Download Effect of Hyperoxia on Left Ventricular Function and Filling

Effect of Hyperoxia on Left Ventricular
Function and Filling Pressures in
Patients With and Without Congestive
Heart Failure*
Susanna Mak, MD; Eduardo R. Azevedo, MD; Peter P. Liu, MD; and
Gary E. Newton, MD
Study objectives: To determine the effects of hyperoxia on left ventricular (LV) function in
humans with and without congestive heart failure (CHF).
Design: An acute physiologic study of the effect of hyperoxia on right-heart hemodynamics, LV
contractility (peak positive rate of rise of LV pressure [ⴙdP/dt]), time constant of isovolumic left
ventricular relaxation (␶), and LV filling pressures.
Setting: Bayer Cardiovascular Clinical Research Laboratory at the Mount Sinai Hospital,
Toronto, Ontario.
Patients: Sixteen patients with stable CHF and 12 subjects with normal LV function received the
hyperoxia intervention.
Interventions: Patients received 21% O2 by a nonrebreather mask, followed by 100% O2 for 20
min, and 21% O2 for a 10-min recovery period.
Results: In response to hyperoxia, there was a 22 ⴞ 6% increase in LV end-diastolic pressure
(LVEDP) in the CHF group and a similar 29 ⴞ 14% increase in LVEDP in the normal LV function
group (p < 0.05 for both; mean ⴞ SEM). Hyperoxia was also associated with a prolongation in ␶
of 10 ⴞ 2% in the CHF group (p < 0.05) and 8 ⴞ 2% in the normal LV function group (p < 0.05).
No changes in ⴙdP/dt were observed in either group.
Conclusions: Hyperoxia was associated with impairment of cardiac relaxation and increased LV
filling pressures in patients with and without CHF. These observations indicate that caution
should be used in the administration of high inspired O2 fractions to normoxic patients, especially
in the setting of CHF.
(CHEST 2001; 120:467– 473)
Key words: congestive heart failure; hyperoxia; ventricular function
Abbreviations: CHF ⫽ congestive heart failure; ⫹dP/dt ⫽ peak positive rate of rise of left ventricular pressure;
LV ⫽ left ventricular; LVEDP ⫽ left ventricular end-diastolic pressure; NO ⫽ nitric oxide; ROS ⫽ reactive oxygen
species; SVR ⫽ systemic vascular resistance; ␶ ⫽ time constant of isovolumic left ventricular relaxation; ␶l ⫽ ␶,
logarithmic method; ␶1/2 ⫽ ␶, pressure half-time method
administration of high-concentration suppleT hemental
oxygen to the point of hyperoxia can be a
frequent occurrence in the hospital setting. It has
been long appreciated that hyperoxia has adverse
*From the Bayer Cardiovascular Clinical Research Laboratory
(Drs. Mak, Azevedo, and Newton), Department of Medicine,
Mount Sinai Hospital, and The Toronto Hospital (Dr. Liu),
University of Toronto, Toronto, Ontario, Canada.
Drs. Mak and Azevedo are Research Fellows of the Heart and
Stroke Foundation, Dr. Liu holds the Heart and Stroke Polo
Chair in Cardiovascular Research, and Dr. Newton is a Research
Scholar of the Heart and Stroke Foundation of Canada.
This work was supported by a Grant in Aid from the Heart and
Stroke Foundation of Ontario (grant No. NA-3469), and by Bayer
Canada, Inc.
Correspondence to: Gary E. Newton, MD, Mount Sinai Hospital,
600 University Ave, Room 1614, Toronto, Ontario, M5G 1X5,
Canada; e-mail: [email protected]
consequences in patients with chronic obstructive
lung disease or acute respiratory failure, as gas
exchange may be worsened by denitrogenation atelectasis and increased intrapulmonary shunting.1,2
Over the last 50 years, it has also been demonstrated3,4 that hyperoxia results in hemodynamic alterations, including increased systemic vascular resistance (SVR) and reduced cardiac output. These
effects are of specific relevance to patients with
congestive heart failure (CHF), a condition with
increasing prevalence and representation in the hospital and intensive care setting. Haque et al5 observed that hyperoxia was also associated with a
reduction in stroke volume and an increase in pulmonary capillary wedge pressure in patients with
CHF. This fall in stroke volume raised the possibility
CHEST / 120 / 2 / AUGUST, 2001
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21965/ on 05/02/2017
467
that high-concentration O2 impaired cardiac function. Importantly, Haque et al5 also demonstrated
that the hemodynamic effects of hyperoxia occurred
in a dose-dependent fashion and were detectable at
inspired O2 concentrations as low as 40%. These
observations highlight the clinical relevance of examining the cardiovascular effects of hyperoxia, especially in the setting of CHF.
This acute hemodynamic investigation explored
the effect of high-concentration O2 on directly measured left ventricular (LV) hemodynamics and isovolumic indexes of LV contractility and relaxation in
patients with CHF and subjects with normal LV
function. We tested the hypothesis that hyperoxia
would impair LV hemodynamics and function in
patients with CHF but not in subjects with normal
LV function.
Materials and Methods
Study Protocol
Following catheter placement, a nonrebreather mask was
applied and 21% O2 was administered for 10 min prior to initial
measurements (control). The reservoir bag was emptied of
residual air, and 100% O2 was then administered by the same
nonrebreather mask for 20 min, at which time repeat measurements were performed (hyperoxia). A final set of measurements
was recorded following a 10-min period during which 21% O2
was again administered (recovery). Subjects in the normoxia
control group received 21% O2 by face mask instead of 100% O2.
Measurements were made at the same time intervals as described
above.
Hemodynamic Measurements
Measurements of heart rate, systemic BP, right atrial pressure,
and pulmonary artery pressure were acquired on a multichannel
chart recorder. At each condition, at least 15 cardiac cycles were
averaged for the final value. Cardiac output was measured by the
thermodilution technique in triplicate. Coronary sinus blood flow
measurements were performed in triplicate according to the
method of Ganz et al.6
Study Population
LV Contractility and Relaxation
Thirty-five patients, separated into three groups, participated
in this study. The CHF group (n ⫽ 16; mean age, 62 ⫾ 2 years;
15 men and 1 woman) all had stable symptoms (New York Heart
Association class II and III) and a LV ejection fraction by
radionuclide ventriculography of ⬍ 35% (mean value, 25 ⫾ 2%).
This group included 5 patients with idiopathic dilated cardiomyopathy and 11 patients with coronary artery disease. Medical
therapy in this group included diuretics (n ⫽ 14), digoxin (n ⫽ 7),
angiotensin-converting enzyme inhibitors or angiotensin II-receptor blockers (n ⫽ 14), ␤-blockers (n ⫽ 6), calcium channel
blockers (n ⫽ 4), and nitrates (n ⫽ 5). The normal LV function
group (n ⫽ 12; mean age, 63 ⫾ 2 years; 10 men and 2 women)
included subjects with stable coronary artery disease. Normal LV
systolic function was documented by two-dimensional echocardiography or quantitative contrast ventriculography. Medical
therapy in this group included angiotensin-converting enzyme
inhibitors (n ⫽ 2), ␤-blockers (n ⫽ 10), calcium channel blockers
(n ⫽ 6), and nitrates (n ⫽ 7). Finally, the normoxia control group
(n ⫽ 7; mean age, 61 ⫾ 3 years; five men and two women) also
included subjects with normal LV function and stable coronary
artery disease. Medical therapy in this group included ␤-blockers
(n ⫽ 7), calcium channel blockers (n ⫽ 2), and nitrates (n ⫽ 3).
This protocol was approved by the University of Toronto Ethics
Review Committee for experimentation involving human subjects. All patients gave written informed consent.
LV pressure and its first derivative (rate of rise of LV pressure)
[continuous electronic differentiation] were digitally recorded
with a sampling rate of 300 Hz. The measure of contractility used
in this study was the peak positive rate of rise of LV pressure
(⫹dP/dt). The measure of relaxation used was the time constant
of isovolumic LV relaxation (␶) (logarithmic method [␶l] and
pressure half-time method [␶1/2]), which was calculated by two
independent methods previously described.7,8 ⫹dP/dt, ␶, and LV
end-diastolic pressure (LVEDP) were calculated off-line using a
customized software program (Labview version 3.0; National
Instruments Corporation; Austin, TX) by a technician blinded to
other data in the investigation. These methods are established
and validated in our laboratory.9
Blood Samples
Blood from the femoral artery was analyzed for pH, Pco2, and
Po2 by an automated pH/blood gas analyzer (model 865; Ciba
Corning Diagnostics; Medfield, MA). Oxygen saturation was
determined with a whole-blood oximeter (Oxicom 3000; Waters
Instruments; Rochester, MN). Oxygen content was calculated as
follows: O2 saturation ⫻ hemoglobin ⫻ 1.34 ⫹ 0.0031 ⫻ Po2.
Arterial and coronary sinus lactate levels were measured at
control and during hyperoxia.
Cardiac Catheterization Procedure
Statistical Analysis
Patients were studied in a nonsedated fasting state following a
diagnostic heart catheterization from the femoral approach.
Treatment with all oral medications was withheld on the morning
of the investigation. For the CHF and normal LV function
groups, the following catheters were inserted under fluoroscopic
guidance: (1) a pulmonary artery catheter, (2) a 7F micromanometer-tipped catheter (Millar Instruments; Houston, TX) in the LV
(11 patients in the normal LV function group and 12 patients in
the CHF group), and (3) a 7F coronary sinus thermodilution flow
catheter (Webster Laboratories; Baldwin City, CA) from an
antecubital vein. For the normoxia control group, a 7F micromanometer-tipped catheter was inserted into the LV.
All data are presented as mean ⫾ SEM. Analysis was performed with a statistical software package (SigmaStat version 1.0;
Jandel Scientific Software; San Rafael, CA). Between-group
comparisons of baseline characteristics were made using a Student’s t test. Within-group comparisons of the responses to
hyperoxia (control, hyperoxia, and recovery) were made with a
one-way repeated-measures analysis of variance. The StudentNewman-Keuls test was used post hoc to identify significant
pairwise differences. The relationship between ␶ and LV pressure
responses to the hyperoxia stimulus was determined using simple
linear regression and a Pearson correlation statistic. A p value
⬍ 0.05 was required for statistical significance.
468
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21965/ on 05/02/2017
Clinical Investigations
Baseline Characteristics
CHF group; p nonsignificant for both), indicating
that myocardial ischemia did not occur despite the
significant fall in coronary sinus blood flow.
The CHF group had significantly elevated heart
rate and filling pressures, as well as impaired LV
contractility and relaxation, compared to the normal
LV function group. Baseline arterial O2 saturation
was similar in the two groups (Table 1).
The Effect of Hyperoxia on LV Hemodynamics,
Contractility, and Relaxation
Results
The Effect of Hyperoxia on Hemodynamics, Blood
Gases, and Lactate Levels
In the CHF group, the hyperoxia stimulus was
associated with several hemodynamic changes, including an increase in SVR, and reductions in cardiac
output, stroke volume, and coronary sinus blood flow
(p ⬍ 0.05 for all; Table 2). In the normal LV function
group, hyperoxia was associated with an increase in
SVR and a reduction in coronary sinus blood flow
(p ⬍ 0.05 for both; Table 3). There were no significant changes in cardiac output and stroke volume in
the normal LV function group.
The hyperoxia stimulus resulted in the expected
increase in the partial pressure of O2 (78 ⫾ 3 to
358 ⫾ 21 mm Hg in the normal LV function group
and 78 ⫾ 4 to 312 ⫾ 17 mm Hg in the CHF group;
p ⬍ 0.05 for both). There was no increase in the
transcardiac lactate gradient during the administration of high inspired O2 fraction (⫺ 0.5 ⫾ 0.1 to
⫺ 0.4 ⫾ 0.1 mmol/L in the normal LV function
group and ⫺ 0.2 ⫾ 0.1 to ⫺ 0.5 ⫾ 0.1 mmol/L in the
Table 1—Baseline Characteristics*
Characteristics
Normal LV Function
Group (n ⫽ 12)
CHF Group
(n ⫽ 16)
Heart rate, beats/min
FAsys, mm Hg
FAdiast, mm Hg
RAP, mm Hg
PAsys, mm Hg
PAdiast, mm Hg
PVR, dyne 䡠 s 䡠 cm⫺5
Cardiac output, L/min
Stroke volume, mL/beat
LVEDP, mm Hg
LVSP, mm Hg
⫹dP/dt, mm Hg/s
␶l, ms
␶1/2, ms
Arterial O2 saturation, %
62 ⫾ 3
146 ⫾ 5
66 ⫾ 3
10 ⫾ 1
30 ⫾ 1
12 ⫾ 1
153 ⫾ 33
4.3 ⫾ 0.3
73 ⫾ 6
11 ⫾ 2
134 ⫾ 5
1,291 ⫾ 40
47 ⫾ 2
37 ⫾ 2
96.5 ⫾ 0.4
73 ⫾ 3†
139 ⫾ 7
69 ⫾ 3
8⫾1
40 ⫾ 3†
18 ⫾ 2†
185 ⫾ 24
4.6 ⫾ 0.4
64 ⫾ 6
21 ⫾ 3†
123 ⫾ 6
1,050 ⫾ 84†
60 ⫾ 3†
46 ⫾ 1†
95.6 ⫾ 0.5
*Data are presented as mean ⫾ SEM. FAsys ⫽ femoral artery systolic pressure; FAdiast ⫽ femoral artery diastolic pressure;
RAP ⫽ right atrial pressure; PAsys ⫽ pulmonary artery systolic
pressure; PAdiast ⫽ pulmonary artery diastolic pressure;
PVR ⫽ pulmonary vascular resistance; LVSP ⫽ left ventricular systolic pressure.
†p value ⬍ 0.05 vs normal LV function group.
There were no significant changes in LV contractility in response to high inspired O2 fraction in
either group, as measured by ⫹dP/dt (Tables 2, 3).
Left ventricular systolic pressure increased significantly in the normal LV function group in response
to the hyperoxia stimulus. In the CHF group, there
was a similar although nonsignificant trend for an
increase in LV systolic pressure.
The 20-min hyperoxia stimulus was associated
with a significant increase in LV filling pressures and
impairment in LV isovolumic relaxation in both the
CHF and the normal LV function groups. In response to hyperoxia, there was a 22 ⫾ 6% increase in
LVEDP in the CHF group and a 29 ⫾ 14% increase
in the normal LV function group (p ⬍ 0.05 for both;
Tables 2, 3 and Fig 1). Hyperoxia was also associated
with prolongation in the time constant of LV ␶ (both
␶l and ␶ 1/2) in the CHF group and in the normal LV
function group (Tables 2, 3 and Fig 1). Neither the
rise in LVEDP nor the prolongation in isovolumic
LV relaxation returned to control values during the
recovery period in both groups. There was a relationship between the prolongation of ␶ and the rise
in LVEDP that was highly significant in the CHF
group (r ⫽ 0.76; p ⬍ 0.01) and in the entire population combined (r ⫽ 0.75; p ⬍ 0.01; Fig 2).
␶ has been reported to be dependent on changes
in afterload, especially in the setting of CHF.10,11
However, there was no relationship between the
prolongation in ␶ (␶l) and the change in LV systolic
pressure in the CHF group (r ⫽ ⫺ 0.13; p ⫽ 0.69)
or in the normal LV function group (r ⫽ ⫺ 0.08;
p ⫽ 0.82). Similarly, there was no relationship between the change in ␶1/2 and changes in LV systolic
pressure in either group. No relationships between ␶
and any measure of afterload, including BP and SVR,
were observed in either group.
Normoxia Control Group
There were no significant changes in heart rate,
BP, ⫹dP/dt, LVEDP, ␶l, or ␶1/2 in the normoxia
control group (data not shown).
Discussion
Our study confirmed that hyperoxia causes systemic vasoconstriction and reduces cardiac output.
We have expanded these findings by demonstrating
CHEST / 120 / 2 / AUGUST, 2001
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21965/ on 05/02/2017
469
Table 2—Effect of Hyperoxia on Hemodynamics and Cardiac Function in Patients With CHF*
Variables
Control
Hyperoxia
Recovery
Heart rate, beats/min
FAsys, mm Hg
FAdiast, mm Hg
RAP, mm Hg
PAsys, mm Hg
PAdiast, mm Hg
PVR, dyne 䡠 s 䡠 cm⫺5
Cardiac output, L/min
SV, mL/beat
SVR, dyne 䡠 s 䡠 cm⫺5
CSBF, mL/min
LVEDP, mm Hg
LVSP, mm Hg
⫹dP/dt, mm Hg/s
␶l, ms
␶1/2, ms
73 ⫾ 3
139 ⫾ 7
69 ⫾ 3
8⫾1
40 ⫾ 3
18 ⫾ 2
185 ⫾ 24
4.6 ⫾ 0.4
64 ⫾ 6
1,626 ⫾ 148
94 ⫾ 6
21 ⫾ 3
123 ⫾ 6
1,050 ⫾ 84
60 ⫾ 3
46 ⫾ 1
72 ⫾ 3
140 ⫾ 7
72 ⫾ 2
9⫾1
40 ⫾ 4
19 ⫾ 2
214 ⫾ 45
4.1 ⫾ 0.3†
57 ⫾ 5†
1,901 ⫾ 181†
72 ⫾ 8†
25 ⫾ 3†
128 ⫾ 6
1,020 ⫾ 79
67 ⫾ 4†
49 ⫾ 2†
74 ⫾ 4
138 ⫾ 9
70 ⫾ 3
9⫾1
41 ⫾ 4
20 ⫾ 2
239 ⫾ 50
4.3 ⫾ 0.3
58 ⫾ 5†
1,715 ⫾ 177
84 ⫾ 6
25 ⫾ 3†
128 ⫾ 8
1,021 ⫾ 80
68 ⫾ 3†
49 ⫾ 2†
*Data are presented as mean ⫾ SEM. CSBF ⫽ coronary sinus blood flow; see Table 1 for definitions of abbreviations.
†p value ⬍ 0.05 vs control.
that hyperoxia is also associated with prolongation in
the time constant of isovolumic LV relaxation, ␶, and
increases in LVEDP. Therefore, our observations
suggest that hyperoxia is associated with disturbances of both early and late phases of LV filling.
Furthermore, these effects were observed not only
in patients with CHF, but also in subjects with
normal LV function and coronary artery disease.
The increase in LVEDP suggests impairment of
one or more aspects of ventricular filling. The determinants of LV filling are numerous and include
ventricular relaxation, diastolic suction, filling volume, distensibility of the ventricular chamber, pericardial restraint, ventricular interaction, and atrial
contribution.12 The observed relationship between
the prolongation of ␶ and the increase in LVEDP
suggests the possibility that impaired relaxation contributed to abnormal LV filling. A potential mechanism by which relaxation was impaired may relate to
the generation of reactive oxygen species (ROS).
Increased concentrations of ROS, as likely occurred
in response to hyperoxia,13,14 have negative effects
on myocardial function,15–18 and can impair myocyte
calcium homeostasis,19,20 as well as cardiac ␤-receptor signaling,21 both of which may prolong ventricular relaxation. By a similar free-radical mechanism,
LV distensibility may have been impaired as a result
of consumption of endothelial-derived nitric oxide
(NO) by ROS in response to hyperoxia. In vitro
studies22 have confirmed that NO and hyperoxia-
Table 3—Effect of Hyperoxia on Hemodynamics and Cardiac Function in Subjects With Normal LV Function*
Variables
Heart rate, beats/min
FAsys, mm Hg
FAdiast, mm Hg
RAP, mm Hg
PAsys, mm Hg
PAdiast, mm Hg
PVR, dyne 䡠 s 䡠 cm⫺5
Cardiac output, L/min
SV, mL/beat
SVR, dyne 䡠 s 䡠 cm⫺5
CSBF, mL/min
LVEDP, mm Hg
LVSP, mm Hg
⫹dP/dt, mm Hg/s
␶l, ms
␶ ⁄ , ms
12
Control
Hyperoxia
Recovery
62 ⫾ 3
146 ⫾ 5
66 ⫾ 3
10 ⫾ 1
30 ⫾ 1
12 ⫾ 1
153 ⫾ 33
4.3 ⫾ 0.3
73 ⫾ 6
1,535 ⫾ 121
89 ⫾ 11
11 ⫾ 2
134 ⫾ 5
1,291 ⫾ 40
47 ⫾ 2
37 ⫾ 2
60 ⫾ 3
148 ⫾ 5
68 ⫾ 2
10 ⫾ 1
26 ⫾ 2†
11 ⫾ 1
135 ⫾ 27
4.2 ⫾ 0.3
72 ⫾ 6
1,663 ⫾ 144†
82 ⫾ 11†
13 ⫾ 2†
142 ⫾ 6†
1,281 ⫾ 43
50 ⫾ 3†
39 ⫾ 2†
63 ⫾ 3
146 ⫾ 5
68 ⫾ 3
10 ⫾ 1
27 ⫾ 2†
12 ⫾ 1
150 ⫾ 29
4.3 ⫾ 0.3
72 ⫾ 7
1,594 ⫾ 125
85 ⫾ 10
13 ⫾ 2†
142 ⫾ 7†
1,300 ⫾ 44
51 ⫾ 3†
40 ⫾ 2†
*Data are presented as mean ⫾ SEM. See Tables 1, 2 for definitions of abbreviations.
†p value ⬍ 0.05 vs control.
470
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21965/ on 05/02/2017
Clinical Investigations
Figure 1. LVEDP at control and during
the hyperoxia stimulus (top, A), and ␶,
the time constant of isovolumic LV relaxation (TL), at control and during the
hyperoxia stimulus (bottom, B) in subjects with normal LV function (solid circles) and patients with CHF (open
squares). *p ⬍ 0.05 vs control.
generated ROS, specifically the superoxide anion,
participate in a reaction that results in the rapid
degradation of NO. Previous investigators have demonstrated that NO enhances diastolic LV distensibility. These studies, in humans with normal ventricular
function,23 CHF,24 and pressure overload25 demonstrated that the local infusion of NO donors resulted
in a fall in LVEDP as well as a shift of the diastolic
pressure-volume curve to the right, confirming improved LV distensibility. The persistence of impaired diastolic performance following removal of
100% O2 in the present study is also consistent with
animal experiments demonstrating that ROS-mediated impairment in cardiac function persists much
longer than the duration of the initial free-radical
insult.17
Our observations did not suggest that the increase
in LVEDP resulted from increased LV diastolic
volume as might have occurred if pulmonary blood
flow had increased. Although ventricular volumes
were not measured in this study, indirect evidence
indicated that pulmonary blood flow did not increase, especially in patients with CHF. In this study
and that of Haque et al,5 cardiac output fell while
pulmonary vascular tone did not change or tended to
increase. Similarly, there was no indication that LV
diastolic volume was altered as a consequence of
diastolic ventricular interaction and pericardial constraint,26 since right atrial pressure, which is a reasonable surrogate for pericardial pressure,27 did not
change in response to high-concentration O2. Both
impaired distensibility and impaired relaxation may
have been a consequence of myocardial ischemia
induced by the observed fall in coronary blood flow
CHEST / 120 / 2 / AUGUST, 2001
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21965/ on 05/02/2017
471
Figure 2. Change in LVEDP (⌬LVEDP, hyperoxia minus
control) vs the change in ␶, the time constant of isovolumic LV
relaxation (⌬␶l, hyperoxia minus control) in subjects with normal
LV function (solid circles) and patients with CHF (open circles).
r ⫽ 0.752, p ⬍ 0.01.
in response to hyperoxia. However, this is unlikely as
there was no decrease in transcardiac lactate extraction, a sensitive index of myocardial ischemia.
The lack of a fall in ⫹dP/dt in response to
high-concentration O2 did not confirm our hypothesis that hyperoxia impairs LV contractile function.
Of note, any hyperoxia-mediated reduction in
⫹dP/dt may have been masked by the concurrent
increases in LVEDP that would have tended to
augment contractility.28 Another interesting possibility is that hyperoxia may have stimulated cardiac
myocyte production of angiotensin I, which is converted to angiotensin II in the blood vessels. Angiotensin II is a potent vasoconstrictor, and also stimulates endothelial cells to secrete endothelin, a major
positive cardiac inotropic substance.29 This mechanism has been demonstrated30 in isolated cardiac
myocytes and could account for both the fall in
coronary blood flow, and the preservation in contractility.
The hyperoxia-associated disturbance in LV filling
and LVEDP observed in our study may have important detrimental clinical sequela. Supplemental O2 is
widely used in the treatment of acute and subacute
cardiac conditions, such as acute pulmonary edema
and unstable coronary syndromes. In an effort to
ensure adequate systemic oxygenation,31 the availability of high-flow enriched O2 mixtures likely leads
to inadvertent hyperoxia, especially in patients without significant pulmonary disease. In patients with
CHF, hyperoxia-associated increase in left-heart filling pressures is of specific concern. The contribution
of increased LV filling pressures to pulmonary venous hypertension and pulmonary congestion are
well known. Although this study only included CHF
patients with stable symptoms based on safety considerations, if a similar rise in LVEDP occurs in
response to hyperoxia in decompensated patients,
this could exacerbate pulmonary congestion. Admittedly, the degree of hyperoxia that occurs in clinical
practice is likely less than that observed in this study
since supplemental O2 is usually administered at
concentrations ⬍ 100%. However, Haque et al5
demonstrated in CHF patients that inspired O2
concentration ⬎ 21% elicited adverse hemodynamic
changes in a dose-dependent manner. Similarly, the
negative effects of hyperoxia on relaxation and
LVEDP may also occur at lower O2 concentrations.
There were limitations in this study with respect to
the assessment of ventricular relaxation. The prolongation in the time constant of isovolumic LV relaxation, ␶, in response to hyperoxia occurred in both
study groups and was significant by two independent
methods. However, the interpretation of this observation is confounded by the increase in LV systolic
pressure that also occurred in response to hyperoxia.
Increases in afterload are associated with slowing of
isovolumic LV relaxation,10,11 although observations
from the present study suggest that the prolongation
in ␶ was not solely due to changes in LV afterload.
There were no relationships between the increases
in LV systolic pressure, or any other measure of
afterload, and the prolongation in ␶. ␶ was also
prolonged in the normal LV function group, despite
the relative insensitivity of ␶ to changes in afterload
in this population.32,33 Indeed, Starling et al32 administered methoxamine to normal subjects achieving
large increases in LV afterload with no changes in ␶.
Despite the lack of a “gold standard” with which to
assess isovolumic relaxation, the marked increase in
LV filling pressures in response to hyperoxia confirms significant impairment of diastolic events.
In summary, the administration of high-concentration O2 was associated with a prolongation of isovolumic LV relaxation and an increase in LV filling
pressures both in patients with CHF and in subjects
with normal LV function. These adverse effects of
hyperoxia on hemodynamics and cardiac function
indicate that caution should be used in the administration of high inspired O2 fractions to normoxic
patients, especially in the setting of CHF.
ACKNOWLEDGMENT: The authors thank the staff of the
Bayer Cardiovascular Clinical Research Laboratory of the Mount
Sinai Hospital, as well as Drs. John Floras and John Parker for
review of the article.
472
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21965/ on 05/02/2017
Clinical Investigations
References
1 Dantzker DR WP, West JB. Instability of lung units with low
VA/Q ratios during O2 breathing. J Appl Physiol 1975;
38:886 – 895
2 Santos C, Ferrer M, Roca J, et al. Pulmonary gas exchange
response to oxygen breathing in acute lung injury. Am J
Respir Crit Care Med 2000; 161:26 –31
3 Whitehorn WV, Edelmann A, Hitchcock FA. Cardiovascular
responses to breathing of 100% oxygen at normal barometric
pressure. Am J Physiol 1946; 146:61– 65
4 Daly WJ, Bondurat S. Effects of oxygen breathing on the
heart rate, blood pressure and cardiac index of normal men resting, with reactive hyperemia and after atropine. J Clin
Invest 1962; 41:126 –132
5 Haque WA, Boehmer J, Clemson BS, et al. Hemodynamic
effects of supplemental oxygen administration in congestive
heart failure. J Am Coll Cardiol 1997; 27:353–357
6 Ganz W, Tamura K, Marcus HS, et al. Measurement of
coronary sinus blood flow by continuous thermodilution in
man. Circulation 1971; 44:181–195
7 Weiss JL, Frederiksen JW, Weisfeldt JM. Hemodynamic
determinants of the time-course fall in canine left ventricular
pressure. J Clin Invest 1976; 58:751–760
8 Mirsky I. Assessment of diastolic function: suggested methods
and future considerations. Circulation 1984; 69:836 – 841
9 Newton GE, Parker AB, Landzberg JS, et al. Muscarinic
receptor modulation of basal and ␤-adrenergic stimulated
function of the failing human left ventricle. J Clin Invest
1996; 98:2756 –2763
10 Eichhorn EJ, Willard JE, Alvarez L, et al. Are contraction and
relaxation coupled in patients with and without congestive
heart failure? Circulation 1992; 85:2132–2139
11 Komamura K, Shannon RP, Pasipoularides A, et al. Alterations in left ventricular diastolic function in conscious dogs
with pacing-induced heart failure. J Clin Invest 1992; 89:
1825–1838
12 Nishimura RA, Tajik J. Evaluation of diastolic filling of left
ventricle in health and disease: Doppler echocardiography is
the clinician’s rosetta stone. J Am Coll Cardiol 1997; 30:8 –18
13 Fridovich I. The biology of oxygen radicals: the superoxide
radical is an agent of oxygen toxicity; superoxide dismutases
provide an important defense. Science 1978; 201:875– 880
14 Jamieson D, Chance B, Cadenas E, et al. The relation of free
radical production to hyperoxia. Ann Rev Physiol 1986;
48:703–719
15 Goldhaber JI, Ji S, Lamp ST, et al. Effects of exogenous free
radicals on electromechanical function and metabolism in
isolated rabbit and guinea pig ventricle. J Clin Invest 1989;
83:1800 –1809
16 Blaustein AS, Schine L, Brooks WW, et al. Influence of
exogenously generated oxidant species on myocardial function. Am J Physiol 1986; 250:H595–H599
17 Bolli R, Zughaib M, Li XY, et al. Recurrent ischemia in the
canine heart causes recurrent bursts of free radical production that have a cumulative effect on contractile function: a
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
pathophysiological basis for chronic myocardial “stunning.”
J Clin Invest 1995; 96:1066 –1084
Schrieir GM, Hess ML. Quantitative identification of superoxide anion as a negative inotropic species. Am J Physiol 1988;
255:H138 –H143
Kaneko M, Beamish RE, Dhalla NS. Depression of heart
sarcolemmal Ca2⫹-pump activity by oxygen free radicals.
Am J Physiol 1989; 256:H368 –H374
Morris TE, Sulakhe PV. Sarcoplasmic reticulum Ca(2⫹)
pump dysfunction in rat cardiomyocytes briefly exposed to
hydroxyl radicals. Free Rad Biol Med 1997; 22:37– 47
Persad S, Rupp H, Jindal R, et al. Modification of cardiacadrenoceptor mechanisms by H2O2. Am J Physiol 1998;
274:H416 –H423
Rubanyi GM, Vanhoutte PM. Superoxide anions and hyperoxia inactivate endothelium-derived relaxing factor. Am J
Physiol 1986; 250:H822–H827
Paulus WJ, Vantrimpont PJ, Shah AM. Acute effects of nitric
oxide on left ventricular relaxation and diastolic distensibility
in humans: assessment by bicoronary sodium nitroprusside
infusion. Circulation 1994; 89:2070 –2078
Heymes C, Vanderheyden M, Bronzwaer JG, et al. Endomyocardial nitric oxide synthase and left ventricular preload
reserve in dilated cardiomyopathy. Circulation 1999; 99:
3009 –3016
Matter CM, Mandinov L, Kaufmann PA, et al. Effect of NO
donors on LV diastolic function in patients with severe
pressure-overload hypertrophy. Circulation 1999; 99:2396 –
2401
Janicki JS. Influence of the pericardium and ventricular
interdependence on left ventricular diastolic and systolic
function in patients with heart failure. Circulation 1990;
81:III-15–III-20
Tyberg JV, Taichman GC, Smith ER, et al. The relationship
between pericardial pressure and right atrial pressure: an
intraoperative study. Circulation 1986; 73:428 – 432
Quinones MA, Gaasch WH, Alexander JK. Influence of acute
changes in preload, afterload, contractile state and heart rate
on ejection and isovolumic indices of myocardial contractility
in man. Circulation 1976; 53:293–302
Mebazaa A, Mayoux E, Maeda K, et al. Paracrine effects of
endocardial endothelial cells on myocyte contraction mediated via endothelin. Am J Physiol 1993; 265:H1841–H1846
Winegrad S, Henrion D, Rappaport L, et al. Self-protection
by cardiac myocytes against hypoxia and hyperoxia. Circ Res
1999; 85:690 – 698
Adult advanced cardiac life support: part III. JAMA 1992;
268:2199 –2235
Starling MR, Montgomery DG, Mancini J, et al. Load
independence of the rate of isovolumic relaxation in man.
Circulation 1987; 76:1274 –1281
Varma SK, Owen RM, Smucker ML, et al. Is tau a preloadindependent measure of isovolumetric relaxation? Circulation
1989; 80:1757–1765
CHEST / 120 / 2 / AUGUST, 2001
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21965/ on 05/02/2017
473