Effects of continuous positive airway pressure on cardiac output and plasma norepinephrine in sedated pigs

Effects of Continuous Positive Airway Pressure on Cardiac Output and Plasma Norepinephrine in Sedated Pigs Steven

M. Schatf,

Ling Chen, David Slamowitz,

Parinam

S. Rao

Continuous positive airway pressure (CPAP) increases cardiac output (CO) in congestive heart failure (CHF). In six sedated pigs that were normovolemic (NV) and hypervolemic (HV), and seven previously instrumented pigs with pacing-induced CHF, we tested the hypothesis that this is associated with decreased total body sympathetic nerve activity (SNA). Hemodynamic variables and plasma norepinephrine level measurements were measured at baseline, CPAP 5 and 10 cm HrO, and recovery. Arterial 0s saturation was maintained at greater than or equal to 90% and PCO~ did not change. For NV baseline plasma norepinephrine level (PNE) was 97 f 61 pg/mL, CO 2.4 2 .5 L/min, and pulmonary wedge pressure (Pw) 10.1 f 2.4 mmHg and did not change with CPAP. HV and CHF were associated with increased baseline Pw (18-21 mmHg). Base-

line CO was increased with HV and unchanged with CHF. Baseline PNE was increased 4 to B-fold with both HV and CHF. CO increased at CPAP 5 compared with baseline with both HV and CHF. However, PNE decreased with CPAP in HV, and increased with CPAP in CHF. Increased CO was always associated with decreased systemic vascular resistance. We conclude the following: (1) increased CO with CPAP can be associated with either increasing or decreasing SNA; (2) CPAP can produce increases in CO when the heart is distended whether baseline LV function is relatively normal (HV) or depressed (CHF); and (3) there are probably a number of different mechanism increasing CO with CPAP and these may vary from condition to condition. Copyright 0 1996 by W.B. Saunders Company

ONTINUOUS POSITIVE airway pressure (CPAP) is often used in the treatment of C patients with congestive heart failure (CHF).

that changes in autonomic function with CPAP could be responsible for LV unloading. Patients with CHF experience a number of alterations in whole body sympathetic nervous system activity (SNA). Chief among these is an elevation in tonic SNA activity, resulting in chronic elevations in baseline plasma norepinephrine levels (PNE). 7,9~31 Other abnormalities of SNA in CHF include downregulation of cardiac adrenergic @receptors4 decreased myocardial norepinephrine levels5 and increased myocardial norepinephrine turnover.16J7 Although enhanced SNA could acutely act to maintain blood pressure and CO, with chronicity, vasoconstriction, myocardial necrosis, and depletion of myocardial p-receptors could actually worsen CHF. 9~31In patients with sleep disordered breathing on the basis of severe CHF, Naughton et a12i recently showed decreased daytime PNE and urine norepinephrine levels after 1 month of nocturnal CPAP therapy.

Studies in humans with severe CHF and CheyneStokes respiration at night showed improved daytime left ventricular (LV) function.20,21 In acute CHF, CPAP leads to improved gas exchange and decreased work of breathing.1$24 It is clear that the acute application of CPAP leads to increased cardiac output (CO) in at least some patients with CHF.1J,28 Recently, we studied the acute effects of CPAP in sedated, previously instrumented pigs.i”J1J5 In normovolemic (NV) animals, low-level CPAP did not change CO, although high-level CPAP decreased CO. However, in hypervolemic (HV) animals10,15 and in animals with CHF,” lowlevel CPAP (5-10 cm H20) was associated with increased CO and improved LV function. The mechanisms by which CO improved were not entirely clear. It has been postulated that because of increased intrathoracic pressure surrounding the heart, LV afterload, as measured by LV end-systolic transmural pressure, may decrease with CPAP.*” However, this did not prove to be the case when LV transmural pressure was measured directly.15 Arterial vasodilation has been observed in both animals1’~‘5 and humans3 with CHF when exposed to CPAP. This could also provide a mechanism whereby LV afterload could decrease, thus leading to increased ejection. This observation suggests JournalofCriricalCare,

Vol 11,No

2(June),1996:

pp 57-64

From the Divisions of Pulmonary and Critical Care and Cardiothoracic Surgery, Long Island Jewish Medical Center, the Long Island Campus for the Albert Einstein College of Medicine, New Hyde Park, NY. Supported by National Institutes of Health Grant No. ROI HL 49808-OlA2. Address correspondence to Steven M. Scharj MD, PhD, Pulmonary and Critical Care Division, Long Island Jewish Medical Center, New Hyde Park, NY 11042. Copyright 0 I996 by K B. Saunders Company O&73-944/19611102-0002$05.00l0 57

SCHARF

58

These investigators suggested that decreased SNA (as reflected in PNE) with CPAP could be the mechanism for improved cardiac function. However, these investigators did not assess the acute response of SNA and correlate this with the acute response of CO. In the present study, we wanted to determine whether changes in total body SNA were correlated with the changes in CO with acute applications of low-level CPAP, and whether the change in CO could be predicted by the change in PNE. We reasoned that if decreased total body SNA, as reflected in decreased PNE, was the mechanism responsible for increased CO with CPAP, then under whatever conditions CPAP led to an increase in CO, and PNE should decrease. Hence, we assessed the cardiovascular and PNE response to CPAP under those conditions under which low-level CPAP had previously been shown to increase CO, namely HV and CHF as well as the normovolemic baseline. To avoid the confounding effects of decreased CO with high levels of CPAP,‘“,11,15 CPAP was limited to the upper limit of its previously shown therapeutic range. MATERIALS

AND

METHODS

Studies were performed on six normal pigs and seven pigs in which CHF was induced by rapid ventricular pacing.6,11,13 All procedures related to surgery, analgesia, anesthesia, and euthanasia were approved by the Institutional Animal Care and Use Committee and were consistent with National Institutes of Health guidelines. At no time in the studies were the animals observed to exhibit any signs of discomfort or pain. Studies in Six Normal Pigs

Conditioned Yorkshire farm pigs, weight 16 to 20 Kg, free of ecto-parasites and endoparasites, were sedated with fentanyl-droperido1 (Innovar-vet, Pittman-Moore; Mandelein, IL) 1 mL/15 kg and ketamine 10 mg/kg intramuscularly (IM) for instrumentation and intubation and maintained on a fentanyl/droperidol drip at approximately .09 ml/kg/h during data collection. The fentanyl/droperidol drip was adjusted to allow the animal to be awake, but sedated and tranquil with an active cornea1 reflex. A cutdown was performed in the groin under local anesthesia and the femoral artery

ET AL

and vein were isolated. A thermistor-tipped catheter was passed via the femoral vein into the pulmonary artery for measurement of pulmonary arterial pressure (PAP), pulmonary occlusion or wedge pressure (Pw) and CO by thermodilution. A catheter was passed via the femoral artery into the ascending aorta for measurement of mean arterial pressure (MAP), and withdrawal of blood for measurement of blood gas tensions. The groin wounds were closed and the animal was placed onto its right side. Pacing-Induced in Seven Pigs with CHF

These animals were prepared in three phases. During phase 1 animals were anesthetized using fentanyl/droperidol (Innovar-vet) 1 mL/15 kg IM and ketamine 10 mg/kg IM and maintained with endotracheal halothane 0.5% to 1.0%. The animals were mechanically ventilated with a tidal volume of 10 mL/kg and a respiratory rate of 12 to 15 per minute. Through a 3- to 4-in long left thoracotomy, the pericardial sac was opened. A unipolar epicardial pacemaker electrode was inserted into the base of the LV and sutured in place. The pericardium was loosely approximated. The pacemaker electrode was attached to a unipolar, externally programmable, pacemaker (Medtronics Inc, Secaucus, NJ) and placed into a separate ventral subcutaneous pocket near the left first nipple. The pacer had been modified to allow for rapid rates in the temporary mode, which was externally programmable. Initial pacer rate was set at the minimum of 30 bpm, far lower than the physiological rate of the animal. Thus, heart rate (HR) was physiological after surgery. The thoracotomy was closed in layers and covered with a dressing soaked in collodion that hardened quickly and kept the wound free of stool contamination for 3 to 4 days and allowed for proper healing. The animals were allowed to awaken and were placed in pens alone with feeding begun on the day after surgery. Penicillin (24,000 U/kg) and dihydrostreptomycin (30 mg/kg) IM was given daily for 3 consecutive days and morphine sulfate 5 mg IM was given every 6 hours for pain for the first 24 hours after surgery. Phase 2 took place 3 days after initial surgery. The animals were sedated with ketamine (1 mL/lO kg) and, by external radio frequency programming, the pacemaker rate increased to 230 bpm. Every 2 days until the data acquisition, the animal was

EFFECTS

OF CONTINUOUS

POSITIVE

AIRWAY

PRESSURE

sedated (ketarnine) and the HR confirmed by electrocardiogram. Phase 3 (data acquisition) took place 7 days after pacemaker activation. The animals were sedated and sedation was maintained as described above for normal animals. The pacemaker rate was lowered to the minimum rate (30 bpm) immediately after sedation. Thus, during this phase the animal set its own heart rate (HR). Under local anesthesia (2% lidocaine), the pacer pocket was incised and the pacer removed. The pocket was sutured closed and the animals instrumented as described above for normal animals. Animals were allowed to breathe from an enriched O2 mixture (25% to 35% 0,). At each data point, CO was measured by thermodilution by injecting 5 mL of iced saline into the right atrium. Measures were repeated in quadruplicate and the mean was taken. HR was obtained from the arterial pressure tracing. Systemic vascular resistance (SVR) was calculated as the ratio of MAP to CO and multiplying by 79.9 to scale to dynes-set-cm-5. Stroke volume (SV) was calculated by dividing CO by HR. Pulmonary vascular resistance (PVR) was calculated as [(mean PAP-Pw)/CO] x 79.9 (dynes-set-cmd5). For measurement of I-norepinephrine levels, at each data collection point, 3 mL of arterial blood were collected in heparinized tubes into which 50 ~J,L of metabisulfite solution 380 mg/mL was added, and spun at 2,000 rpm in a refrigerated centrifuge for 20 minutes. One milliliter of plasma was drawn off and added to 50 uL of 4M perchloric acid, spun at 3,000 rpm and decanted into plastic storage tubes. These were kept at -70°C for later measurement by a previously published high-pressure liquid chromatographic (HPLC) technique.27 Twenty microliters of sample was injected into the HPLC and catecholamines separated on a l&s column, mobile phase (85:15 H20-methanol, 1% ethylenediaminetetraacetic acid (EDTA), 2 mM I-heptane sulfonic acid and 1% acetic acid, pH 3.8) flow rate 1 mL/min. Standards were included along with each assay. CPAP was applied via the endotracheal tube using a high-flow system and a spring-loaded expiratory threshold valve (Vital Sign Inc; Totowa, NJ). High-flow rates were used to keep CPAP levels constant throughout the respiratory cycle, avoiding flow dependency of airway pressure.26 A pneumotach was placed in the

59

expiratory limb of the CPAP set-up, which allowed for counting respiratory rate (RR). EXPERIMENTAL

PROTOCOL

After instrumentation, animals were allowed to stabilize while breathing spontaneously for 30 to 45 min. Measurements were taken at CPAP = 0 (baseline). CPAP was then increased to 5 and then 10 cm H20. At each CPAP level, 20 to 30 min were allowed before any measurements were taken. After measurements at CPAP = 10 cm H20, CPAP was returned to 0 cm HZ0 for 20 to 30 minutes and repeat measurements taken (recovery). In the normal animals, after these measurements HV was produced by infusing hetastarch 35 mL/kg over approximately 20 minutes. After a stabilization period of 20 to 30 min the above protocol was repeated. At the end of the experiments, animals were killed by injection of 10 mL of 5 g/mL pentobarbital solution in saturated potassium chloride. After euthanasia, the position of all catheters was verified visually at necropsy. Data Analysis

Data were digitized in 30 second epochs at a frequency of 100 Hz and downloaded onto a personal computer (Compaq 386/25; Houston, TX) using commercially available software (DASA-Gould; Cleveland, OH). Data were expressed as mean ? standard deviation. Data were then analyzed by 2-factor (2-way) analysis of variance (ANOVA) (Sigmastat, Jandel; San Rafael, CA) one factor being status-NV, HV, and CHF-the other being CPAP level. When the trends between the status groups seemed different and/or the interaction term was significant, separate one-way ANOVA for repeated measures were performed for each status. If trends reached statistical significance, a posthoc analysis (Newman-Keuls test) was performed to determine the levels at which significance occurred. The null hypothesis was rejected at the 5% level. RESULTS

Table 1 shows changes in respiratory measurements and hemoglobin for the three conditions of NV, HV, and CHF. Hypoxemia did not occur. Baseline hypercapnia was noted under all conditions, but there were no significant differences between NV, HV, or CHF, nor

SCHARF

60

Table Variable PO, (mmHg)

Pcoz (mmHg)

1. Respiratory

CPAP level

Hgb (vol %)

CHF

128 + 47 105 2 51

115 2 68 105 + 56

160 + 68 211 2 116

10 cm H,O Recovery

122 +- 52 116 + 47

83 2 24 217 2 185

200 2 101 201 k 104

Baseline 5 cm HZ0

52 2 7

50 + 6

49 2 9

53 + 6 5228 55k 11

51 + 5 53 2 7 48 k a

49 2 a 51 + a 51 -ca

Baseline 5 cm Hz0 1Ocm Hz0 Recovery Baseline 5 cm HZ0 10cm Hz0

RR bm-d

HV

Baseline 5 cm H,O

10cmH20 Recovery PH W)

Measurements NV

Recovery Baseline 5 cm H,O 10 cm Hz0 Recovery

Abbreviations: NV, normovolemia;

7.31 + 0.4 7.30 ? .05

733 2 .07

7.31 2 .03 7.30 k .05 7.30 2 .05 7.29 ? .04 7.29 + .08 7.33 2 .05

7.33 2 .06 7.31 +- .05 7.32 r .06

a.3 + 0.7 a.4 ? 0.9

6.0 2 o.a* 6.3 + 0.8

a.4 -t 1.2 a.5 -t 1.3

a.2 + 0.8 a.3 + 0.9

6.6 2 1.1 6.1 k 0.9

a.3 t 1.2 a.4 2 1.2

39.7 + 7.2 34.3 + 1.5 39.7 r 4.7 32.0 + 6.9

43.0 r 9.6 40.5 r 10.2

31.7 k 2.9 30.0 2 12.7 38.7 k 4.7 34.0 + 7.9

26.2 2 6.0 43.0 -t 10.3

CPAP, continuous positive airway pressure; HV, hypervolemia; CHF, congestive heart

failure; Hgb, hemoglobin; RR, respiratory rate. *at baseline value signifies P < .Ol for difference of that volume status from NV (2 way analysis of variance, NewmanKeuls test). differences.

There

were

no significant

between

CPAP

level

between different CPAP levels. Arterial hemoglobin was unchanged with CPAP, but fell as expected with volume infusion by approximately 30%. RR was increased at baseline for all conditions, but there were no significant differences between NV, HV, or CHF, or between CPAP levels. Table 2 shows the results of measurements of hemodynamic variables. This table shows that HV was associated with an increase in HR compared with the other conditions. Both HV and CHF were associated with an increase in Pw and mean PAP compared with NV, but no change in PVR. Although there was no overall trend for any of the variables with CPAP, subgroup analysis showed that with HV and CHF, 5 cm Hz0 CPAP was associated with increased SV for both HV and CHF. There was a further increase in SV with recovery compared with baseline for CHF. Figure 1 shows the changes in CO. HV was associated with an increase in baseline CO compared with NV. There were increases in CO for CPAP 5 cm Hz0 compared with baseline for HV and CHF. After CPAP removal, CO increased further at recovery for CHF. Figure 2

ET AL

shows changes in SVR. For both HV and CHF, there were decreases in SVR at CPAP = 5 cm Hz0 compared with baseline. For CHF, SVR decreased even further on removal of CPAP (recovery) compared with baseline pre-CPAP values. Figure 3 shows PNE values. For NV there were no changes in PNE with CPAP. With both HV and CHF, there was an increase in baseline PNE compared with NV. However, PNE changed differently with CPAP for HV than with CHF. For HV, there was a progressive decrease in PNE values with increasing CPAP. On removal of CPAP, PNE values returned to a value not significantly different from baseline. However, in the CHF animals, PNE values increased significantly at CPAP = 5 cm Hz0 and then returned at the higher CPAP level to a value not significantly different from baseline. After removal of CPAP, mean PNE value fell to a value not significantly different from baseline. DISCUSSION

The principle finding in this study was that we could not find a consistent association between changing SNA as measured by PNE and the change in CO induced by CPAP. Experimental Preparation

As in previous studies, 10~11~15 we observed mild baseline hypercapnia and increased baseline RR. This could be a response to paradoxic central respiratory stimulation caused to fentanyl.* The resulting shallow breathing could have led to an increase in dead space ventilation and arterial Pco2. Because there was no significant change in arterial Pco2 or pH with CPAP, respiratory acidosis could not account for our findings. Similarly, changes in CO could not be attributed to changes in sedation. The sedative was given at a constant infusion rate. Baseline and recovery hemodynamic and blood gas values were the same, suggesting no lightening of sedation with time, and no behavioral changes were observed in the animals suggesting change in sedative level. One difference between the present and our previous studies1°J5 is that in the normal pigs there was no prior instrumentation. Previously, HV produced no increase in baseline CO, whereas in the present study there was an increase in baseline CO with HV. We had

EFFECTS

OF CONTINUOUS

POSITIVE

AIRWAY

PRESSURE

Table Variable

CPAP level

MAP (mmHg)

HR (bpm)

SV (mL)

Pw (mmHg)

PAP (mmHg)

PVR (dyne-set-cm-5)

61

2. Hemodynamics

Measurements

NV

HV

Base 5 cm H,O

72.4 2 10.6 81.4 r 12.2

10 cm HZ0 Recovery

79.6 + 7.6 85.2 + 15.6

Base 5 cm Hz0 10cm H,O

97 * 37 942 18 99 + 20

CHF

94.8 2 18.3 87.9 + 11.1 85.6 + 13.1

74.3 t 15.5 80.2 t 14.3 77.9 t 10.2

77.8 + 13.6 118 + 19*

78.5 t 13.9 103 t 30

113 + 17 120 2 21 116? 13

102 + 35 101 2 25 109 + 32

28.7 + 17.4 33.4 rt 16.9’

24.0 zk 6.6 27.4 k 6.8*

25.6 c 14.0 29.9 -c 15.1

27.9 2 12.1 29.5 + 10.41

20.9 + 1.6’ 20.3 k 1.1 21.1 + .5

18.5 + 1.9* 19.6 + 2.9 21 _f 3.2

Recovery Base

97 k 22 28.6 -c 15.5

5 cm H,O 10 cm Hz0 Recovery

27.8 k 12.1 33.1 + 18.6 28.5 k 16.6

Base 5 cm H,O

10.1 + 2.4 11.7 + 2.0

10 cm Hz0 Recovery

11.6 + 2.3 10.8 -c 1.8

23 k 1.7

18.6 k 3.4

Base 5 cm H,O

12.1 ? 24 16.7 ? .2

29.1 AZ 11.7* 29.2 k 4.5

26.1 + 3.6* 27.4 + 5.4

10 cm H,O Recovery

16.6 ? 2.3 14.8 k 1.8

Base 5 cm H,O 10cm H,O

304 2 162 443 ? 140 326 t 202

29.8 + 7.4 28.2 + 3.9 265 2 124

29.9 f 3.5* 27.9 t- 4.0 285 + 124

Recovery

338 + 221

149 k 113 201 + 212 214 2 94

260 xk 153 312 + 222 257 ‘- 116

Abbreviations: MAP, mean arterial pressure; HR, heart rate; pulmonary arterial pressure; PVR, pulmonary vascular resistance;

P”,,,

PWV

NS

NS

<.05

NS

NS

NS

< .OOOl

NS

i .OOOl

NS

NS

NS

SV, stroke volume; Pw, pulmonary wedge pressure; PAP, mean CPAP, continuous positive airway pressure; NV, normovolemia; HV,

hypervolemia; CHF, congestive heart failure; ANOVA, analysis of variance. The last two columns show the P values for the two factor ANOVA, one factor being volume status (Pstat), the other being CPAP level (Pcpap). *after the baseline values signifies P < .05 for difference of the volume state from NV (Z-way ANOVA, Newman-Keuls test). tat any

P < .05 for difference

CPAP level signifies

from

baseline

for that volume

condition

(l-way

repeated

measures

ANOVA,

Newman-Keuls

test)

I

CPAP

m D IB%

CPAP 5 CPAP 10 RECOV

0 * + I

i WB ISSJ EZZB

T 35of

CPAP 0 CPAP 5 CPAP 10 RECOV

3ooc “E y Y 0 e g

25oc

2000

z 1500

NV

HV

CHF

STATUS Fig 1. Effect of CPAP and volume status on CO with NV, HV, and CHF. Recov, recovery. l =P =z .05 compared with CPAP 0 [one-way ANOVA for repeated measures, Newman-Keuls test). + on the CPAP 0 bar = P < .05 compared with normovolemia (2 way ANOVA, Newman-Keuls test),

1000 NV

HV

CHF

STATUS Fig 2. Fig 1.

Effect

of CPAP and volume

status

on SVR. Key as in

SCHARF

62

800

I m

CPAP 0 CPAP 5

700

ES! B

CPAP 10 RECOV

from adrenergic nerve endings and only a small fraction of that released into the synaptic cleft gains access to the circulation. However, the prompt changes in PNE in response to changes in CPAP suggests that PNE levels were in fact representative of SNA. PNE levels are also influenced by rate of biochemical breakdown and renal excretion. Theoretically, increased CO with low-level CPAP could lead to increased renal perfusion and excretion. This mechanism would tend to decrease PNE whenever CO increased. In our study, PNE had no relationship to CO. Thus, we believe that PNE levels represented overall SNA with CPAP and were not the secondary results of other factors such as changes in metabolic rate, accumulation, or changes in renal excretion.

600

NV

I HV

CHF

STATUS Fig 3. asinFig1.

Effect

of CPAP and volume

status

on PNE levels.

ET AL

Key

suggested lo that the trauma of cardiac instrumentation for later data recording could have led to pericardial constriction, thus limiting the increase in CO with HV. However, the present study shows that the CO response to CPAP does not depend on prior surgery. CHF produced baseline changes consistent with previously described changes in both animal and clinical studies of CHF.6J1J3s29These include the increased Pw6,13,29and PNE levels29 observed here. Of interest is the increase in baseline PNE measured in HV. Increased baseline SNA in HV was also suggested by increased baseline HR. In humans, central HV produces either no change or a decrease in PNE levels.8,23 However, in our animals, HV was severe enough to lead to Pw levels associated with pulmonary congestion and edema. Possibly, these changes led to increased baseline PNE levels on a reflex basis. On the other hand, decreased hemoglobin associated with volume infusion could also have contributed to increased SNA. HR was not increased with CHF at baseline even though PNE was elevated relative to NV. This was likely the result of the downregulation of cardiac p-receptors associated with chronic CHF.9,16J7,31 We have used PNE as an index of overall SNA. This is consistent with the generally accepted notion that PNE reflects overall SNA.7J8x30Norepinephrine is released primarily

Cardiac Output and PNE Response to CPAP

It must be concluded that PNE is not a reliable predictor of the CO response to CPAP under all conditions. This suggests that changes in total body integrated SNA activity are not responsible per se for the increase in CO with low-level CPAP. Our data on the increase in PNE with CPAP in CHF are consistent with our earlier findings that in CHF, but not with HV there was an increase in myocardial contractility with CPAP.lOJr Thus, increased myocardial contractility may be one of the mechanisms contributing to increased CO with CPAP in CHF. This mechanism may work additively with others (discussed below) to increase CO in CHF but is not likely to be the sole mechanism responsible. Our findings on the changes in PNE with CPAP in the CHF animals were different than those in humans with CHF given CPAP by Naughton et al. 22 These investigators observed decreases in venous PNE during CPAP administration. However, obvious differences between species, chronicity and severity of CHF, and concomitant drug administration (sedation in the present study, cardiac medications in the Naughton study), were all uncontrolled variables, which might have rendered the results different. The fact that PNE may either increase or decrease with CPAP-associated increases in CO does not mean that changes in the autonomic nervous system have nothing to do with the regulation of CO with CPAP. PNE, although representing overall SNA, may not adequately

EFFECTS

OF CONTINUOUS

POSITIVE

AIRWAY

PRESSURE

represent local changes in sympathetic tone in specific vascular beds. For example, in the study of Naughton et al, 22PNE decreased with CPAP, but peroneal SNA did not change. In the present study, whenever CO increased with CPAP, SVR decreased. It is reasonable to postulate that decreased SVR was on the basis of reflex vasodilation with CPAP, which in turn reduced LV afterload and led to increased CO with HV and CHF. Vasodilation could have resulted from withdrawal of sympathetic tone, enhancement of parasympathetic tone, or nonsympathetic nonadrenergic vasodilatory mechanisms. Of course, vasodilation could have been baroreceptor mediated and thus the result of rather than the cause of increased CO with CPAP. Finally, it has been postulated that both sympathetic antagonists and agonists can improve cardiac function in heart failure.9 This is because in CHF cardiac sympathetic innervation can be heterogeneous. This can have profound effects on the temporal coordination of myocardial contraction and the duration and configuration of the cardiac action potential. Any therapy that makes sympathetic responses to the heart more uniform could improve cardiac function. Perhaps this explains why both increased (CHF) and decreased (HV) SNA were associated with improved CO. Changes in CO After Removal of CPAP

As in our previous study,” we observed a secondary increase in CO in the CHF animals after CPAP removal. The mechanism for these changes has not been clarified. However, reflex vasodilation associated with removal of CPAP in CHF could be responsible. Alternatively, the situation could be analogous to intermittent infusions of inotropic agents in which cardiac function may remain improved for a time even after removal of infusion of inotropic agents.12y25 If this were the case, then nocturnal administration of CPAP to patients with CHF could be like a nightly infusion of an inotropic agent with continued improvement during the day as reported previously.4 Mechanisms of Improved CO with CPAP

It is likely that increased CO with CPAP is multifactorial, with different factors assuming

63

different degrees of importance under different conditions. For example, increased SNA may be a contributing factor to increased CO in this model of CHF but not with HV. As discussed above, arterial vasodilation with CPAP could be a primary mechanism leading to LV unloading and increased CO with CPAP in any model of cardiac congestion. There is another mechanism whereby increased intrathoracic pressure with CPAP could lead to increased CO. This can best be understood in terms of the coupling of the heart to the peripheral circulation as modeled by Permutt and Wise.25 Normally, LV end-systolic elastance is far greater than end-diastolic elastance. However, with severe cardiac congestion, the heart may reach the limit of diastolic filling, especially if constrained by the pericardium.25 Under these conditions, diastolic elastance may be greater than systolic elastance. If volume were transferred out of the LV during CPAP because of increased intrathoracic pressure, volume could well decrease more at end-systole than at enddiastole. This would be accompanied by an increase in SV. Thus, the observation that CO increases with CPAP in the presence of cardiac congestion is analogous to earlier observations that phlebotomy or placing tourniquets on the thighs is accompanied by decreased SV in normals, but was often associated with increased SV in patients with CHF.14J9 With CPAP 5 cm Hz0 esophageal pressure increases less than 1 cm Hz0.15 It remains to be determined whether this is sufficient to transfer enough volume out of the thorax for this mechanism to account for increased CO. Thus, increased CO with CPAP may not depend on depression of myocardial function, but only on the LV having reached the limits of diastolic filling (severe congestion). It would make no difference whether baseline myocardial function was normal, depressed, or even enhanced. In the present study, CO was increased with CPAP in the HV animals even in the presence of increased baseline CO. According to this hypothesis, cardiac congestion to the limits of ventricular filling is the necessary condition for CPAP to enhance CO. ACKNOWLEDGMENTS The authors would David Wilson for their

like to thank able technical

Michael assistance.

DiBlasi

and

64

SCHARF

ET AL

REFERENCES chemical evidence of cardiac sympathetic activation and increased central nervous system norepinephrine turnover in severe congestive heart failure. J Am Coil Cardiol 23:570-578,1994 17. Kendall MJ, Lynch KP, Hjalmarson A, Kjekshus J: P-blockers and sudden cardiac death. Ann Int Med 123:358367,1995 18. Lake CR, Ziegler MG, Kopin IJ: Use of plasma norepinephrine for evaluation of sympathetic neural function in man. Life Sci 18:1315-1326,1976 19. McMichael J, Sharpey-Schaefer EP: The action of intravenous digoxin in man. Q J Med 13:123,1944 20. Naughton MT, Liu PP, Bernard DC, et al: Treatment of congestive heart failure and Cheyne-Stokes respiration during sleep by continuous positive airway pressure. Am J Respir Crit Care Med 151:92-97,1995 21. Naughton MT, Bernard DC, Liu PP, et al: Effects of nasal CPAP on sympathetic activity in patients with heart failure and central sleep apnea. Am J Respir Crit Care Med 152:473-479,1995 22. Naughton MT, Floras JS, Rahman MA, Bradley TD: Acute sympathoneural response to CPAP in heart failure. Am J Resp Crit Care Med 151:A706,1995 (abstr) 23. Ogihara TJ, Shima H, Hara Y, et al: Changes in plasma atria1 natriuretic polypeptide concentration during head-out water immersion and saline infusion in normal men. Jpn Heart J 28:41-51, 1987 24. Perel A, Williamson DC, Model1 JH: Effectiveness of CPAP by mask for pulmonary edema associated with hypercarbia. Int Care Med 9:17-19,1983 25. Permutt S, Wise RA: The control of cardiac output through coupling of heart and blood vessels, in Yin FCP (ed): Ventricular/Vascular Coupling, New York, NY, Springer-Verlag, 1987, pp 159-179 26. Pinsky MR, Herhocik D, Culpepper JA, Snyder JV: Flow resistance of expiratory positive-pressure systems. Chest 94:788-791,1988 27. Rao PS, Rujikarn N, Luber J Jr: A specific sensitive HPLC method for determination of plasma dopamine. Chromatographia 28:309-312,1989 28. Rasanen J, Heikkila J, Downs J, et al: Continuous positive airway pressure by facemask in cardiogenic pulmonaty edema. Am J Cardiol55:296-300,1985 29. Riegger AJG, Liebau G: The renin-angiotensinaldosterone system, antidiuretic hormone and sympathetic nerve activity in an experimental model of congestive heart failure in the dog. Clin Sci 62:465-469, 1982 30. Robertson D, Johnson GA, Nies AS, et al: Comparative assessment of stimuli that release neuronal and adrenomedullary catecholamines in man. Circulation 59:637-643, 1979 31. Swedberg K, Enroth P, Kjekshus J, Wilhelmsen L: Hormones regulating cardiovascular function in patients with severe CHF and their relation to mortality. Circulation 82:1730-1736,199O

1. Bersten AD, Holt AW, Vedig AE, et al: Treatment of severe cardiogenic pulmonary edema with continuous positive airway pressure delivered by face mask. N Engl J Med 32518251830,199l 2. Booth NH: Neuroleptics, in Booth NH, McDonald E (eds): Veterinary Pharmacology. Ames, IA, Iowa State University, 1977, pp 268-269 3. Bradley TD, Holloway RM, Mclaughlin PR, et al: Cardiac output response to continuous positive airway pressure in congestive heart failure. Am J Respir Crit Care Med 145377-382, 1992 4. Bristow MR, Ginsburg P, Minobe W, et al: Decreased catecholamine sensitivity and B-adrenergic receptor density in failing human hearts. N Engl J Med 307:205-211,1982 5. Chidsey CA, Braunwald E, Morrow AG: Catecholamine excretion and cardiac stores of norepinephrine in congestive heart failure. Am J Med 39:442-451, 1965 6. Chow E, Woodard JC, Farrar DJ: Rapid ventricular pacing in pigs: An experimental model of congestive heart failure. Am J Physiol258:H1603-H1605,1990 7. Cohn JN, Levine TB, Oliveri MT, et al: Plasma norepinephrine as a guide to prognosis in patients with chronic congestive heart failure. N Engl J Med 311:819-823, 1984 8. Convertino VA, Mack GA, Nadel ER: Elevated central venous pressure: A consequence of exercise traininginduced hypervolemia? Am J Physiol260:R273-R277,1991 9. Daly PA, Sole MJ: Myocardial catecholamines and the pathophysiology of heart failure. Circulation 82134-143, 1990 (suppl 1) 10. Genovese J, Moskowitz M, Tarasiuk A, et al: Effects of continuous positive airway pressure on cardiac output in normal and hypervolemic unanesthetized pigs. Am J Respir Crit Care Med 150:752-758,1994 11. Genovese J, Huberfeld SI, Tarasiuk A, et al: Effects of CPAP on cardiac output in pigs with pacing induced congestive heart failure. Am J Resp Crit Care Med 152:18471853,1995 12. Gibelin P, Sbirrazzuoli V, Drici M, et al: Effects of short-term administration of dobutamine on left ventricular performance, exercise capacity, norepinephrine levels and lymphocyte adrenergic receptor density in congestive heart failure. Cardiovasc Drugs Ther 4:1105-1111,199O 13. Hendrick DA, Smith AC, Kratz JM, et al: The pig as a model of tachycardia and dilated cardiomyopathy. Lab Anim Sci 40:495-501,199O 14. Howarth S, McMichael J, Sharpey-Schaefer EP: Effects of venesection in low output heart failure. Clin Sci 6:41-50, 1946 15. Huberfeld SI, Genovese J, Tarasiuk A, Scharf SM: Effect of CPAP on pericardial pressure and respiratory system mechanics in pigs. Am J Respir Crit Care Med 152:142-147, 1995 16. Kaye DM, Lambert GW, Lefkovitz J, et al: Neuro