Evidence that Circulatory Oscillations Accompany Ventilatory Oscillations during Exercise in Patients with Heart Failure1- 3
ISSAHAR BEN-DOV, KATHY E. SIETSEMA, RICHARD CASABURI, and KARLMAN WASSERMAN
Introduction Periodic breathing (PB) resembling Cheyne-Stokes respiration (CSR) has been described in patients with congestive heart failure (CHF) at rest and during exercise (1-3). This has been attributed to unstable ventilatory control caused by slow circulation, leading to delayed transmission of humoral ventilatory stimuli to the chemoreceptors (4-7). However, blood pressure and cerebral perfusion also oscillate during CSR (1, 8), suggesting that circulatory control is also involved. Because O 2 uptake (V02) depends on pulmonary blood flow, insight into the role of the circulation in PB can be gained by studying V02, which is also oscillatory (9), breath by breath, in patients with CHF during exercise. Oscillations of V02 measured from respired gas could represent true fluctuations of O 2 across the alveolar-pulmonary capillary interface. Alternatively, these fluctuations might merely reflect cyclic changes in lung gas stores, because of changing end-expiratory lung volume (EELV) and O2 concentration, induced by ventilatory fluctuations. True transalveolar fluctuations in V02 cannot take place unless pulmonary blood flow and/or metabolic rate oscillate, since arterial blood is nearly fully saturated with O 2 at the Pao, usually present in these patients during exercise (2). The Fick relatjonship, V02 = Q(Cao2 - Cv02), where Q = cardiac output, C = content, a = arterial and v = mixed venous, dictates that true alveolar V02 fluctuates when pulmonary blood flow or CV02oscillates. We hypothesized that lung gas stores do not account for the prominent oscillations in V01 observed during exercise in patients with CHF, and that these oscillations are more likely attributed to oscillations of pulmonary blood flow. To test this hypothesis we analyzed the fluctuation and phase relationship of ventilation (VE) and V02 in 17 patients with CHF who showed marked spontaneous PB during exercise. To examine to what 776
SUMMARY Periodic breathing (PB) during exerclH In patients with congestive heart failure (CHF) hypothellzed that the V0 2 Is associated with prominent oscillations (OSC) of O2 Uptalc8 (V02)' OSC represent OSC In true O2 exchange, resulting from concomitant cardiac output fluctuations and are not merely due to OSC of lung O2 stores. compared the amplitude of the OSC of V0 2 , ventilation (VE), and end-explratory lung volume (EELV)In 17patients with CHF and PB and In seven healthy control subjects who volltlonally simulated PB. Sublects underwent an Incremental and/or a constant work-rate exerclH test. VE and v0 2 were measured breath by breath. EELVchange was estimated by summing the difference between Inspiratory and expiratory tidal volumes for each breath. The amplitude of the OSc, a, Is expre888d as the ratio of the difference between the peak (peak - nadlr)/mean]. In CHF,during and nadir of the oscillating variable divided by Its mean [a Incremental testing, the amplitude of the VE OSC was smaller than that of the v0 2 OSC (aVE = 49 ± 15% [SD], a v0 2 63 ± 25%, P < 0.01). In contrast, during volitional PB In the control subJects, VE OSC were larger than v0 2 OSC (aVE = 48 ± 12%, a v0 2 = 25 ± 11%, P < 0.01). This suggests that changing VE Itself cannot account for the marlc8d v0 2 OSC seen In CHF. In the patients, EELVshowed no systematic Osc, did not correlate with a v0 2, and wei not IlgnHlcantly different from zero. In contrast, In the control subjects, EELV tended to be lower at peak VE (-0.14 ± 0.5 L, P = 0.0&), and It was well correlated with a V0 2 (r = 0.6, P < 0.01). In conclusion, v0 2 oscillations In patients with CHF are more prominent than the associated VE OSC, are greater than the v0 2 OSC that can be Induced by volitional PB, and are not associated with changes In EELV. This supports the hypothesis that cardiac output OSC contribute to v0 2 OSC.
we
we
=
=
AM REV RESPIR DIS 1912; 145:778-781
extent these gas exchange oscillations could be induced by the breathing pattern itself, we compared the V02 fluctuations observed in the patients with CHF with the V02 oscillations induced in control subjects who volitionally oscillated their breathing pattern to simulate the natural PB under the same exercise conditions. We found that in the patients with CHF who exhibited PB, the amplitude of the V02 oscillations during exercise were more prominent relative to the concomitant VEoscillations and greater than the V02 oscillations induced by the simulated PB in the control group. Methods Protocols 1\vo protocols were employed: a progressive incremental exercisetest performed on either a treadmill or a cycle ergometer, and a constant work-rate (WR) low intensity test performed on a cycle ergometer. Fourteen patients with CHF who had oscillatory breathing during exercise and four control subjects who simulated oscillatory breathing pattern performed the incremental exercise.Five patients with CHF who had oscillatory breath-
ing during exercise and five control subjects who simulated oscillatory breathing pattern participated in the constant WR protocol.
Incremental Exercise Studies 'Ireadmill tests. All treadmill studiesweredone according to a modified Naughton protocol (10). After a short period of collecting resting data, the speed and/or the slope of the treadmill were increased every 2 min from 1 mile/h and zero degrees to a maximum of 3.5 miles/h and 16 degrees or to the limit of tolerance. Cycle ergometry. After 2 to 3 min of unloaded pedaling, the WR was increased ev(Received in original form April 15, 1991 and in revised form September 16, 1991) 1 From the Division of Respiratory and Critical Care, Physiology and Medicine, Harbor-UCLA Medical Center, Torrance, California. 2 Supported in part by Grant HL-l1907 from the National Heart, Lung, and Blood Institute and by the Research and Education Institute, HarborUCLA Medical Center. 3 Correspondence and requests for reprints should be addressed to Dr. Issahar Ben-Dov, Division of Respiratory and Critical Care, Physiology and Medicine,Harbor-UCLA Medical Center, 1000 W. Carson Street. Torrance, CA 90509.
PERIODIC BREATHING AND O. UPTAKE OSCILLATIONS IN CHF
ery 1 min in steps, or continuously in a ramp pattern, at a rate of 10 or 15 W/min. During the incremental tests subjects respired through a mouthpiece connected to a low resistance valve (dead space = 100 ml) with the nose occluded. Expiratory flow was recorded by a pneumotachograph and integrated to determine volume. Fractional concentrations of expired CO 2and O2weremeasured by infrared CO 2and fuel cellO 2analyzer. End-tidal pressure of CO 2(pETe02), VE, V02, and Veo2were measured breath by breath at the mouth using a computerized system (System2001or CPX; Medical Graphics Corp., St. Paul, MN). In addition, Veo2IVo2 (R), mean heart rate (HR), and O 2 pulse (Vo2/HR) were also calculated breath by breath.
Constant Work-Rate Exercise Studies The constant WR exercise tests were performed for 6 min on a cycle ergometer. The WR was 25 W for the patients and 50 to 70 W for the control subjects. In each case, the WR chosen was below the subject's anaerobic threshold (AT) as determined from a prior incremental study. Measurements of VE and gas exchange were made using a system designed in this institution (11). A mass spectrometer (Perkin-Elmer Medical Instruments, Oakbrook, IL) was used to measure respired O2, CO 2, and N 2 concentrations. A turbine flow-meter (SensorMedics, Anaheim, CA) was used to measure inspiratory and expiratory tidal volume (VT).
Voluntary Periodic Breathing during Exercise Healthy subjects were guided to volitionally simulate PB during exercise. This was done in order to contrast the gas exchange induced by pure ventilatory oscillations with that of spontaneous PB as found in patients with CHF. Subjects were instructed to oscillate their VT (respiratory rate was maintained constant) while exercising. The period of the voluntary oscillation was 1 min/cycle,approximating that observed in the patients. The amplitude of the VT oscillation was chosen so that the resulting voluntarily induced VE fluctuations would match the mean fluctuation of the patients' spontaneous PB, which was also primarily induced by VT changes. Subjects were assisted to achieve the desired ventilatory pattern by a metronome and a display of the actual and the target VT (Ventilation Monitor LS-80; Bear Medical Products, Riverside, CA), which was connected to the breathing circuit through low-resistance tubing.
Subjects Patients. The first 14 patients listed in table 1 participated in the incremental protocol. They were selected from a population of 48 patients with CHF based on the finding of PB during an incremental treadmill or a cycle ergometer exercise test. Tests were done for various clinical indications or as a part of research protocols. All patients had an es-
777 TABLE 1 SUBJECT CHARACTERISTICS Age (yr)
Sex
Ethnic Origin
Diagnosis
EF (%)
Treatment
61 52 36 73 77 67 63 71 65 63 69 51 48 36 54 47 59
M M M M M F M M M M M M M M M M M
C A C C H C C C C H C C A C A C H
CAD CAD IHD CAD CAD CAD CAD CAD CAD HTN HTN CAD VHD IHD IHD IHD HTN
20 26 20 NO 25 38 20 40 18 24 24 18 30 20 15 20 28
Dg Dt Vd Dt Vd Dg Dt Vd Dg Dt Nt Dg Dt Nt Vd Dg Dt Nt Dg Dt Dg Vd Dg Dt Dt Nt Vd Dg Dt Vd Dg Dt Dg Dg Dt Nt Dg Dt Nt Vd Dg Dt Nt Vd Dg Dt Nt Vd
1
37
2 3 4 5 6 7
34
F M M M M M F
C C C C C
Exercise Mode
Patients
1 2 3 4 5 6 7 8 9 10 11 12 13* 14 15 16 17 Control subjects
35 42 39 40 31
0 C
C T C T T T T T T T T T C+T C+T C
C C C C C T+C T+C T T
Definitionof abbreviations: EF .. resting ejection fraction; A • African American; C .. Caucasian; H - Hispanic; 0 - Oriental; CAD .. coronary artery disease; IHD • idiopathic heart disease; HTN .. hypertension; VHD .. valvular heart disease; NO _ no data; Dg • digoxin; Dt • diuretics; Nt .. nitrates, Vd .. vasodilator; C .. cycle ergometer; T .. treadmill. • Chronic atrial fibrillation.
timated resting left ventricular ejection fraction (echocardiography or radionuclear or contrast ventriculography) of ~ 40070. The five patients (patients 13through 17)who participated in the constant WR protocol were selectedon the same basis as the first 14.Clinical data on all 17patients are shown in table 1. Controlsubjects. Seven healthy volunteers participated (table 1).Four of them (Subjects 1 to 4) performed the incremental protocol on a treadmill, and five of them (Subjects 3 to 7) performed the constant WR protocol on a cycleergometer. The study was approved by the Institutional Human Subjects Committee, and all subjects signed an informed consent.
Data Analysis Incremental tests. From the graphic plots of studies of 48 patients with CHF (246studies), 14patients were found to have a clear oscillatory breathing. To minimize the effect of random variability in breathing pattern, breathby-breath data were smoothed using an eightbreath moving average filter, Only studies with a marked oscillation of 30 to 60 s duration were selected for analysis (see below).
Amplitudeandphaserelationship measurements. PB cycles weredisplayedon an expanded time scale for VE, V02, Veo2' R, HR, O 2 pulse, and PETeo2' From these plots, the amplitude of oscillation of each variable was computed. The magnitude of the oscillations of the PB (A) was defined as the difference
between the peak and nadir value of the variable divided by the mean value of the variable over the time period of the oscillation [A = (peak - nadir)/mean]. If AVE was ~ 25% in at least two consecutive PB cycles during exercise,the study was included in the analysis. The study showing the largest oscillations was used for a given patient. Phase relationships of the oscillations were also determined from expanded time scale plots. The length of each cycle (nadir to nadir) was determined visually and was normalized to 360 degrees. The phase difference (degree) between a given pair of variables was calculated from the time difference between the peaks or nadirs of the variables. Oscillations of a given variable were referred to as positive or negative with respect to the VE change, depending on whether the change was in the same ( + )A or the opposite (- )A direction as the change in VE. Magnitude of oscillations were calculated for multiple PB cycles for each subject (two to seven per patient and six to eight per control subject), and average values for each subject were used for calculating the reported amplitude and phase relationship.
Constant Work-Rate Tests Calculation of the breath-by-breath change in EELV during PB. For calculation of the dynamic change of EELV, the breath-bybreath expiratory VT(VTE) (expressedat BTPS) was subtracted from the corresponding in-
778
:J
BEN-DOV, SIElSEMA, CASABURI, AND WASSERMAN
50
0
... te
~...J
-4
'
CU .~
"e......
0.6
=
0.4
ILl
.>
l:
0.8 30 20
N
0
c-
0.2
-8
~
0 4
2
0
~
E
120 0 0
"2 l'll
Ql
E
0
~
3.0
.~
C\l
0
0
90 DO 0
Q:lo
60
Q:]CW
°eP
-12
o
30
::; > ..J
L&J L&J
.:
l:
~
2.0
70
ILl
">
1.0
40
0 .>
cu
3
6
9
TIME (min)
0
.c
Co)
120
240
360
Time (sec) Fig. 1. Calculation of the changein end-expiratory lung volume(EELV) in a normalsubjectcycling at 70 W.The cumulativebreath-by-breath differencebetweenthe inspiratory and expiratory tidal volume (VTI - VTe) is shownby the solid line.This differencereachesa negative valueof -12 L at end-exercise becauseof the higher temperature and humidity of the expired air. However, this curveclearlyoscillatesas it falls, and five distinct cycles are seen.Tocorrect for the negativetrend resulting from the higher temperature and humidity of VTe(a v due toTO andH20), a linearregressionthrough the cumulative difference is calculated (dsshed line). The value of this regression on the y axis at any given time is subtracted fromthe actualcumulative VTI - VTe. The difference shown on the lower panel on a larger scaledescribesthe actualchangesin EELV during this interval(seetext).Hatchedareas = EELV rise; stippled areas = EELVfall.
spiratory VT(Vn) (expressedat ambient temperature, pressure, and humidity). The difference (Vn - VTE) showed breath-to-breath fluctuations. However, the cumulative difference (Vn - VTE) throughout the exercisehad a negative slope because of the higher temperature and humidity of the expired gas. Although there was on average a higher volume exhaled than inhaled, it was assumed that there was no net change in EELV over the course of the exercise. Thus, the breath-bybreath change in EELV is represented by the difference between Vn and VTE, corrected for the instantaneous change in volume attributed to temperature and water vapor. The cumulative volume difference and the component of the difference attributed to water vapor and temperature are shown in figure 1. The sum of (Vn - VTE) minus the component of this cumulative value, which is due to temperature and humidity, represents the fluctuation of EELV, plus a small effect caused by instantaneous change in R because of hypoventilation or hyperventilation.
Statistics The incremental and constant WR studies were analyzed independently. Within-group
Fig. 2. Representative studies showing spontaneous periodic breathing (PB) in Patient5 with CHF (A) and volitionalPB in control SUbject 5 (8) during incremental exercisetesting. In both cases,exercisestarts at Minute 2. The amplitude of the ventilatory(Ve)oscillations are 52 ± 17 and 57 ± 11% of the respective mean values.The amplitude of the V02 oscillationsare 76 ± 27 and31 ± 6% of the meanfor the patientand control subject, respectively. Despitea similar amplitude in Ve for both subjects, the amplitude of Vo2 is larger in the patient.Notethat in the patientwith CHF, smalleroscillations are present at rest.
analysis was done using two-tailed paired t tests. Analysis of variance with means testing by the Student-Neuman-Keul method was used for multiple comparisons. Between groups comparisons weredone using the unpaired t test. Differences wereaccepted as significant at p < 0.05. Findings are shown as mean ± 1 SD.
Results
Incremental Studies Magnitudeofoscillations. In incremental exercise studies from 14 patients with CHF, 50 PB cycles were selected (mean, 3.5; range, 2 to 7 cycles/patient). Thirtyfour PB cycles (mean, 8.5; range, 5 to 12 cycles/subject) were analyzed in the four control subjects. Examples of exercise spontaneous PB in a patient and simulated PB in a control subject are shown in figure 2. In the patient the amplitude of the oscillations of VE and V02 were 52 ± 12070 and 76 ± 270/0, respectively, with a period of 0.76 min/cycle. In the control subject, the amplitude of the simulated oscillations of VE and V02 were 57 ± 110/0 and 31 ± 60/0, respectively. The major difference between the natural PB (CHF) and the simulated PB (control) was that the fluctuation of V02 was usually more prominent than the simultaneous VE oscillations in the former, and smaller than the VE in the latter. This is
V
11'
...J N
a.
·i• •: ••••
•••«*:.-:
l:
0
ao -0.5 c
0
D.
~
...J
0.5
identityline
0
a
6
100
~
(,)
150
40
C
&IJ
~
1.2
A
0
0
30
60
90
120
150
tilE (% of mean) Fig. 3. The amplitude, a [a = (peak - nadir)/mean], of theventilatory (VE) oscillationsduringincremental exercise tests is plotted against the amplitude of the O2 uptake(\toJl oscillationsof the corresponding cyclesfor the patientswith CHF (50 spontaneous PB cycles)and for the control subjects (34 simulated PB cycles). The line of identityis shown. On average, a V02 was higher in the patientswith CHF. Opensquares = patientswith CHF; asterisks = control subjects.
illustrated in figure 3 for all subjects of both groups. The mean values of the nadir, peak, and the corresponding percent of mean oscillation (a 0/0) for each cardiorespiratory variable for patients and for control subjects are summarized in table 2. Despite similar aVE (49 versus 48070), a V02 was higher in the patients (63 versus 25070). Phase relationship. The phase relationships between VE and PETC02 oscillations in Patient 5 (spontaneous PB) and in Subject 6 (volitional PB) are shown in figure 4. In the control subject, a decline in PETC02 coincides with a rise in VE (peak PETC02 lags nadir VE by 0 ± 0 degrees), as expected if the change in PETC02 were due solely to a change in alveolar ventilation. In contrast, in the patient with CHF there was 48 ± 8 degrees phase delay between the nadir of VE and the peak of PETC02. The distribution of phase differences between VE and PETC02 for the 50 cycles of the patients with CHF and the 34 cycles of the control subjects is shown in figure 5. In the control subjects the mean phase difference between nadir VE and peak PETC02' 4 ± 8 degrees, is not significantly different from zero. In the patients, the difference of 44 ± 14 degrees is significantly larger than zero (P < 0.01). This suggests that CO2 flow to the lungs during the cyclicbreathing remained relatively constant in the control subjects, but CO2 flow to the lung is cyclic in the patients with CHF.
779
PERIODIC BREATtiNG AND O. UPTAKE OSCILLAnONS IN CHF
TABLE 2 MAGNITUDE OF OSCILLATIONS IN CARDIORESPIRATORY VARIABLES DURING INCREMENTAL EXERCISE IN PATIENTS WITH CHF (SO CYCLES) AND IN CONTROL SUBJECTS (34 CYCLES) CHF; Spontaneous PB Variable
Nadir
22 466 386 0.78 PETC02, mm Hg 35 HR, beat/min * 123 VoJHR, mllmin/beat 4.0 VE, Llmin V0 2, mllmin vco; mllmin R
± ± ± ± ± ± ±
Control; Voluntary PB
&(%)
Peak
7 36 ± 11 223 839 ± 293 185 725 ± 244 0.1 0.93 ± 0.1 7 30 ± 6 23 127 ± 22 1.5 6.5 ± 1.5
49 ± 15 63 ± 25 65 ± 29 16 ± 11 -14 ± 10 2±6 22 ± 30
Nadir
42 ± 13 1,570 ± 448 1,3n ± 525 0.86 ± 0.08 41 ± 6 136 ± 13 11 ± 3
Peak
69 2,009 2,057 1.09 36 147 14
± ± ± ± ± ± ±
24 548
613 0.15 6 14 3
&(%)
48 25 43 23 -13 8 18
± ± ± ± ± ± ±
12 11tt: 27tt: 11t 8 7t 15t
~nlt#on of tJbbf'eV/8f!
ISSAHAR BEN-DOV, KATHY E. SIETSEMA, RICHARD CASABURI, and KARLMAN WASSERMAN
Introduction Periodic breathing (PB) resembling Cheyne-Stokes respiration (CSR) has been described in patients with congestive heart failure (CHF) at rest and during exercise (1-3). This has been attributed to unstable ventilatory control caused by slow circulation, leading to delayed transmission of humoral ventilatory stimuli to the chemoreceptors (4-7). However, blood pressure and cerebral perfusion also oscillate during CSR (1, 8), suggesting that circulatory control is also involved. Because O 2 uptake (V02) depends on pulmonary blood flow, insight into the role of the circulation in PB can be gained by studying V02, which is also oscillatory (9), breath by breath, in patients with CHF during exercise. Oscillations of V02 measured from respired gas could represent true fluctuations of O 2 across the alveolar-pulmonary capillary interface. Alternatively, these fluctuations might merely reflect cyclic changes in lung gas stores, because of changing end-expiratory lung volume (EELV) and O2 concentration, induced by ventilatory fluctuations. True transalveolar fluctuations in V02 cannot take place unless pulmonary blood flow and/or metabolic rate oscillate, since arterial blood is nearly fully saturated with O 2 at the Pao, usually present in these patients during exercise (2). The Fick relatjonship, V02 = Q(Cao2 - Cv02), where Q = cardiac output, C = content, a = arterial and v = mixed venous, dictates that true alveolar V02 fluctuates when pulmonary blood flow or CV02oscillates. We hypothesized that lung gas stores do not account for the prominent oscillations in V01 observed during exercise in patients with CHF, and that these oscillations are more likely attributed to oscillations of pulmonary blood flow. To test this hypothesis we analyzed the fluctuation and phase relationship of ventilation (VE) and V02 in 17 patients with CHF who showed marked spontaneous PB during exercise. To examine to what 776
SUMMARY Periodic breathing (PB) during exerclH In patients with congestive heart failure (CHF) hypothellzed that the V0 2 Is associated with prominent oscillations (OSC) of O2 Uptalc8 (V02)' OSC represent OSC In true O2 exchange, resulting from concomitant cardiac output fluctuations and are not merely due to OSC of lung O2 stores. compared the amplitude of the OSC of V0 2 , ventilation (VE), and end-explratory lung volume (EELV)In 17patients with CHF and PB and In seven healthy control subjects who volltlonally simulated PB. Sublects underwent an Incremental and/or a constant work-rate exerclH test. VE and v0 2 were measured breath by breath. EELVchange was estimated by summing the difference between Inspiratory and expiratory tidal volumes for each breath. The amplitude of the OSc, a, Is expre888d as the ratio of the difference between the peak (peak - nadlr)/mean]. In CHF,during and nadir of the oscillating variable divided by Its mean [a Incremental testing, the amplitude of the VE OSC was smaller than that of the v0 2 OSC (aVE = 49 ± 15% [SD], a v0 2 63 ± 25%, P < 0.01). In contrast, during volitional PB In the control subJects, VE OSC were larger than v0 2 OSC (aVE = 48 ± 12%, a v0 2 = 25 ± 11%, P < 0.01). This suggests that changing VE Itself cannot account for the marlc8d v0 2 OSC seen In CHF. In the patients, EELVshowed no systematic Osc, did not correlate with a v0 2, and wei not IlgnHlcantly different from zero. In contrast, In the control subjects, EELV tended to be lower at peak VE (-0.14 ± 0.5 L, P = 0.0&), and It was well correlated with a V0 2 (r = 0.6, P < 0.01). In conclusion, v0 2 oscillations In patients with CHF are more prominent than the associated VE OSC, are greater than the v0 2 OSC that can be Induced by volitional PB, and are not associated with changes In EELV. This supports the hypothesis that cardiac output OSC contribute to v0 2 OSC.
we
we
=
=
AM REV RESPIR DIS 1912; 145:778-781
extent these gas exchange oscillations could be induced by the breathing pattern itself, we compared the V02 fluctuations observed in the patients with CHF with the V02 oscillations induced in control subjects who volitionally oscillated their breathing pattern to simulate the natural PB under the same exercise conditions. We found that in the patients with CHF who exhibited PB, the amplitude of the V02 oscillations during exercise were more prominent relative to the concomitant VEoscillations and greater than the V02 oscillations induced by the simulated PB in the control group. Methods Protocols 1\vo protocols were employed: a progressive incremental exercisetest performed on either a treadmill or a cycle ergometer, and a constant work-rate (WR) low intensity test performed on a cycle ergometer. Fourteen patients with CHF who had oscillatory breathing during exercise and four control subjects who simulated oscillatory breathing pattern performed the incremental exercise.Five patients with CHF who had oscillatory breath-
ing during exercise and five control subjects who simulated oscillatory breathing pattern participated in the constant WR protocol.
Incremental Exercise Studies 'Ireadmill tests. All treadmill studiesweredone according to a modified Naughton protocol (10). After a short period of collecting resting data, the speed and/or the slope of the treadmill were increased every 2 min from 1 mile/h and zero degrees to a maximum of 3.5 miles/h and 16 degrees or to the limit of tolerance. Cycle ergometry. After 2 to 3 min of unloaded pedaling, the WR was increased ev(Received in original form April 15, 1991 and in revised form September 16, 1991) 1 From the Division of Respiratory and Critical Care, Physiology and Medicine, Harbor-UCLA Medical Center, Torrance, California. 2 Supported in part by Grant HL-l1907 from the National Heart, Lung, and Blood Institute and by the Research and Education Institute, HarborUCLA Medical Center. 3 Correspondence and requests for reprints should be addressed to Dr. Issahar Ben-Dov, Division of Respiratory and Critical Care, Physiology and Medicine,Harbor-UCLA Medical Center, 1000 W. Carson Street. Torrance, CA 90509.
PERIODIC BREATHING AND O. UPTAKE OSCILLATIONS IN CHF
ery 1 min in steps, or continuously in a ramp pattern, at a rate of 10 or 15 W/min. During the incremental tests subjects respired through a mouthpiece connected to a low resistance valve (dead space = 100 ml) with the nose occluded. Expiratory flow was recorded by a pneumotachograph and integrated to determine volume. Fractional concentrations of expired CO 2and O2weremeasured by infrared CO 2and fuel cellO 2analyzer. End-tidal pressure of CO 2(pETe02), VE, V02, and Veo2were measured breath by breath at the mouth using a computerized system (System2001or CPX; Medical Graphics Corp., St. Paul, MN). In addition, Veo2IVo2 (R), mean heart rate (HR), and O 2 pulse (Vo2/HR) were also calculated breath by breath.
Constant Work-Rate Exercise Studies The constant WR exercise tests were performed for 6 min on a cycle ergometer. The WR was 25 W for the patients and 50 to 70 W for the control subjects. In each case, the WR chosen was below the subject's anaerobic threshold (AT) as determined from a prior incremental study. Measurements of VE and gas exchange were made using a system designed in this institution (11). A mass spectrometer (Perkin-Elmer Medical Instruments, Oakbrook, IL) was used to measure respired O2, CO 2, and N 2 concentrations. A turbine flow-meter (SensorMedics, Anaheim, CA) was used to measure inspiratory and expiratory tidal volume (VT).
Voluntary Periodic Breathing during Exercise Healthy subjects were guided to volitionally simulate PB during exercise. This was done in order to contrast the gas exchange induced by pure ventilatory oscillations with that of spontaneous PB as found in patients with CHF. Subjects were instructed to oscillate their VT (respiratory rate was maintained constant) while exercising. The period of the voluntary oscillation was 1 min/cycle,approximating that observed in the patients. The amplitude of the VT oscillation was chosen so that the resulting voluntarily induced VE fluctuations would match the mean fluctuation of the patients' spontaneous PB, which was also primarily induced by VT changes. Subjects were assisted to achieve the desired ventilatory pattern by a metronome and a display of the actual and the target VT (Ventilation Monitor LS-80; Bear Medical Products, Riverside, CA), which was connected to the breathing circuit through low-resistance tubing.
Subjects Patients. The first 14 patients listed in table 1 participated in the incremental protocol. They were selected from a population of 48 patients with CHF based on the finding of PB during an incremental treadmill or a cycle ergometer exercise test. Tests were done for various clinical indications or as a part of research protocols. All patients had an es-
777 TABLE 1 SUBJECT CHARACTERISTICS Age (yr)
Sex
Ethnic Origin
Diagnosis
EF (%)
Treatment
61 52 36 73 77 67 63 71 65 63 69 51 48 36 54 47 59
M M M M M F M M M M M M M M M M M
C A C C H C C C C H C C A C A C H
CAD CAD IHD CAD CAD CAD CAD CAD CAD HTN HTN CAD VHD IHD IHD IHD HTN
20 26 20 NO 25 38 20 40 18 24 24 18 30 20 15 20 28
Dg Dt Vd Dt Vd Dg Dt Vd Dg Dt Nt Dg Dt Nt Vd Dg Dt Nt Dg Dt Dg Vd Dg Dt Dt Nt Vd Dg Dt Vd Dg Dt Dg Dg Dt Nt Dg Dt Nt Vd Dg Dt Nt Vd Dg Dt Nt Vd
1
37
2 3 4 5 6 7
34
F M M M M M F
C C C C C
Exercise Mode
Patients
1 2 3 4 5 6 7 8 9 10 11 12 13* 14 15 16 17 Control subjects
35 42 39 40 31
0 C
C T C T T T T T T T T T C+T C+T C
C C C C C T+C T+C T T
Definitionof abbreviations: EF .. resting ejection fraction; A • African American; C .. Caucasian; H - Hispanic; 0 - Oriental; CAD .. coronary artery disease; IHD • idiopathic heart disease; HTN .. hypertension; VHD .. valvular heart disease; NO _ no data; Dg • digoxin; Dt • diuretics; Nt .. nitrates, Vd .. vasodilator; C .. cycle ergometer; T .. treadmill. • Chronic atrial fibrillation.
timated resting left ventricular ejection fraction (echocardiography or radionuclear or contrast ventriculography) of ~ 40070. The five patients (patients 13through 17)who participated in the constant WR protocol were selectedon the same basis as the first 14.Clinical data on all 17patients are shown in table 1. Controlsubjects. Seven healthy volunteers participated (table 1).Four of them (Subjects 1 to 4) performed the incremental protocol on a treadmill, and five of them (Subjects 3 to 7) performed the constant WR protocol on a cycleergometer. The study was approved by the Institutional Human Subjects Committee, and all subjects signed an informed consent.
Data Analysis Incremental tests. From the graphic plots of studies of 48 patients with CHF (246studies), 14patients were found to have a clear oscillatory breathing. To minimize the effect of random variability in breathing pattern, breathby-breath data were smoothed using an eightbreath moving average filter, Only studies with a marked oscillation of 30 to 60 s duration were selected for analysis (see below).
Amplitudeandphaserelationship measurements. PB cycles weredisplayedon an expanded time scale for VE, V02, Veo2' R, HR, O 2 pulse, and PETeo2' From these plots, the amplitude of oscillation of each variable was computed. The magnitude of the oscillations of the PB (A) was defined as the difference
between the peak and nadir value of the variable divided by the mean value of the variable over the time period of the oscillation [A = (peak - nadir)/mean]. If AVE was ~ 25% in at least two consecutive PB cycles during exercise,the study was included in the analysis. The study showing the largest oscillations was used for a given patient. Phase relationships of the oscillations were also determined from expanded time scale plots. The length of each cycle (nadir to nadir) was determined visually and was normalized to 360 degrees. The phase difference (degree) between a given pair of variables was calculated from the time difference between the peaks or nadirs of the variables. Oscillations of a given variable were referred to as positive or negative with respect to the VE change, depending on whether the change was in the same ( + )A or the opposite (- )A direction as the change in VE. Magnitude of oscillations were calculated for multiple PB cycles for each subject (two to seven per patient and six to eight per control subject), and average values for each subject were used for calculating the reported amplitude and phase relationship.
Constant Work-Rate Tests Calculation of the breath-by-breath change in EELV during PB. For calculation of the dynamic change of EELV, the breath-bybreath expiratory VT(VTE) (expressedat BTPS) was subtracted from the corresponding in-
778
:J
BEN-DOV, SIElSEMA, CASABURI, AND WASSERMAN
50
0
... te
~...J
-4
'
CU .~
"e......
0.6
=
0.4
ILl
.>
l:
0.8 30 20
N
0
c-
0.2
-8
~
0 4
2
0
~
E
120 0 0
"2 l'll
Ql
E
0
~
3.0
.~
C\l
0
0
90 DO 0
Q:lo
60
Q:]CW
°eP
-12
o
30
::; > ..J
L&J L&J
.:
l:
~
2.0
70
ILl
">
1.0
40
0 .>
cu
3
6
9
TIME (min)
0
.c
Co)
120
240
360
Time (sec) Fig. 1. Calculation of the changein end-expiratory lung volume(EELV) in a normalsubjectcycling at 70 W.The cumulativebreath-by-breath differencebetweenthe inspiratory and expiratory tidal volume (VTI - VTe) is shownby the solid line.This differencereachesa negative valueof -12 L at end-exercise becauseof the higher temperature and humidity of the expired air. However, this curveclearlyoscillatesas it falls, and five distinct cycles are seen.Tocorrect for the negativetrend resulting from the higher temperature and humidity of VTe(a v due toTO andH20), a linearregressionthrough the cumulative difference is calculated (dsshed line). The value of this regression on the y axis at any given time is subtracted fromthe actualcumulative VTI - VTe. The difference shown on the lower panel on a larger scaledescribesthe actualchangesin EELV during this interval(seetext).Hatchedareas = EELV rise; stippled areas = EELVfall.
spiratory VT(Vn) (expressedat ambient temperature, pressure, and humidity). The difference (Vn - VTE) showed breath-to-breath fluctuations. However, the cumulative difference (Vn - VTE) throughout the exercisehad a negative slope because of the higher temperature and humidity of the expired gas. Although there was on average a higher volume exhaled than inhaled, it was assumed that there was no net change in EELV over the course of the exercise. Thus, the breath-bybreath change in EELV is represented by the difference between Vn and VTE, corrected for the instantaneous change in volume attributed to temperature and water vapor. The cumulative volume difference and the component of the difference attributed to water vapor and temperature are shown in figure 1. The sum of (Vn - VTE) minus the component of this cumulative value, which is due to temperature and humidity, represents the fluctuation of EELV, plus a small effect caused by instantaneous change in R because of hypoventilation or hyperventilation.
Statistics The incremental and constant WR studies were analyzed independently. Within-group
Fig. 2. Representative studies showing spontaneous periodic breathing (PB) in Patient5 with CHF (A) and volitionalPB in control SUbject 5 (8) during incremental exercisetesting. In both cases,exercisestarts at Minute 2. The amplitude of the ventilatory(Ve)oscillations are 52 ± 17 and 57 ± 11% of the respective mean values.The amplitude of the V02 oscillationsare 76 ± 27 and31 ± 6% of the meanfor the patientand control subject, respectively. Despitea similar amplitude in Ve for both subjects, the amplitude of Vo2 is larger in the patient.Notethat in the patientwith CHF, smalleroscillations are present at rest.
analysis was done using two-tailed paired t tests. Analysis of variance with means testing by the Student-Neuman-Keul method was used for multiple comparisons. Between groups comparisons weredone using the unpaired t test. Differences wereaccepted as significant at p < 0.05. Findings are shown as mean ± 1 SD.
Results
Incremental Studies Magnitudeofoscillations. In incremental exercise studies from 14 patients with CHF, 50 PB cycles were selected (mean, 3.5; range, 2 to 7 cycles/patient). Thirtyfour PB cycles (mean, 8.5; range, 5 to 12 cycles/subject) were analyzed in the four control subjects. Examples of exercise spontaneous PB in a patient and simulated PB in a control subject are shown in figure 2. In the patient the amplitude of the oscillations of VE and V02 were 52 ± 12070 and 76 ± 270/0, respectively, with a period of 0.76 min/cycle. In the control subject, the amplitude of the simulated oscillations of VE and V02 were 57 ± 110/0 and 31 ± 60/0, respectively. The major difference between the natural PB (CHF) and the simulated PB (control) was that the fluctuation of V02 was usually more prominent than the simultaneous VE oscillations in the former, and smaller than the VE in the latter. This is
V
11'
...J N
a.
·i• •: ••••
•••«*:.-:
l:
0
ao -0.5 c
0
D.
~
...J
0.5
identityline
0
a
6
100
~
(,)
150
40
C
&IJ
~
1.2
A
0
0
30
60
90
120
150
tilE (% of mean) Fig. 3. The amplitude, a [a = (peak - nadir)/mean], of theventilatory (VE) oscillationsduringincremental exercise tests is plotted against the amplitude of the O2 uptake(\toJl oscillationsof the corresponding cyclesfor the patientswith CHF (50 spontaneous PB cycles)and for the control subjects (34 simulated PB cycles). The line of identityis shown. On average, a V02 was higher in the patientswith CHF. Opensquares = patientswith CHF; asterisks = control subjects.
illustrated in figure 3 for all subjects of both groups. The mean values of the nadir, peak, and the corresponding percent of mean oscillation (a 0/0) for each cardiorespiratory variable for patients and for control subjects are summarized in table 2. Despite similar aVE (49 versus 48070), a V02 was higher in the patients (63 versus 25070). Phase relationship. The phase relationships between VE and PETC02 oscillations in Patient 5 (spontaneous PB) and in Subject 6 (volitional PB) are shown in figure 4. In the control subject, a decline in PETC02 coincides with a rise in VE (peak PETC02 lags nadir VE by 0 ± 0 degrees), as expected if the change in PETC02 were due solely to a change in alveolar ventilation. In contrast, in the patient with CHF there was 48 ± 8 degrees phase delay between the nadir of VE and the peak of PETC02. The distribution of phase differences between VE and PETC02 for the 50 cycles of the patients with CHF and the 34 cycles of the control subjects is shown in figure 5. In the control subjects the mean phase difference between nadir VE and peak PETC02' 4 ± 8 degrees, is not significantly different from zero. In the patients, the difference of 44 ± 14 degrees is significantly larger than zero (P < 0.01). This suggests that CO2 flow to the lungs during the cyclicbreathing remained relatively constant in the control subjects, but CO2 flow to the lung is cyclic in the patients with CHF.
779
PERIODIC BREATtiNG AND O. UPTAKE OSCILLAnONS IN CHF
TABLE 2 MAGNITUDE OF OSCILLATIONS IN CARDIORESPIRATORY VARIABLES DURING INCREMENTAL EXERCISE IN PATIENTS WITH CHF (SO CYCLES) AND IN CONTROL SUBJECTS (34 CYCLES) CHF; Spontaneous PB Variable
Nadir
22 466 386 0.78 PETC02, mm Hg 35 HR, beat/min * 123 VoJHR, mllmin/beat 4.0 VE, Llmin V0 2, mllmin vco; mllmin R
± ± ± ± ± ± ±
Control; Voluntary PB
&(%)
Peak
7 36 ± 11 223 839 ± 293 185 725 ± 244 0.1 0.93 ± 0.1 7 30 ± 6 23 127 ± 22 1.5 6.5 ± 1.5
49 ± 15 63 ± 25 65 ± 29 16 ± 11 -14 ± 10 2±6 22 ± 30
Nadir
42 ± 13 1,570 ± 448 1,3n ± 525 0.86 ± 0.08 41 ± 6 136 ± 13 11 ± 3
Peak
69 2,009 2,057 1.09 36 147 14
± ± ± ± ± ± ±
24 548
613 0.15 6 14 3
&(%)
48 25 43 23 -13 8 18
± ± ± ± ± ± ±
12 11tt: 27tt: 11t 8 7t 15t
~nlt#on of tJbbf'eV/8f!
