary ventilation; FVC ⫽ forced vital capacity; Dlco ⫽ Diffusing capacity of
the lung for carbon monoxide; MIP ⫽ maximum inspiratory pressure; MEP ⫽ maximum expiratory pressure; % Normal ⫽ percent
of standard predicted for normal values.
Statistical Analysis
Paired t test comparisons were made between the two treatments for all dependent variables. Due to the number of t tests
performed, we subjected all analyses to a Bonferroni correction,
and p ⬍ 0.0125 was considered to be significant.
Data Acquisition
Preparatory Phase: Subjects were seated, at rest, and postabsorptive for at least 4 h and had refrained from consuming alcohol
and caffeinated beverages for at least 12 h prior to the test. A
multilead ECG was attached to the subject (model Max-1;
Marquette Medical Systems; Milwaukee, WI). Respiration was
recorded via amplified signals recorded from a thermistor (YSI
reusable temperature probe; Yellow Springs, OH) placed under
the subject’s nostril. A radial artery pulse sensor for BP was
attached to the subject’s wrist to record beat-to-beat BP (model
7000; Colin Medical Instruments; San Antonio, TX). The BP was
measured with the BP cuff at heart level on the right arm. Pulse
oximetry was recorded using an index finger pulse oximeter
(Sat-Trak Pulse Oximeter 767589 –103; SensorMedics; Yorba
Linda, CA). The data were gathered while the subject was seated
in an upright position. The environment in which they were
seated was quiet with dimmed lighting, and subjects equilibrated
to the environmental conditions for 15 min prior to the initiation
of any data collection.
Data Collection: To remove bias, a Venturi mask was placed on
the subject and either SuppO2 or compressed air (CA) was
delivered to the subject in a random order. There were two
data-gathering protocols. Protocol 1 comprised the following: 15
min of equilibration while breathing O2; 5 min of data collection
while breathing O2; followed by 15 min of equilibration while
breathing CA; and 5 min of data collection while breathing CA.
Protocol 2 comprised the following: 15 min of equilibration while
breathing CA; 5 min of data collection while breathing CA;
followed by 15 min of equilibration while breathing O2; and 5
min of data collection while breathing O2. Randomization to a
protocol was based on the patient’s hospital identification number (odd number, protocol 1; even number, protocol 2). The total
test duration was approximately 40 min (two trials with 15 min of
equilibration on a new gas mixture in between). All data collection epochs were 5 min in duration, which is in accordance with
the standards put forth by the Task Force on Standards of Heart
Rate Variability Procedures.9
Data gathering sessions included ECG recordings using lead II
and lead V5, beat-by-beat BP recordings, oxygen saturation
(SatO2) levels, and respiration. During data collection, the subjects were instructed to breathe at a rate of 12 breaths/min for all
collection periods. Breathing was regulated through the use of a
pacing light that cycled 12 times a minute. All biopotentials were
channeled through an interface board (BNC 2080; National
Instruments; Austin TX) and were fed into a 12-bit analog-todigital converter (DAQCard-700; National Instruments) and then
into a computer (Vision Book Plus; Hitachi; San Jose, CA) with a
high-speed processor (Pentium; Intel; San Jose, CA). Postacquisition data analyses were carried out separately. All data were
stored on the hard drive of the computer and were backed up
daily using cartridges for exterior computer drives (Zip drive, Jaz
drive; Iomega; Roy, UT).
Data Analysis
Time-Frequency Analysis: The current investigation used time
and frequency or the Wigner-Ville distribution to decompose
heart rate variability. This analysis is preferred over the traditional frequency spectral analysis due to the nonstationary nature
692
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21952/ on 05/15/2017
Clinical Investigations
of respiration in COPD patients. Specifically, the Wigner-Ville
distribution decomposes a signal expressed as a function of time
into a signal expressed as a function of both time and frequency.
In this approach, the signal is divided into a series of short
windows, and the Fourier transform then is calculated for each
section. The problem with this approach is that as the time
window is made shorter the frequency resolution decreases,
resulting in a lessened ability to accurately determine the frequency content. In other words, there is a tradeoff between time
resolution and frequency resolution. The time-frequency (TF)
methods used herein do not suffer from the resolution problems
described above and produce a distribution that is accurate in
both time and frequency. The TF analysis employed herein,
namely, the Wigner-Ville distribution, uses a modified Fourier
transform applied to overlapping sections of the time signal to
develop a joint density function that is dependent on both time
and frequency. It has been shown to provide a reliable estimate
of spectral powers without generating undesirable cross-terms.
Furthermore, the Wigner-Ville distribution provides the most
accurate estimate and the highest resolution.10,11 For a more
detailed description on the Wigner-Ville distribution method of
analysis, we refer the reader to Novak.12 Due to the skewness of
the data, a log transformation was performed.
BRS Assessment: The spontaneous beat-by-beat interactions of
systolic BP (SBP) and R-R interval reflect true baroreflex events
rather than random interactions.13 The R-R intervals were delineated via a peak detection algorithm described in previous work
from our laboratory.14 The spontaneous baroreflex response was
determined from the beat-to-beat changes in the R-R interval
and SBP by a modification of the technique of Bertinieri et al.15
Linear regression was performed on any episode of three consecutive beats with R-R interval lags in the same direction (either
up or down), and the resultant data yield an evaluation of the
baroreflex sensitivity.13 Verification of the correct breathing
frequency of 12 breaths/min was made by visually counting the
respiratory trace on the computer display panel as well as by
spectrally decomposing the respiratory wave.
BP Variability Analysis
TF analysis was performed in a similar fashion to the heart rate
variability analysis. In this technique, the systolic peaks are
detected via a peak detection algorithm and then are interpolated
and spectrally decomposed via the Wigner-Ville distribution.
Mean Heart Rate Analysis: Using the peak detection algorithm
described in the article by Volterrani et al,8 the mean interbeat
intervals were calculated for the sample and were averaged to
arrive at a mean heart rate.
Mean Pulse Pressure Analysis: Using the peak detection algorithm described in the article by De Meersman et al,14 the mean
SBP and diastolic BP were calculated as SBP-diastolic BP from
an average of the entire sample.
Figure 1. Comparison of the HFln of heart rate while breathing
SuppO2 and while breathing CA.
CA, 2.8 ⫾ 1.8 msec/mm Hg; p ⬍ 0.015) following
the SuppO2 treatment. See Figure 1 for HFln data
and Figure 2 for BRS data. The low-frequency
modulation/HFln ratio of heart rate variability revealed significant differences between the two treatments (SuppO2, 2.7 ⫾ 1.2; CA, 3.1 ⫾ 1.3; p ⬍ 0.008)
(Fig 3). The analysis of BP variability data revealed
that there was a significant decrease in the logtransformed HFln (CA, 6.9 ⫾ 1.0 modulations/mm
Hg2/Hz; SuppO2, 6.5 ⫾ 1.2 ln mm Hg2/Hz;
p ⬍ 0.0001) (Fig 4). The log-transformed low-frequency modulation of BP variability did not reveal
any significant differences.
Hemodynamic data revealed a decrease in mean
heart rate while breathing SuppO2 compared to that
while breathing CA (CA, 87.3 ⫾ 13.3 beats/min;
SuppO2, 85.0 ⫾ 12.4 beats/min; p ⬍ 0.0004) (Fig 5).
The arterial pulse pressure significantly decreased
during breathing with SuppO2 vs breathing with CA
(CA, 57.2 ⫾ 13.5 mm Hg; SuppO2, 53.3 ⫾ 13.0 mm
Hg; p ⬍ 0.0023) (Fig 6).
Results
Significant differences were seen in SatO2 levels
following breathing with SuppO2 compared to those
following breathing with CA (mean [⫾ SD] SuppO2,
96.4 ⫾ 1.5%; CA, 92.8 ⫾ 2.9%; p ⬍ 0.0001). For the
autonomic parameters, significant increases were
seen in the natural log-transformed high-frequency
modulation (HFln) (SuppO2, 10.8 ⫾ 1.3 natural logarithm [ln] ms2/Hz; CA, 10.6 ⫾ 1.3 ln ms2/Hz;
p ⬍ 0.028) and BRS (SuppO2, 3.3 ⫾ 2.2 ms/mm Hg;
Figure 2. BRS on a beat-to-beat basis while breathing SuppO2
and while breathing CA.
CHEST / 118 / 3 / SEPTEMBER, 2000
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21952/ on 05/15/2017
693
Figure 3. Comparison of the ratio of natural log-transformed
low-frequency modulation (LFln) to HFln of heart rate while
breathing SuppO2 and CA.
Figure 5. Comparison of the mean heart rate while breathing
SuppO2 and CA.
Discussion
The results of this investigation indicate that oxygen supplementation in COPD patients significantly
and favorably alters autonomic modulation. Specifically, following breathing with SuppO2 there is an
augmentation in high-frequency cardiac modulation,
an increased BRS, a decrease in heart rate, a decrease in high-frequency BP modulation, and a
decrease in the mean pulse pressure. The increase in
high-frequency cardiac modulation is representative
of an increase in vagal tone. The lack of change in
low-frequency modulation may be due to the presence of high sympathetic tone and the use of sympathetic agonists in many of these patients. The
decrease in high-frequency modulation in BP is also
consistent with improved vagal tone and with a
possible modulation of the sympathetic tone.16 Highfrequency BP modulation is known to be dependent
on a combination of sympathetic and vagal tone,
while low-frequency modulation is more purely sympathetic. Since there is little alteration of the lowfrequency component of BP variability, there is also
probably little change in the sympathetic component
of the high-frequency modulation. This leaves the
vagal component of the high-frequency modulation
as the most likely cause of the significant change
seen in the high-frequency modulation. The decreases in pulse pressure and heart rate support
these findings. This autonomic modulation may be
due to a combination of factors, including reduced
work of breathing, increased chemoreceptor sensitivity, and a reduction in pulmonary artery pressures.
Although the change in arterial pulse pressure was
small, it is in the range of values found to correlate
favorably with improved outcomes in a population of
patients with cardiac disease.17,18 The heart rate is
improved significantly, and larger heart rate changes
have been shown to improve cardiac morbidity.19
However, the significance of this small change is
Figure 4. Comparison of the log-transformed HFln of BP while
breathing SuppO2 and CA.
Figure 6. Comparison of the mean arterial pulse pressure while
breathing SuppO2 and CA.
694
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21952/ on 05/15/2017
Clinical Investigations
unclear,20 other than as a marker of increased vagal
tone. The BRS is in the range of high-risk patients in
a population of patients with cardiac disease. These
numbers are in the range of poor outcome expectations as seen in the recent Autonomic Tone and
Reflexes After Myocardial Infarction trial.21 The
improvement in the BRS is approximately 17%,
which is within the range that has been shown to
improve outcomes in cardiac patients.22 Although we
are concerned with a different population in this
study, the ability to induce a change with treatment
may imply an improvement as well for individuals
with COPD.
As hypercapnia and hypoxemia are strong stimulants to sympathetic activity, a higher basal autonomic tone could be expected between the subjects
who were frankly hypoxemic (Pao2, ⱕ 60) and/or
hypercapnic (Paco2, ⱖ 45) and those subjects who
were less affected. The hypoxemic/hypercapnic
group also may be expected to have different responses to the treatment with supplemental oxygen.
Unfortunately, the number of subjects who fell into
the severely affected group was too small to allow for
a clear evaluation of these differences. The collection
of more subjects in both the hypercapnic and hypoxemic groups should be undertaken to look at these
issues.
Since the subject’s respiratory rate, inspiratory
time, and expiratory time were controlled for in this
experiment, we would not expect significant changes
in the work of breathing or in arterial CO2 levels
after SuppO2 administration. For this reason, we
would not expect this to cause stimulation of the
chemoreceptors, which would affect autonomic
modulation.
Pulmonary hypertension has been shown to be, at
least in part, responsible for the attenuation of
baroreflex responses in COPD patients.4 It is for this
reason that we speculate that the normalization of
BRS in these individuals is partially due to a reduction in pulmonary arterial tension following breathing with SuppO2. The improved central venous
SatO2 level during SuppO2 administration might
decrease the level of pulmonary arterial vasoconstriction and right ventricular wall stress, in turn causing
a decrease in sympathetic tone. This might happen
via the nitric oxide-mediated pathway, which causes
a direct relaxation of the pulmonary artery smooth
muscle in the presence of increased oxygen tension.
Since chronic hypoxemia and its associated pulmonary hypertension have been shown to be major
causes of morbidity and mortality,23 these findings of
improved autonomic modulation have particular
clinical relevance. The long-term oxygen treatment
trial and nocturnal oxygen treatment trial demonstrated the long-term benefits of SuppO2 in the
reduction of mortality in individuals with severe
hypoxemia. Although the mechanisms of the reduction of mortality are not totally clear, these results
established supplemental oxygen as the standard of
care for these individuals.24,25 While these studies
failed to demonstrate a reduction in pulmonary
hypertension and cor pulmonale in patients with
chronic hypoxemia, there appeared to be a reduction
of the progression of pulmonary hypertension. This
may have occurred via a reduction of hypoxic vasoconstriction and a reduction in basal vascular tone
with the prevention of vascular smooth muscle hypertrophy.26,27 Our findings of increased vagal modulation may be an indication of a possible mechanism
to describe the clinical finding mentioned above.
In conclusion, this study demonstrates that
SuppO2 in COPD patients significantly and favorably alters autonomic modulation. The exact mechanism by which this happens is not clear; however,
future research involving measurements of pulmonary arterial pressures, nitric oxide levels, and endtidal CO2 levels might hold the answer. These
findings may be of special value in further supporting
evidence of a cardioprotective role of SuppO2 in
patients with COPD.
ACKNOWLEDGMENT: We thank the VIDDA foundation for
its continued support.
References
1 Saito M, Mano T, Abe H, et al. Responses in muscle
sympathetic activity to acute hypoxia in humans. J Appl
Physiol 1986; 55:493– 498
2 Somers VK, Mark AL, Abboud FM. Potentiation of sympathetic nerve responses to hypoxia in borderline hypertensive
subjects. Hypertension 1988; 11:608 – 612
3 Somers VK, Mark AL, Abboud FM. sympathetic activation by
hypoxia and hypercapnia: implications for sleep apnea. Clin
Exp Hypertens 1988; A10(suppl):413– 422
4 Bhattachaya J, Cunningham DJC, Howson MG, et al. The
effects of mild hypoxia with constant PACO2 and ventilation
on the setting and sensitivity of the baroreceptor-cardiacdepressor reflex in man. J Physiol 1973; 234(suppl):112p–
114p
5 Bigger JT Jr, Kleiger RE, Fleiss JL, et al. Components of
heart rate variability measured during healing of acute myocardial infarction. Am J Cardiol 1988; 61:208 –215
6 Billman GE, Hoskins RS. Time-series analysis of heart rate
variability during submaximal exercise: evidence for reduced
cardiac vagal tone in animals susceptible to ventricular fibrillation. Circulation 1989; 80:146 –157
7 Bilmann GE, Schwartz PJ, Stone HL. Baro-receptor reflex
control of heart rate: a predictor of sudden cardiac death.
Circulation 1982; 66:874 – 880
8 Volterrani M, Scalvini S, Mazzuero G, et al. Decreased heart
rate variability in patients with chronic obstructive pulmonary
disease. Chest 1994; 106:1432–1437
9 Task Force of the European Society of Cardiology and North
American Society of Pacing and Electrophysiology. Heart rate
variability: standards of measurement, physiological interpretation, and clinical use. Circulation 1996; 93:1043–1065
CHEST / 118 / 3 / SEPTEMBER, 2000
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21952/ on 05/15/2017
695
10 Kitney RI. Techniques for studying short-term changes in
cardiorespiratory data. In: di Rienzo M, ed. Computer analysis of cardiovascular signals. Washington, DC: IOS Press,
1995; 41–52
11 Novak P. Time/frequency mapping of the heart rate, blood
pressure and respiratory signals. Med Biol Eng Comput 1993;
31:103–110
12 Novak V. Slow cardiovascular rhythms at tilt and syncope.
J Clin Neurophysiol 1995; 12:64 –71
13 Blaber AP, Yamamoto Y, Hughson RL. Methodology of
spontaneous baroreflex relationship assessed by surrogate
data analysis. Am J Physiol 1995; 268 (suppl):H1682–H1687
14 De Meersman R, Reisman S, Daum R, et al. Influence of
respiration on metabolic, hemodynanic, psychometric, and R-R
interval spectra. Am J Physiol 1995; 269(suppl):H1437–H1440
15 Bertinieri G, Dirienzo M, Cavallazzi A, et al. Evaluation of
baroreceptor reflex by blood pressure monitoring in unanesthetized cats. Am J Physiol 1988; 254(suppl):H377–H383
16 Murphy CA, Sloan, RP, Myers MM. Pharmacologic responses and spectral analyses of spontaneous fluctuations in
heart rate and blood pressure in SHR rats. J Auton Nerv Syst
1991; 36:237–250
17 Smuylan H, Safar ME. Systolic blood pressure revisited. J Am
Coll Cardiol 1997; 29:1407–1413
18 Madhavan S, Ooi WL, Cohen H, et al. Relation of pulse
pressure and blood pressure reduction to the incidence of
myocardial infarction. Hypertension 1994; 23:395– 401
19 SaCunha R, Pannier B, Benetos A, et al. Association between
high heart rate and high arterial rigidity in normotensive and
hypertensive subjects. J Hypertens 1997; 15(suppl):1423–1430
20 La Rovere MT, Specchia G, Mortara A, et al. Baroreflex
sensitivity, clinical correlates, and cardiovascular mortality
among patients with a first myocardial infarction: a prospective study. Circulation 1988; 78:816 – 824
21 La Rovere MT, Bigger JT Jr, Marcus FI, et al. Baroreflex
sensitivity and heart-rate variability in prediction of total
cardiac mortality after myocardial infarction: ATRAMI (Autonomic Tone and Reflexes After Myocardial Infarction)
Investigators. Lancet 1998; 351:478 – 484
22 Pagani M, Somers V, Furlan R, et al. Changes in autonomic
regulation induced by physical training in mild hypertension.
Hypertension 1988; 12:600 – 610
23 Weitzenblum E, Mammosser M, Ehrhart M. Evolution and
prognosis of pulmonary hypertension in chronic obstructive
pulmonary diseases. Herz 1986; 11:147–154
24 Medical Research Council Working Party. Long term domiciliary oxygen therapy in chronic hypoxic cor pulmonale
complicating chronic bronchitis and emphysema. Lancet
1981; 1:681– 685
25 Nocturnal Oxygen Therapy Trial Group. Continuous or nocturnal oxygen therapy in hypoxemic chronic obstructive airways disease: a clinical trial. Ann Intern Med 1980; 93:391–
398
26 Fletcher EC, Luckett RA, Goodnight-White S, et al. A
double-blind trial of nocturnal supplemental oxygen for sleep
desaturation in patients with chronic obstructive pulomonary
disease and a daytime PaO2 above 60 mm Hg. Am Rev Respir
Dis 1992; 145:1070 –1076
27 Stewart AG, Howard P. Indications for long term oxygen
therapy. Respiration 1992; 59(suppl):8 –13
696
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21952/ on 05/15/2017
Clinical Investigations