Am J Physiol Heart Circ Physiol 313: H408–H420, 2017.

First published May 26, 2017; doi:10.1152/ajpheart.00081.2017.

RESEARCH ARTICLE Integrative Cardiovascular Physiology and Pathophysiology

Right atrial pressure and venous return during cardiopulmonary bypass

Per W. Moller,1,2 Bernhard Winkler,3 Samuel Hurni,3 Paul Philipp Heinisch,3 Andreas Bloch,1
Soren Sondergaard,4 Stephan M. Jakob,1 Jukka Takala,1 and David Berger1
1
Department of Intensive Care Medicine, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland;
2
Department of Anaesthesiology and Intensive Care Medicine, Institute of Clinical Sciences at the Sahlgrenska Academy,
University of Gothenburg, Sahlgrenska University Hospital, Gothenburg, Sweden; 3Department of Cardiovascular Surgery,
Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland; and 4Centre of Elective Surgery, Silkeborg
Regional Hospital, Denmark
Submitted 8 February 2017; accepted in final form 22 May 2017

Moller PW, Winkler B, Hurni S, Heinisch PP, Bloch A, Son- is directly measureable after pressure equilibration at zero flow
dergaard S, Jakob SM, Takala J, Berger D. Right atrial pressure and represents the stressed vascular volume at a given vascular
and venous return during cardiopulmonary bypass. Am J Physiol compliance (32). According to Guyton, venous return (VR) is
Heart Circ Physiol 313: H408 –H420, 2017. First published May 26, driven by the pressure difference between MSFP and right
2017; doi:10.1152/ajpheart.00081.2017.—The relevance of right atrial (RA) pressure (RAP) (20). The heart decreases RAP and
atrial pressure (RAP) as the backpressure for venous return (QVR) and
mean systemic filling pressure as upstream pressure is controversial
maintains the stressed volume in the systemic vessels by left
during dynamic changes of circulation. To examine the immediate ventricular stroke work (32). This model of circulation was
response of QVR (sum of caval vein flows) to changes in RAP and originally formulated for steady states, but it has subsequently
pump function, we used a closed-chest, central cannulation, heart been used to interpret dynamic changes in VR and cardiac
bypass porcine preparation (n ⫽ 10) with venoarterial extracorporeal output (16, 47). Pinsky (40) observed good correlation between
membrane oxygenation. Mean systemic filling pressure was deter- MSFP estimated from instantaneous VR curves and MSFP
mined by clamping extracorporeal membrane oxygenation tubing measured during ventricular fibrillation, suggesting that Guy-
with open or closed arteriovenous shunt at euvolemia, volume expan- ton’s model was applicable to dynamic conditions as well.
sion (9.75 ml/kg hydroxyethyl starch), and hypovolemia (bleeding Versprille and Jansen (46) showed that an increase in RAP
19.5 ml/kg after volume expansion). The responses of RAP and QVR preceded the reduction of VR in response to inspiratory hold
were studied using variable pump speed at constant airway pressure and supported Guyton’s concept of RAP as backpressure for
(PAW) and constant pump speed at variable PAW. Within each volume VR (19).
state, the immediate changes in QVR and RAP could be described with
Conversely, others have considered changes in RAP as
a single linear regression, regardless of whether RAP was altered by
pump speed or PAW (r2 ⫽ 0.586 – 0.984). RAP was inversely propor- being a result of changes in flow (28). These contrasting views
tional to pump speed from zero to maximum flow (r2 ⫽ 0.859 – have not been reconciled in a subsequent theoretical debate (2,
0.999). Changing PAW caused immediate, transient, directionally 9, 28). The control of VR is highly relevant in clinical treat-
opposite changes in RAP and QVR (RAP: P ⱕ 0.002 and QVR: P ⱕ ment of circulatory failure, especially in the widespread use of
0.001), where the initial response was proportional to the change in modern circulatory assist devices (24).
QVR driving pressure. Changes in PAW generated volume shifts into We used a closed-chest, heart bypass porcine model with
and out of the right atrium, but their effect on upstream pressure was venoarterial extracorporeal membrane oxygenation (VA-
negligible. Our findings support the concept that RAP acts as back- ECMO) and pulmonary artery ligation. This allowed indepen-
pressure to QVR and that Guyton’s model of circulatory equilibrium dent control of both RA outflow and RAP in the absence of
qualitatively predicts the dynamic response from changing RAP. interaction with the pulmonary circulation, external volume
NEW & NOTEWORTHY Venous return responds immediately to reservoirs, and Starling resistors (2, 9, 31). We hypothesized
changes in right atrial pressure. Concomitant volume shifts within the that RAP acts as backpressure to VR. To test this hypothesis,
systemic circulation due to an imbalance between cardiac output and we studied the effect of acute airway pressure (PAW)-induced
venous return have negligible effects on mean systemic filling pres- changes of RAP on VR and MSFP and the effect of acute
sure. Guyton’s model of circulatory equilibrium can qualitatively changes in ECMO flow (QECMO) on RAP and VR. This
predict the resulting changes in dynamic conditions with right atrial allowed us to evaluate separately the effect of RAP and pump
pressure as backpressure to venous return. function (as modeled by QECMO) on VR and to evaluate
right atrial pressure; mean systemic filling pressure; venous return; whether Guyton’s model of VR is valid during dynamic
cardiac output; extracorporeal membrane oxygenation changes, when the concomitant volume shifts and their effects
on MSFP and RAP are accounted for.
MEAN SYSTEMIC FILLING PRESSURE (MSFP) is the aggregate pres- MATERIALS AND METHODS
sure caused by elastic recoil of the systemic vessels (3, 22). It
The study complied with the Guide for the Care and Use of
Laboratory Animals (1996) and Swiss National guidelines and was
Address for reprint requests and other correspondence: D. Berger, Dept. of approved by the Commission of Animal Experimentation of Canton
Intensive Care Medicine, Inselspital, University Hospital Bern, Bern CH-3010, Bern (Bern, Switzerland) (approval no. BR83/15). Fourteen domestic
Switzerland (e-mail: [email protected]). pigs were fasted for 12 h with free access to water. The first three pigs

H408 0363-6135/17 Copyright © 2017 the American Physiological Society http://www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS H409
were used in pilot studies to establish the instrumentation and feasi- Pulmonary
bility of the study procedures. Eleven pigs were included in the study. artery ligated
As previously described (4), after premedication with intramuscular
ketamine (20 mg/kg) and xylazine (2 mg/kg), induction and intubation
followed with midazolam (0.5 mg/kg). Anesthesia was maintained
with propofol (4 mg·kg⫺1·h⫺1) and fentanyl (5 ␮g·kg⫺1·h⫺1), and the
SVC flow
depth was controlled by repeatedly testing the response to nose pinch. probe
Additional injections of fentanyl (50 ␮g) were given as needed. Arterial
reperfusion
Amiodarone was given as a bolus of 200 mg followed by infusion of cannula
25 mg/h. Cefuroxime (1.5 g) was given at induction and 4 h later.
Muscle relaxation was maintained with rocuronium (100 mg/h) for Venous
the study measurements. The pigs were mechanically ventilated with drainage Position for RAP
a volume-controlled mode (Servo-I, Maquet Critical Care, Solna, cannula measurement

Sweden) using a positive end-expiratory pressure (PEEP) of 5

cmH2O, a fraction of inspired O2 (FIO2) of 0.30, and a tidal volume of IVC flow
probe Aortic flow
7 ml/kg body wt. Before the start of ECMO, the respiratory rate was probe
adjusted to maintain an end-tidal CO2 of 40 mmHg. During ECMO,
tidal ventilation was continued with respiratory rate fixed at 16
breaths/min (inspiration-to-expiration ratio: 1:2). The flow rate of
oxygenator sweep gas (100% O2) was adjusted to keep arterial PCO2
in the normal range (ABL90Flex, Radiometer Medical, Brønshøj,
Denmark). venous clamp ECMO
AV shunt
flow probe
clamp
Installations
arterial clamp
The following catheters were surgically placed: a right carotid
artery catheter, a double lumen catheter with the tip in the RA, a ECMO
three-lumen catheter in the superior vena cava, and a wide-bore
introducer inserted into the external jugular vein. A median sternot-
omy was used to enter the thoracic cavity, and the pericardium was Fig. 1. Extracorporeal membrane oxygenation (ECMO) circuit including
opened. After administration of 5,000 U of heparin, the RA and cannulas, arteriovenous (AV) shunt, and sites for flow probes and right atrial
pressure (RAP) measurement. Items below the dashed line are located outside
ascending aorta were cannulated (right angle metal tip 28-Fr venous
the chest. During “stop flow maneuvers,” venous and arterial clamps were
and wire-reinforced 18-Fr arterial cannulas, Medtronic, Minneapolis, applied as indicated, and the AV shunt clamp can be either open or closed.
MN) and connected to an ECMO circuit (centrifugal pump, nonpul- SVC, superior vena cava; IVC, inferior vena cava.
satile flow, Cardiohelp, HLS Set Advanced 5.0, Maquet, Rastatt,
Germany). Intermittent heparin boluses were used to keep an activated
clotting time of ⬎180 s. Appropriately sized transit time ultrasonic
customized analysis software (Soleasy, Alea Solutions, Zürich, Swit-
flow probes were placed around the intrathoracic parts of the superior
zerland).
and inferior vena cava and descending aorta and on the arterial ECMO
The tip of the catheter for RAP measurement, the venous drainage
tubing (Transonic Systems, Ithaca, NY). The pulmonary artery was
cannula, the inlet port of the ECMO pump, and all pressure transduc-
ligated after adequate QECMO was confirmed. Catheters and cables
ers were fixed to the height of the mid RA and verified by intraoper-
were guided outside the thoracic cavity. Passive pleural drains were
ative palpation and fluoroscopy. Pressures were zeroed against atmo-
placed with the free ends connected to an underwater seal. The
sphere and two-point calibrated using a water manometer. Flow
pericardium, sternum, and wound layers were closed. The urinary
probes were zeroed and calibrated electronically before the study
bladder was drained using cystostomy.
measurements. Baseline drift for pressures and flows was checked at
ECMO Circuit the end of the experiment.

The VA-ECMO circuit had a shunt between the arterial and venous Experimental Protocol
tubing (Fig. 1). Clamping the ECMO inlet and outlet tubing while The protocol consisted of three interventions: 1) stepwise altera-
opening the shunt enabled rapid pressure and volume equilibration. tions of ECMO pump speed (“pump speed maneuvers”), 2) clamping
Fluid Administration of ECMO tubing for the determination of mean systemic filling
pressure at zero flow (“stop flow maneuvers”), and 3) changing RAP
During surgery, Ringer lactate was infused at a rate of 10 by varying PAW against a fixed pump speed (“PAW maneuvers”) (Fig.
ml·kg⫺1·h⫺1 and, thereafter, reduced to 3 ml·kg⫺1·h⫺1. Hydroxyethyl 2). All interventions were performed at three volume states as de-
starch (HES; 6% Voluven, Fresenius Kabi, Bad Homburg, Germany) scribed below.
was supplemented for measured blood loss and added in 50-ml
boluses to allow basal QECMO without RA collapse [425 (247) ml]. Volume States and ECMO Pump Speed
After the chest had been closed, in the euvolemia volume state,
Pressure Measurement and Data Acquisition
QECMO was adjusted during tidal ventilation to achieve a venous O2
Intravascular pressure and PAW were measured using transducers saturation (SVO2) of 50% [lower normal limit for pigs (21)]. If
(xtrans, Codan Medical, Lensahn, Germany) and a modular multipa- necessary, HES boluses were given to allow sufficient QECMO to
rameter monitor (S/5 Critical Care Monitor, Datex-Ohmeda, GE reach the SVO2 target and to allow inspiratory hold without RA
Healthcare, Helsinki, Finland), including ECG and end-tidal CO2. collapse. This pump speed resulted in baseline QECMO (baseline rpm)
Output from pressure transducers and ultrasonic blood flow probes in euvolemia. Changing pump speed results in linear changes in
were recorded at 100 Hz using a data-acquisition system (LabView, QECMO under the study conditions. This was verified by comparing
National Instruments, Austin, TX) and were processed offline using a the measured QECMO to pump speed [r2 ⫽ 0.999 (0.965– 0.9999)].

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
H410 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS

Euvolemia Volume Expansion Hypovolemia

Fig. 2. Abbreviated experimental protocol. For Pump speed maneuvers Pump speed maneuvers Pump speed maneuvers
each volume state (order not randomized to mini-
mize reflex impact of hypovolemia), inspiratory Airway pressure maneuvers Airway pressure maneuvers Airway pressure maneuvers
and expiratory pump speed maneuvers were per- Stop flow maneuvers Stop flow maneuvers Stop flow maneuvers
formed in randomized order. Airway pressure
(PAW) maneuvers preceded stop flow maneuvers.
HES, hydroxyethyl starch.
Infusion of HES 9.75 Bleeding up to 19.5
mL/kg mL/kg

After infusion of 9.75 ml/kg HES (volume expansion), QECMO was reflex-mediated increase in arterial blood pressure (ABP) were seen.
again adjusted during tidal ventilation to reach the SVO2 target of 50% MSFP was taken as the mean value of RAP during three beats of
and to allow inspiratory hold without RA collapse to determine the equilibrium defined from ABP nadir (4). At least 3 min were allowed
baseline flow after volume expansion. After bleeding up to 19.5 ml/kg for heart rate and blood pressure to return to baseline.
or until mean arterial pressure ⫽ 35 mmHg (hypovolemia), QECMO PAW maneuvers. ECMO speed was set to 75% of maximum (75%
was reduced during tidal ventilation to allow inspiratory hold without rpm) and an expiratory hold was done at a PEEP of 5 cmH2O for 10 s.
RA collapse to determine the baseline flow in hypovolemia. In each A tidal inspiration followed by an inspiratory hold for 10 s was then
of the volume states, an attempt was made to further increase performed (inspiratory hold maneuver; Fig. 4). Tidal breathing at
QECMO from baseline until RA collapse started to occur during baseline ECMO speed was resumed for 1 min. The next maneuver
inspiratory hold. The pump speed resulting in this QECMO was consisted of 10 s of inspiratory hold (also at 75% rpm) followed by
considered as maximum (maximum rpm) for each condition and expiration to zero PEEP and expiratory hold for 10 s (zero PEEP
used for reference.
maneuver; Fig. 4). Mean values for RAP and flows were selected for
Pump speed maneuvers. The maximum ECMO pump speed and
the three last beats preceding the holds (“pre”), for the beat displaying
pump speeds of 75% and 50% thereof were used during inspiratory
and expiratory holds. The order of inspiratory versus expiratory holds maximum caval flow change (“⌬QVCmax”), for the three following
was randomized for each animal. Data were extracted for three beats beats (“early”), and for the three last beats before tidal breathing and
9 s into each hold after flow had reached its new steady state (Fig. 3) baseline ECMO speed were resumed (“late”) (Fig. 4, bottom). With
(4). Maneuvers lasted for 30 s, after which pump speed was reset to constant ECMO pump speed, PAW changes would lead to a transient
baseline rpm and tidal ventilation for at least 1 min until heart rate and imbalance between inflow and outflow from the RA (46), associated
blood pressure returned to baseline. with volume shifts between the RA and the systemic vascular com-
Stop flow maneuvers. MSFP was measured by clamping the ECMO partment. We estimated the acute intrasystemic volume shift by
circuit either in expiratory hold, inspiratory hold, or with ongoing tidal integrating the immediate inflow to the RA during the four beats of
ventilation (order randomized) and the shunt either open or closed “⌬QVCmax” and “early” from the measured VR flow (QVR), assum-
(Fig. 3). Flow was resumed after 30 s or earlier if signs of a ing a change in QECMO equal to zero.

Pump speed maneuver Internal order randomized:

Inspiratory hold
20 10000 Expiratoryhold
PAW
Fig. 3. Top left: pump speed maneuvers. For QECMO
7500
definitions of different pump speeds, see 10
METHODS. In brief, RAP was extracted in mid-
expiration (purple bar) or midinspiration at 5000
0
various pump speeds in randomized order.
Bottom left: stop flow maneuvers. Clamping 2500
the ECMO tubing with opened shunt (black
line) led to rapid pressure and volume equili- 0
bration. Arterial blood pressure (ABP) ap- mmHg mL/min
proached RAP asymptotically. Mean systemic 5s

filling pressure (MSFP) was taken as mean

RAP during three beats of the ABP nadir Stop flow maneuver Stop flow maneuver
(purple bar). These stop flow maneuvers were
performed at all volume states, in inspiratory 80 Exp/Insp hold
PAW
and expiratory hold and tidal ventilation in RAP 50% rpm
randomized order, and with open or closed ABP
Exp/Insp hold
ECMO shunt (fixed sequence). The RAP-ro-
tations per minute (rpm) data pairs from pump 75% rpm
40
speed and stop flow maneuvers were com- Exp/Insp hold
bined (Fig. 3, top right, and Fig. 5, A and B) Baseline rpm
for a linear regression of RAP versus rpm. Exp/Insp hold

0 Maximum rpm
mmHg 5s Exp/Insp hold

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS H411
Inspiratory hold maneuver Zero PEEP maneuver
20 10,000

7,500
10

5,000 Fig. 4. Left: inspiratory hold maneuver. Right:

0 zero positive end-expiratory pressure (PEEP)
maneuver. ECMO speed was 75% throughout
2,500 both maneuvers. For the inspiratory hold ma-
neuver, after 10 s of expiratory hold at a
0 PEEP of 5 cmH2O, the ventilator was set to
mmHg
mL/min inspiratory hold. Mean values were selected
5s for the three last beats preceding inspiration
late late (“pre,” blue shade), for the beat in which
maximum change in caval flow was seen
PAW (mmHg) QIVC (mL/min) (“⌬QVCmax,” red shade), for the three fol-
RAP (mmHg) QSVC (mL/min) lowing beats (“early,” purple shade), and for
ECG (mV) QAorta (mL/min) the three last beats of the hold (“late,” green
20 10,000 shade), after which tidal breathing was re-
sumed and pump speed reset. For the zero
15
PEEP maneuver, after 10 s of inspiratory
7,500 hold, the ventilator circuit was disconnected
10
(zero PEEP). Mean values were selected as
5 for the inspiratory hold maneuver with the
5,000 exception of “pre” representing the three last
0 beats preceding zero PEEP.

2,500

0
mmHg mL/min
pre ΔQVCmax early pre ΔQVCmax early 1s

Tidal Ventilation and Combining the Results Into Individual To analyze the effect of control variable on the VR function, we
VR Plots used a generalized estimating equation. Data were grouped according
to the experimentally controlled variable as 1) QVR-RAP pairs from
In all volume states, beat-to-beat extraction of QVR-RAP data pairs pump speed and stop flow maneuvers, representing a changing pump
was performed for three consecutive respiratory cycles during tidal effect, and 2) QVR-RAP pairs obtained during tidal ventilation and the
ventilation at baseline rpm. One breath, including at least two beats in immediate effects of PAW maneuvers, representing changing RAP.
inspiration and expiration, was integrated into a common VR plot (one VR plots combining all animals in each volume state (dependent
for each volume state and animal), including tidal ventilation, pump variable QVR, covariate RAP; exchangeable variance structure) were
speed, and stop flow maneuvers (in expiration and inspiration) and the constructed and analyzed for the interaction of choice of control
immediate effects (“pre” and “⌬QVCmax”) of PAW maneuvers. variable (pump speed vs. PAW) on the VR function. P values of ⬍
Statistical Analysis and Calculations 0.05 were considered significant. Systemic vascular compliance
(Cvasc) was calculated as the inverse of the slope of the linear
To detect a change in MSFP and venous return driving pressure regression line for MSFP (expiration and open shunt) and blood
(VRdP) induced by changes in RAP, we assumed a SD of the volume changes.
difference of 0.7 mmHg based on previous data on the effects of PEEP The time delay between changes in RAP and QVR (sum of inferior
(4). A sample size of n ⫽ 9 would allow one to detect a significant vena cava flow and superior vena cava flow) was characterized by
change in these variables with a power of 80%. Data were analyzed cross correlation. Cross correlation of RAP and QVR during PAW
using SPSS software (version 21, SPSS, Chicago IL). Normality was maneuvers was done from beginning of “pre” to the end of “early.”
assessed using the Shapiro-Wilk test. The effects of volume state on All animals were analyzed individually using customized analysis
MSFP and the effects of changing PAW were analyzed using one-way software (Soleasy, Alea Solutions, Zürich, Switzerland). Signal prop-
repeated-measures ANOVA or a Friedman test, with the Bonferroni agation delay of pressure versus flow in the measurement setup [41
correction for multiple comparisons. The effects of respiratory cycle (7.5) ms, as quantified at 1,000-Hz sampling] was corrected by
and shunt state on MFSP were analyzed by repeated-measures shifting the RAP signal, sampled at 100 Hz, four samples earlier. To
ANOVA (within-subject factors: respiratory cycle and shunt). Pres- focus on the immediate waveforms of RAP and QVR, these signals
sures and flows were measured as mean values over the selected beats. were individually normalized in amplitude from ⫺1 to ⫹1. For the
Results are reported as means (SD) or medians (range) as appropriate. summary plots (Fig. 7), the time axis was also normalized to allow
Linear regressions were done for RAP versus rpm data pairs for each averaging of all animals.
animal and volume state. For each animal and volume state, linear
regressions were constructed of QVR-RAP data pairs from pump RESULTS
speed and stop flow maneuvers (in expiration and inspiration) and the
immediate effects of PAW maneuvers (“pre” and “⌬QVCmax”) and One animal died prematurely, and the autopsy showed
one breath of tidal ventilation (including at least two beats in inspi- complex heart malformations. In total, 10 animals were in-
ration and expiration). cluded [body weight: 42 (2) kg]. For all animals, in hypovo-

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
H412 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS

Table 1. Baseline hemodynamics

n Euvolemia Volume Expansion Hypovolemia P Value

rpm 10 3,095 (175) b

3,224 (247) 2,550 (153)d
⬍0.0005
ECMO flow, ml/min 10 3,497 (76) 3,782 (160)b 2,605 (84)d ⬍0.0005
QVR, ml/min 10 3,126 (412) 3,329 (654)b 2,307 (297)d ⬍0.0005
Venous O2 saturation, % 9 52 (4) 55 (6)b 38 (6)d ⬍0.0005
RAP, mmHg† 10 4.5 (1.8⫺6.0) 5.7 (3.6⫺8.0)b 2.4 (⫺1.6 to 4.6) ⬍0.0005†
Lactate, mmol/l 8 2.6 (0.8) 3.5 (1.2) 3.9 (2.1) 0.042
Mean arterial pressure, mmHg† 10 60 (51⫺74) 60 (53⫺73)b 45 (36⫺57)d ⬍0.0005†
Heart rate, beats/min 10 93 (15) 96 (19) 107 (36) NS
MSFP, mmHg 9 10.3 (0.5)a 12.3 (1.2)b 8.7 (1.0)d ⬍0.0005
Venous return driving pressure, mmHg† 9 5.2 (4.9⫺9.1) 6.6 (4.8⫺9.5) 5.9 (4.7⫺10.7) NS†
Resistance to venous return, mmHg/l† 9 1.9 (1.3⫺2.9) 1.9 (1.3⫺2.8)c 2.5 (2.0⫺5.6) 0.013†
Data are means (SD) or medians (range) if not normally distributed. Data were averaged over three heart beats in expiratory hold at a positive end-expiratory
pressure of 5 cmH2O and baseline pump speed. Venous O2 saturation and lactate were sampled at baseline pump speed at the beginning of each volume state.
rpm, Revolutions per min; RAP, right atrial pressure; QVR, venous return flow (sum of caval vein flows); MSFP, mean systemic filling pressure. P values were
determined by one-way repeated-measures ANOVA for effect of volume state or a Friedman test (†), as appropriate. Post hoc analysis with the Bonferroni
correction is indicated as follows: aeuvolemia vs. volume expansion (P ⫽ 0.002), bvolume expansion vs. hypovolemia (P ⱕ 0.001; cfor venous return driving
pressure, P ⫽ 0.029), and dhypovolemia vs. euvolemia (P ⱕ 0.012). NS, not significant.

lemia, the bleeding was discontinued due to hypotension, and collapse, negative RAP values were observed during offline
the actual bled volume was 12 (9 –17) ml/kg. MSFP and QVR analysis in animals 10 and 11 (animal 11 only in inspira-
increased during volume expansion and decreased in hypovo- tion), suggesting caval collapse upstream of the ECMO
lemia (Table 1). The differences between QECMO and QVR cannula (4).
were small and attributable to coronary sinus flow.
Stop Flow Maneuvers Tidal Ventilation and VR

After the ECMO circuit was clamped, ABP approached The relationship between QVR and RAP, including the open
RAP asymptotically (Fig. 3). In euvolemia, expiration and shunt MSFP at expiration, was highly linear [total 90 breaths,
open shunt, pressure equilibrium was reached in 18 (5) s. r2 ⫽ 0.979 (0.595– 0.9997)].
MSFP was higher with open shunt versus closed shunt. In-
spiratory hold increased MSFP compared with expiratory hold Pressure-Flow Transients in Response to PAW Maneuvers
only in euvolemia (Table 2). Cvasc was 183 (75) ml/mmHg or
4.3 (1.8) ml·kg⫺1·mmHg⫺1 [n ⫽ 9, r2 ⫽ 0.915 (0.472– 0.999)]. Nine animals had complete sets for all volume states (Fig. 6
and Tables 3 and 4). Changing PAW resulted in a change in
Pump Speed Maneuvers RAP in the same direction. However, during the hold maneu-
vers, RAP returned toward its “pre” values earlier than did
The relation between RAP and pump speed was highly
linear, independent of volume state and breath cycle [euvol- PAW. Changing RAP by means of PAW caused immediate,
emia expiration: r2 ⫽ 0.978 (0.894 – 0.999), inspiration: r2 ⫽ transient, directionally opposite changes in QVR (Fig. 6). Dur-
0.953 (0.811– 0.996), volume expansion expiration: r2 ⫽ 0.987 ing the inspiratory hold maneuver, QVR partially recovered
(0.890 – 0.995), inspiration: r2 ⫽ 0.931 (0.859 – 0.981), hypo- already during the three beats (“early”) after the maximum
volemia expiration: r2 ⫽ 0.969 (0.859 – 0.999), and inspiration: decrease in QVC, whereas RAP still remained elevated. This
r2 ⫽ 0.908 (0.760 – 0.988); Fig. 5, A and B]. Despite setting pattern was reversed during the zero PEEP maneuver.
the maximum ECMO pump speed as the highest without RA The change in RAP evoked by PAW maneuvers was associ-
ated with a transient imbalance between QVR and cardiac
output at the time of maximum change (“⌬QVCmax” in Fig. 4).
Table 2. MSFP
QVR decreased during inspiratory hold but increased during
MSFP, mmHg P Value zero PEEP maneuvers. The respective aortic flow was un-
Respiratory changed or even slightly increased (inspiratory hold in euvol-
n Open shunt Closed shunt Cycle Shunt state Interaction emia). During the first four beats (“⌬QVCmax” and “early”),
Euvolemia 10 the median volume shifted upstream was 9 (⫺2 to 35) ml in
Expiration 10.0 (1.1) 9.0 (1.0) 0.005 0.002 NS
Inspiration 10.4 (1.0) 9.4 (0.9) inspiratory hold, and the median volume shifted downstream in
Volume expansion 10 expiratory hold at zero PEEP was 12 (4 – 68) ml.
Expiration 11.9 (1.6) 10.8 (1.7) NS ⬍0.0005 NS The cross correlations (n ⫽ 9) revealed a peak negative
Inspiration 12.2 (1.7) 10.8 (1.7)
Hypovolemia 9 correlation between RAP and QVR close to zero time lag.
Expiration 8.7 (1.0) 7.9 (0.9) NS 0.004 NS Changes in RAP slightly preceded changes in QVR in the
Inspiration 8.8 (1.2) 8.4 (0.9)
inspiratory hold maneuvers [euvolemia: 3.2 (2.7), volume
Data are means (SD); n ⫽ 9 for hypovolemia as one animal lacked a valid expansion: 1.7 (3.4), and hypovolemia: 1.2 (3.2) samples] and
MSFP expiration open shunt. MSFP was taken as the mean value of RAP
during three beats of equilibrium defined from the arterial blood pressure nadir.
in the zero PEEP maneuvers [euvolemia: 3.8 (2.4), volume
P values were determined by repeated-measures ANOVA (within-subject expansion: 1.7 (2.9), and hypovolemia: 2.4 (3.2) samples].
factors respiratory cycle and shunt state) with the Bonferroni correction. Summary plots are shown in Fig. 7.

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS H413

A 1 3 4 5
Euvol r2 0.979 Euvol r2 0.947 Euvol r2 0.964 Euvol r2 0.999
2 2
VE r 0.949 VE r 0.978 2
RAP (mmHg)

2
VE r 0.986 VE r 0.992
Hypo r2 0.946 Hypo r2 0.999 Hypo r2 0.993 Hypo r2 0.925

6 7 8 9
Euvol r2 0.998 Euvol r2 0.894 Euvol r2 0.982 Euvol r2 0.965
RAP (mmHg)

2
2
VE r 0.995 VE r 0.960 2
VE r 0.988 2
VE r 0.989
Hypo r2 0.951 Hypo* r2 0.859 Hypo r2 0.997 Hypo r2 0.987
*MSFP missing, 4-pt regr.

10 11
Euvol r2 0.978 Euvol r2 0.994
2 2
VE r 0.890 VE r 0.991
Hypo r2 0.989 Hypo r2 0.859
RAP (mmHg)

Euvolemia (Euvol)
Volume Expansion (VE)
Hypovolemia (Hypo)

rpm (1/min)

rpm (1/min)
B
1 Euvol r2 0.901 3 Euvol r2 0.987 4 Euvol r2 0.954 5 Euvol r2 0.983
VE r2 0.933 VE r2 0.929 VE r2 0.970 VE r2 0.979
Hypo r2 0.880 Hypo r2 0.988 Hypo r2 0.988 Hypo r2 0.863
RAP (mmHg)

6 Euvol r2 0.958 7 Euvol r2 0.843 8 Euvol r2 0.907 9 Euvol r2 0.952

VE r2 0.893 VE r2 0.952 VE r2 0.859 VE r2 0.981
RAP (mmHg)

Hypo r2 0.935 Hypo r2 0.938 Hypo r2 0.760 Hypo r2 0.983

10 Euvol r2 0.996 11 Euvol r2 0.811

VE r2 0.892 VE r2 0.895
Hypo r2 0.809 Hypo r2 0.811
RAP (mmHg)

+ Euvolemia (Euvol)
+
Volume Expansion (VE)
+ Hypovolemia (Hypo)

rpm (1/min) rpm (1/min)

Fig. 5. A: individual plots of RAP-rpm data pairs from pump speed maneuvers (at baseline rpm, maximum rpm, 75% rpm, and 50% rpm) and MSFP with open
shunt from stop flow maneuvers, all taken during expiration. B: corresponding maneuvers in inspiration.

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
H414 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS

Inspiratory hold maneuver Zero PEEP maneuver

RAP (mmHg)
QVR (mL/min)
Euvolemia

Fig. 6. Behavior of RAP () and venous

return (o) during inspiratory hold (left) and RAP (mmHg)
QVR (mL/min)

zero PEEP maneuvers (right). P values rep-

resent within-subject effects of one-way re-
peated-measures ANOVA for PAW. †Fried-
man test for nonnormal distribution. *P ⬍
0.05 for post hoc pairwise comparison to
“pre.” QVR, venous return flow.
Volume Expansion

RAP
QVR
RAP (mmHg)
QVR (mL/min)

Hypovolemia
pre ΔQVCmax early late pre ΔQVCmax early late

Combining Results Into Individual VR Plots sures changed. These findings are consistent with RAP acting
2 as backpressure to VR. They imply that VR driving pressure is
The r values for euvolemia, volume expansion, and hypo- itself dynamically changing in the transit between steady
volemia were 0.821 (0.614 – 0.984), 0.857 (0.748 – 0.950), and states, mainly due to changes in RAP.
0.857 (0.586 – 0.961), respectively, indicating an overall linear The linear decrease of RAP with increasing pump speed
relation between QVR and RAP (Fig. 8). over the observed range was a result of volume shift away from
Effect of Control Variable on the VR Function the RA (37). Increased pump speed is equivalent to improved
cardiac function (5, 38). The dynamic behavior of RAP and
Two animals had negative RAP values, suggesting caval QVR during PAW maneuvers further demonstrates the effect of
collapse (4): animal 10 in the volume expansion and hypovo- RAP on QVR. The initial decrease or increase in QVR occurred
lemia segment and animal 11 (only in inspiration) during despite constant aortic flow, as confirmed in three animals with
euvolemia and hypovolemia, and they were excluded from direct measurement of QECMO using a high-resolution flow
analysis in these volume states. The choice of control variable probe (data not shown). Therefore, in the inspiratory hold
(pump speed vs. PAW) had no effect on the QVR-RAP relation maneuvers, QVR was reduced as a consequence of acutely
(generalized estimating equation; Fig. 9 and Table 5). increased RAP. In the zero PEEP maneuvers, QVR was in-
creased as a consequence of acutely decreased RAP. Simulta-
DISCUSSION
neously, volume was shifted out of the RA by the constant
The main findings of this study were that 1) an inverse pump flow or into the RA by increased QVR, respectively. This
relationship between RAP and QVR was present regardless of volume shift is likely to contribute to the partial restoration of
whether RAP was altered by PAW or pump speed; 2) changes RAP despite sustained alteration in PAW. Furthermore, QVR
in QVR in response to changes in RAP were immediate and reacted immediately to changes in RAP, regardless of whether
transient; 3) a perturbation of the system simultaneously initi- these were generated by changes in pump speed or changes in
ated volume shifts and flow restoration; and 4) as a conse- PAW, so that RAP acts as backpressure for VR. The cross
quence of volume shifts, both downstream and upstream pres- correlation also demonstrated an inverse relationship between

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS H415
Table 3. Inspiratory hold maneuver
Pre ⌬QVCmax Early Late P Value

Euvolemia
PAW, mmHg† 4.4 (4.1⫺5.5) 12.4 (8.9⫺15.1)a 13.6 (11.5⫺17.2)a 12.0 (10.6⫺15.5) ⬍0.0005†
RAP, mmHg 5.2 (1.2) 6.1 (1.2)a,c 5.9 (1.2)a,c 5.6 (1.4) 0.001
QVR, ml/min 2,263 (323) 1,635 (272)a,b,c 2,135 (249)c 2,297 (234) ⬍0.0005
QAorta, ml/min 1,758 (191) 1,815 (182)a 1,772 (195) 1,819 (203) 0.014
Volume expansion
PAW, mmHg† 4.3 (3.9⫺5.0) 13 (8.9⫺19.0)a 14.3 (13.0⫺18.7)a 13.1 (12.1⫺16.9) ⬍0.0005†
RAP, mmHg† 6.7 (2.0⫺7.9) 8.0 (3.6⫺9.4)a,c 7.6 (2.7⫺9.2)a 6.9 (1.1⫺9.0) ⬍0.0005†
QVR, ml/min 2,501 (305)c 1,946 (393)a,b,c 2,440 (379)c 2,596 (328)a 0.001
QAorta, ml/min 2,109 (382) 2,096 (366) 2,114 (418) 2,170 (410) NS
Hypovolemia
PAW, mmHg 4.3 (0.4)c 13.8 (1.7)a,b 15.4 (1.5)a,c 13.8 (1.3)a ⬍0.0005
RAP, mmHg† 4.1 (1.0⫺5.9) 5.4 (2.0⫺6.9)a 5.0 (1.6⫺6.8) 5.2 (1.0⫺6.6) 0.002†
QVR, ml/min 1,508 (345) 913 (306)a,b,c 1,385 (306)c 1,517 (276) ⬍0.0005
QAorta, ml/min 1,261 (215) 1,300 (199) 1,269 (202) 1,285 (190) NS
Data are means (SD) or medians (range) if not normally distributed. Data were averaged according to Fig. 4, left; n ⫽ 9. ⌬QVCmax, maximum caval flow
change; QAorta, aortic flow. P values were determined by one-way repeated-measures ANOVA (or †Friedman test if appropriate) for effect of changing airway
pressure (PAW). Post hoc analysis with the Bonferroni correction is indicated as follows: asignificant difference vs. pre, bsignificant difference vs. early, and
c
significant difference vs. late.

RAP and QVR with changes in RAP slightly preceding changes shift. In this study, we excluded the possible confounding
in QVR. All of these findings indicate that RAP acts as an effects of right ventricular function and the pulmonary circu-
independent variable for VR while it is, itself, dependent on lation. Our present findings, therefore, represent the direct
pump function. The backpressure role of RAP constitutes the effects of altered RAP on the QVR and provide further support
physical link between changes in pump or cardiac function and of the role of RAP as the backpressure for QVR.
the VR function and is in line with the earlier findings of Starling (42), Guyton et al. (20), Grodins et al. (17), and
Versprille and Jansen (46) and Pinsky (40) with beating heart Levy (28) all agreed that RAP represents the node of interac-
preparations and the original experiments of Guyton (19, 20). tion between circuit and pump functions, reflecting the oper-
We have recently observed a rightward shift of VR curves ating point of the equilibrated cardiovascular system, as can be
during inspiratory hold maneuvers in an animal model with a seen in Guyton’s graphical analysis (18, 28) or variants based
beating heart (4), and we assumed that RAP acted as backpres- on Levy’s interpretation (9, 28). Transient imbalances between
sure to VR during inspiratory holds (29, 46). We were criti-
VR and cardiac output, for example, induced by volume
cized for regarding the pressure difference between MSFP and
expansion or changes in PAW, as in our experiments (4), move
RAP (the VR driving pressure) as relevant for QVR (3, 6): a
central issue in the ongoing theoretical debate on the validity of this operating point toward a new steady state (28). These
Guyton’s concept of circulation (7–9, 30, 31). The pulmonary volume shifts will act to, at least partially, restore both RAP
circulation, its blood volume, and their respective effects on and driving pressure toward the levels preceding the imbalance
right heart function interact with the changes in the systemic (46) and reflect the hydraulic, physical self-regulating proper-
circulation induced by changes in PAW (11, 34, 39, 40, 45), and ties of the cardiovascular system, unrelated to any active
they may have influenced our previous findings of VR curve homeostatic adaption.

Table 4. Zero positive end-expiratory pressure maneuver

Pre ⌬QVCmax Early Late P Value

Euvolemia
PAW, mmHg† 12.4 (11.0⫺15.7)c 0.7 (0.1⫺4.3) 0.2 (⫺0.1 to 0.3)a 0.1 (⫺0.1 to 0.5)a ⬍0.0005†
RAP, mmHg 5.9 (1.7) 4.6 (1.9)a 4.8 (1.7)a 5.2 (1.7) ⬍0.0005
QVR, ml/min† 2,292 (1,777⫺2,410) 2,876 (2,513⫺4,703)a,c 2,476 (1,895⫺3,115) 2,404 (1,877⫺2,997) ⬍0.0005†
QAorta, ml/min 1,743 (234) 1,772 (265) 1,814 (186) 1,841 (225) NS
Volume expansion
PAW, mmHg† 12.9 (10.6⫺16.2)c 0.3 (⫺0.5 to 7.1) 0.2 (⫺0.3 to 0.4)a 0.1 (⫺0.1 to 0.2)a ⬍0.0005†
RAP, mmHg† 6.8 (1.1⫺8.9) 6.4 (0.1⫺7.2)a,c 5.9 (0.6⫺7.5) 6.7 (1.3⫺8.0) 0.002†
QVR, ml/min 2,563 (338) 3,198 (571)a,b,c 2,668 (391) 2,528 (358) 0.001
QAorta, ml/min 2,110 (363) 2,135 (441) 2,138 (414) 2,147 (421) NS
Hypovolemia
PAW, mmHg† 13.2 (10.8⫺16.5)c 0.5 (⫺0.3 to 2.8) 0.2 (⫺0.1 to 0.3)a 0.0 (⫺0.1 to 0.2)a ⬍0.0005†
RAP, mmHg 4.8 (1.7) 3.5 (1.7)a 3.7 (1.8)a 4.1 (2.1) ⬍0.0005
QVR, ml/min† 1,447 (970⫺1,945) 2,115 (1,491⫺3,500)a,c 1,537 (1,339⫺2,544) 1,459 (981⫺2,253) ⬍0.0005†
QAorta, ml/min 1,255 (203) 1,227 (220) 1,278 (210) 1,293 (215) NS
Data are means (SD) or medians (range) if not normally distributed. Data were averaged according to Fig. 4, right; n ⫽ 9. P values were determined by one-way
repeated-measures ANOVA (or †Friedman test if appropriate) for effect of changing PAW. Post hoc analysis with the Bonferroni correction is indicated as
follows: asignificant difference vs. pre; bsignificant difference vs. early, and csignificant difference vs. late.

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
H416 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS

Inspiratory hold maneuver Zero PEEP maneuver

40 0.8
Euvolemia
QVR and RAP (normalized)

0.6
20
Cross Correlation

0.4

0.2
0
0.0
-20 -0.2

-0.4
-40
-0.6

-60 -0.8
-1000 -500 0 500 1000 -1000 -500 0 500 1000
40 0.8
Volume Expansion
QVR and RAP (normalized)

0.6
20
Cross Correlation

0.4

0.2
0
0.0
-20 -0.2

-0.4
-40
-0.6

-60 -0.8
-1000 -500 0 500 1000 -1000 -500 0 500 1000
40 0.8
Hypovolemia
QVR and RAP (normalized)

0.6
20
Cross Correlation

0.4
RAP
0.2
0 QVR
0.0 Cross Correlation
-20 -0.2

-0.4
-40
-0.6

-60 -0.8
-1000 -500 0 500 1000 -1000 -500 0 500 1000

Time (normalized, samples)

Fig. 7. Cross correlations between RAP and QVR during PAW maneuvers. Time was normalized to 1,000 samples to allow averaging of the animals. RAP and
QVR were individually normalized in amplitude (⫺1 to ⫹1). RAP and QVR were negatively correlated with changes in RAP slightly preceding changes in QVR.
The technical propagation delay of pressure and flow signals was experimentally quantified with a roller pump and closed tube circuit with low compliance.
Clamping caused simultaneous changes in flow and pressure. The flow increase as a reference point was visually estimated as a step change in the slope from
baseline. The pressure change was estimated as an increase of more than two SDs from baseline. RAP was corrected for the propagation delay before cross
correlation.

Guyton saw RAP as a cause of altered flow (20), whereas model with a ligated pulmonary artery excludes the possibility
Levy interpreted RAP as an effect of altered flow. The con- that RAP reacts to increases in right ventricular afterload (39),
trolled variable in Guyton’s experiment was the height of a as could be the case with right heart function during PAW
Starling resistor or pump speed. RAP itself could, therefore, changes.
not be an independent variable (9, 28). We excluded these Even those who oppose Guyton’s model (1) agree that
factors in our heart-lung bypass model. In that respect, our VR MSFP at zero flow should represent the elastic recoil pressure
model solely reflects the resistive and elastic properties of the of the systemic vascular stressed volume (3, 33). Stressed
systemic vasculature. Our results demonstrate that, using con- volume is redistributed along vascular beds during flow; nev-
stant pump speed as an equivalent of unchanged cardiac ertheless, it is still present, and so must be its related pressure,
function, VR responds immediately and inversely to changes in MSFP (3, 25, 32). One central question relates to the applica-
RAP caused by altered PAW. In our experiment, the change in bility of Guyton’s VR model in the transit between steady
RAP is the cause and the change in VR is the effect, as RAP states, a condition in which it is often used implicitly (40). A
was altered by an extracirculatory event. This cause-and-effect temporary imbalance between VR and cardiac output implies a
scenario is supported by the cross-correlation analysis. In volume shift, as we observed during the PAW maneuvers. In
addition to Pinsky (40) and Versprille and Jansen (46), our intact circulations, this can occur between regions within the

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS H417

1 Euvol r2 0.937 3 Euvol r2 0.778 4 Euvol r2 0.731 5 Euvol r2 0.984

VE r2 0.918 VE r2 0.807 VE r2 0.748 VE r2 0.925
QVR (mL/min)

Hypo r2 0.933 Hypo r2 0.838 Hypo r2 0.907 Hypo r2 0.819

6 Euvol r2 0.881 7 Euvol r2 0.780 8 Euvol r2 0.861 9 Euvol r2 0.752

VE r2 0.898 VE r2 0.861 VE r2 0.816 VE r2 0.950
QVR (mL/min)

Hypo r2 0.961 Hypo r2 0.835 Hypo r2 0.876 Hypo r2 0.776

10 Euvol r2 0.960 11 Euvolr2 0.614

VE r2 0.854 VEr2 0.777
QVR (mL/min)

Hypo r2 0.889 Hypor2 0.586

Euvolemia (Euvol)
Volume Expansion (VE)
Hypovolemia (Hypo)

RAP (mmHg) RAP (mmHg)

Fig. 8. Individual venous return plots with QVR-RAP data pairs from pump speed maneuvers (in expiration and inspiration), tidal ventilation (all beats in one
breath), and the immediate effects of PAW maneuvers (inspiratory hold and zero PEEP; “pre” and “⌬QVCmax”) and stop flow maneuvers (MSFP; open shunt,
expiration, and inspiration). Euvolemia: r2 0.821 (0.614 – 0.984); volume: expansion: r2 0.857 (0.748 – 0.950); hypovolemia: r2 0.857 (0.586 – 0.961).

systemic compartment (4) as well as between the systemic and reported increased MSFP during inspiratory hold maneuvers
cardiopulmonary compartments (34). In the present experi- and positive pressure tidal ventilation, respectively. Increase of
ment, volume is shifted between upstream and downstream upstream pressure with higher PEEP has been shown by
regions within the systemic compartment, as the cardiopulmo- Fessler et al. (14, 15) and us (4), and it has been attributed to
nary compartment is physically excluded. Increasing PAW will larger stressed volume by Nanas and Magder (35). Additional
move part of the RA volume upstream. proposed mechanisms in intact circulations include a hepatos-
The upstream pressure effect of this shifted volume by the planchnic waterfall (4, 10) or pressurization of the splanchnic
end of the “early” phase of increased PAW was ⬍5% of MSFP bed via downward movement of the diaphragm (44), but the
(data not shown) and of a similar magnitude as the increased results are not consistent (15).
static MSFP during inspiration over expiration in euvolemia
(Table 2). Volume shifts from respiratory maneuvers reported Limitations of the Study
by others and ourselves (4, 11) are small. The large vascular
compliance buffers their effect on MSFP (4, 32). In our PAW After ligation of the pulmonary artery, some of the volume
maneuvers, these shifts restored RAP despite sustained altera- contained in the pulmonary circulation and left heart could
tion in PAW. The predominant mechanism for restoring VRdP have been transferred to the systemic compartment by the
and QVR is, therefore, attributed to the restoration of RAP in contracting ventricles, with MSFP approaching mean circula-
the downstream area, where the RA and adjacent large veins tory filling pressure (MCFP). The bronchial circulation con-
are characterized by a lower compliance and larger pressure tributes around 2% of cardiac output, of which ~10% drains
effect of volume change, again demonstrating the central role into the RA (12, 26). During cardiopulmonary bypass, bron-
of RAP as backpressure to VR. chial perfusion may be reduced to 13% of normal (41). Bron-
Increasing or releasing PAW leads to a rightward or leftward chial drainage to the both atria would, therefore, neither in-
shift of the respective pressure-flow relationship between RAP crease systemic stressed volume nor affect the estimated vol-
and QVR (4, 14, 15, 35), which was reproduced here in the late ume drained from the RA via ECMO. The volume shifted with
phases of the PAW maneuvers and could be only partially tidal breathing is small (11), with small effects on MSFP (4,
explained by shifted volume or direct effects of increased PAW 32). The backpressure role of RAP would not be influenced by
on upstream pressure. In addition to the downstream pressure volume shifts. The changes in MSFP were small. Since metic-
role of RAP, we found, in static measurement conditions, ulous care was taken to exclude technical errors, we believe
increased MSFP during inspiration in euvolemia. Whether PAW these changes are real. The study design does not allow
during tidal ventilation may directly influence MSFP is an evaluation of other mechanisms than volume shifts as the cause
unresolved question. Jellinek et al. (23) and Chihara et al. (13) of altered MSFP. The reported volume shifts, though small,

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
H418
QVR (mL/min) RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS

Euvolemia

Euvolemia; Pump speed maneuver and MSFP

Euvolemia; Tidal ventilation and immediate Airway pressure maneuvers*
Volume Expansion; Pump speed maneuver and MSFP
Volume Expansion; Tidal ventilation and Airway pressure maneuvers*
QVR (mL/min)

Hypovolemia; Pump speed maneuver and MSFP

Hypovolemia; Tidal ventilation and Airway pressure maneuvers*
*Immediate effects ("pre" and "ΔQVCmax") included

Volume Expansion
QVR (mL/min)

Hypovolemia

RAP (mmHg)
Fig. 9. QVR-RAP data pairs from pump speed and stop flow maneuvers (solid symbols), representing changing pump effects, and the corresponding data pairs
obtained during tidal ventilation and the immediate effects of PAW maneuvers (open symbols), representing changing RAP. Using generalized estimating
equations (GEE) (Table 5), interactions of the controlled variable (change in pump speed vs. change in PAW) on the QVR-RAP relation were analyzed for the
volume states. Model parameters of GEE, including 95% confidence intervals, are shown as thick and thin lines, respectively (see also Table 5). Animal 10 was
excluded from the analysis during the volume expansion and hypovolemia segments of the experiment, and animal 11 was excluded from the euvolemia and
hypovolemia segments of the experiment, due to caval collapse.

Table 5. Generalized estimating equations

Dependent Interaction Covariate
Method Covariate Variable ⫻ Method Equation Parameters

rpm, Slope m, ml/mmHg Intercept b, ml

n (%) PAW, n (%) RAP, mmHg QVR, ml/min P value (95% confidence interval) (95% confidence interval) P value

Euvolemia 90 (52) 83 (48) 4.8 (⫺0.8 to 12.1) 2,626 (0⫺5,097) 0.149 ⫺417.4 (⫺475.9 to 4,417 (3,887⫺4,949) ⬍0.0005*
358.9)
Volume Expansion 90 (52) 84 (48) 6.7 (2.8⫺14.6) 2,783 (0⫺5,159) 0.815 ⫺475.4 (⫺539.1 to 5,820 (5,106⫺6,634)
411.6)
Hypovolemia 79 (50) 80 (50) 4.1 (0.1⫺10.6) 2,560 (0⫺3,500) 0.180 ⫺335.4 (⫺280.3 to 3,019 (2,600⫺3,438)
390.4)
Because of the markedly negative RAPs and possible caval collapse, animal 10 was excluded from analysis in volume expansion and hypovolemia, and animal
11 was excluded in euvolemia and hypovolemia. Values for RAP and QVR are given as medians (range). For rpm, data pairs from pump speed and stop flow
maneuvers are shown; for PAW, data pairs from tidal ventilation and the immediate effect of PAW maneuvers (pre and ⌬QVCmax) are shown. *P values for
equations are valid for both slopes and intercepts.

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS H419
were within the resolution of the in situ flow probes. A further REFERENCES
limitation is that high-resolution flow probes for evaluation of 1. Beard DA, Feigl EO. CrossTalk opposing view: Guyton’s venous return
transient changes in ECMO were only available in three curves should not be taught. J Physiol 591: 5795–5797, 2013. doi:10.1113/
animals. Remaining right-left shunts after pulmonary artery jphysiol.2013.260034.
ligation could not be controlled for and may have influenced 2. Beard DA, Feigl EO. Understanding Guyton’s venous return curves. Am
our results. Aortic flows at time of pressure equilibrium in the J Physiol Heart Circ Physiol 301: H629 –H633, 2011. doi:10.1152/
ajpheart.00228.2011.
stop flow maneuvers were small and close to the resolution 3. Berger D, Moller PW, Takala J. Reply to “Letter to the editor: Why
limit of our flow probes. The cross correlations are limited by persist in the fallacy that mean systemic pressure drives venous return?”.
the sampling frequency of 100 Hz. Accordingly, only phase Am J Physiol Heart Circ Physiol 311: H1336 –H1337, 2016. doi:10.1152/
shifts exceeding 10 ms between RAP and QVR could be ajpheart.00622.2016.
detected. 4. Berger D, Moller PW, Weber A, Bloch A, Bloechlinger S, Haenggi M,
We reported a higher compliance compared with models Sondergaard S, Jakob SM, Magder S, Takala J. Effect of PEEP, blood
volume, and inspiratory hold maneuvers on venous return. Am J Physiol
using an intact circulation (4, 36). This may be the result of Heart Circ Physiol 311: H794 –H806, 2016. doi:10.1152/ajpheart.00931.
systemic inflammation and vasoplegia caused by extracorpo- 2015.
real circulation and could influence the pressure effects of our 5. Berlin DA, Bakker J. Starling curves and central venous pressure. Crit
estimated volume shifts (27). Effects of nonpulsatile flow on Care 19: 55, 2015. doi:10.1186/s13054-015-0776-1.
vessel tone are small (43, 48). 6. Brengelmann GL. Letter to the editor: Why persist in the fallacy that
mean systemic pressure drives venous return? Am J Physiol Heart Circ
Conclusions Physiol 311: H1333–H1335, 2016. doi:10.1152/ajpheart.00536.2016.
7. Brengelmann GL. The classical Guyton view that mean systemic pres-
RAP exerts its role as backpressure for VR at any time point, sure, right atrial pressure, and venous resistance govern venous return
yet it represents the operating point of an equilibrated cardio- is/is not correct. J Appl Physiol 101: 1532–1535, 2006. doi:10.1152/
japplphysiol.00698.2006.
vascular system only after volume shifts caused by transient 8. Brengelmann GL. Counterpoint: the classical Guyton view that mean
imbalance between VR and cardiac output have settled. The systemic pressure, right atrial pressure, and venous resistance govern
pressure effects of such imbalances are small and dominated by venous return is not correct. J Appl Physiol 101: 1525–1527, 2006.
RAP as downstream pressure. Guyton’s model can be used to doi:10.1152/japplphysiol.00698a.2006.
qualitatively predict the direction of dynamic transit between 9. Brengelmann GL. A critical analysis of the view that right atrial pressure
determines venous return. J Appl Physiol 94: 849 –859, 2003. doi:10.1152/
steady states. japplphysiol.00868.2002.
ACKNOWLEDGMENTS 10. Brienza N, Ayuse T, O’Donnell CP, Permutt S, Robotham JL. Re-
gional control of venous return: liver blood flow. Am J Respir Crit Care
We thank Olgica Beslac, Michael Lensch, Kay Nettelbeck, and Daniel Med 152: 511–518, 1995. doi:10.1164/ajrccm.152.2.7633700.
Mettler for their expert technical assistance during the experiments and 11. Brower R, Wise RA, Hassapoyannes C, Bromberger-Barnea B, Per-
Hansjörg Jenni for valuable support with the ECMO. Sven Trelle from the mutt S. Effect of lung inflation on lung blood volume and pulmonary
Clinical Trials Unit of the University of Bern offered excellent statistical venous flow. J Appl Physiol 58: 954 –963, 1985.
advice. 12. Charan NB, Thompson WH, Carvalho P. Functional anatomy of bron-
Part of the data in this paper was presented as a poster at the 37th chial veins. Pulm Pharmacol Ther 20: 100 –103, 2007. doi:10.1016/j.pupt.
International Symposium on Intensive Care and Emergency Medicine, March 2006.03.004.
21–24, 2017, in Brussels, Belgium. 13. Chihara E, Hashimoto S, Kinoshita T, Hirose M, Tanaka Y, Mo-
rimoto T. Elevated mean systemic filling pressure due to intermittent
GRANTS
positive-pressure ventilation. Am J Physiol Heart Circ Physiol 262:
This work was supported by Grant 23/2015 from the “Stiftung für For- H1116 –H1121, 1992.
schung in Anaesthesiologie und Intensivmedizin” (Funds for Research in 14. Fessler HE, Brower RG, Wise RA, Permutt S. Effects of positive
Anesthesia and Intensive Care Medicine) in Bern, Switzerland (to D. Berger, end-expiratory pressure on the canine venous return curve. Am Rev Respir
P. W. Moller, and S. Jakob). Dis 146: 4 –10, 1992. doi:10.1164/ajrccm/146.1.4.
15. Fessler HE, Brower RG, Wise RA, Permutt S. Effects of positive
DISCLOSURES end-expiratory pressure on the gradient for venous return. Am Rev Respir
The Department of Intensive Care Medicine, Inselspital, Bern University Dis 143: 19 –24, 1991. doi:10.1164/ajrccm/143.1.19.
Hospital, University of Bern, has or has had research and development/ 16. Funk DJ, Jacobsohn E, Kumar A. Role of the venous return in critical
consulting contracts with Orion Corporation, Abbott Nutrition International, B. illness and shock: part II-shock and mechanical ventilation. Crit Care Med
Braun Medical, Centre Suisse d’Electronique et de Microtechnique SA, Ed- 41: 573–579, 2013. doi:10.1097/CCM.0b013e31827bfc25.
wards Lifesciences Services, Kenta Biotech, Maquet Critical Care, Omnicare 17. Grodins FS, Stuart WH, Veenstra RL. Performance characteristics of
Clinical Research, and Nestlé. Educational grants were received from Frese- the right heart bypass preparation. Am J Physiol 198: 552–560, 1960.
nius Kabi, Glaxo-Smith-Kline, Merck, Sharp and Dome, Eli Lilly, Baxter, 18. Guyton AC. Determination of cardiac output by equating venous return
Astellas, AstraZeneca, B. Braun Medical, Commonwealth Serum Laboratories curves with cardiac response curves. Physiol Rev 35: 123–129, 1955.
Behring, Maquet, Novartis, Covidien, Nycomed, Pierre Fabre Pharma (Roba 19. Guyton AC, Lindsey AW, Abernathy B, Richardson T. Venous return
Pharma), Pfizer, and Orion Pharma. No personal financial benefits resulted at various right atrial pressures and the normal venous return curve. Am J
from the respective development/consulting contracts and/or educational Physiol 189: 609 –615, 1957.
grants. 20. Guyton AC, Lindsey AW, Kaufmann BN. Effect of mean circulatory
filling pressure and other peripheral circulatory factors on cardiac output.
AUTHOR CONTRIBUTIONS Am J Physiol 180: 463–468, 1955.
P.W.M., S.S., S.M.J., J.T., and D.C.B. conceived and designed research; 21. Hannon JP, Bossone CA, Wade CE. Normal physiological values for
P.W.M., B.W., S.H., P.P.H., and D.C.B. performed experiments; P.W.M. and conscious pigs used in biomedical research. Lab Anim Sci 40: 293–298,
D.C.B. analyzed data; P.W.M., J.T., and D.C.B. interpreted results of experi- 1990.
ments; P.W.M. and D.C.B. prepared figures; P.W.M. and D.C.B. drafted 22. Henderson WR, Griesdale DE, Walley KR, Sheel AW. Clinical review:
manuscript; P.W.M., B.W., A.M.B., S.S., S.M.J., J.T., and D.C.B. edited and Guyton—the role of mean circulatory filling pressure and right atrial
revised manuscript; P.W.M., B.W., S.H., P.P.H., A.M.B., S.S., S.M.J., J.T., pressure in controlling cardiac output. Crit Care 14: 243, 2010. doi:10.
and D.C.B. approved final version of manuscript. 1186/cc9247.

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.
H420 RIGHT ATRIAL PRESSURE AND VENOUS RETURN DURING CARDIOPULMONARY BYPASS

23. Jellinek H, Krenn H, Oczenski W, Veit F, Schwarz S, Fitzgerald RD. 36. Ogilvie RI, Zborowska-Sluis D, Tenaschuk B. Measurement of mean
Influence of positive airway pressure on the pressure gradient for venous circulatory filling pressure and vascular compliance in domestic pigs. Am
return in humans. J Appl Physiol 88: 926 –932, 2000. J Physiol Heart Circ Phyiol 258: H1925–H1932, 1990.
24. Kakino T, Saku K, Sakamoto T, Sakamoto K, Akashi T, Ikeda M, Ide 37. Parkin WG. Volume state control—a new approach. Crit Care Resusc 1:
T, Kishi T, Tsutsui H, Sunagawa K. Prediction of hemodynamics under 311–321, 1999.
left ventricular assist device. Am J Physiol Heart Circ Physiol 312: 38. Patterson SW, Starling EH. On the mechanical factors which determine
H80 –H88, 2017. doi:10.1152/ajpheart.00617.2016. the output of the ventricles. J Physiol 48: 357–379, 1914. doi:10.1113/
25. Keener JS. Compliance. In: Mathematical Physiology: Systems Physiol- jphysiol.1914.sp001669.
ogy. edited by Keener J, Sneyd J. New York: Springer, 2009, p. 476 – 479. 39. Pinsky MR. Determinants of pulmonary arterial flow variation during
respiration. J Appl Physiol Respir Environ Exerc Physiol 56: 1237–1245,
doi:10.1007/978-0-387-75847-3.
1984.
26. Kotoulas C, Panagiotou I, Tsipas P, Melachrinou M, Alexopoulos D,
40. Pinsky MR. Instantaneous venous return curves in an intact canine
Dougenis D. Experimental studies in the bronchial circulation. Which is preparation. J Appl Physiol Respir Environ Exerc Physiol 56: 765–771,
the ideal animal model? J Thorac Dis 6: 1506 –1512, 2014. doi:10.3978/ 1984.
j.issn.2072-1439.2014.09.32. 41. Schlensak C, Doenst T, Preusser S, Wunderlich M, Kleinschmidt M,
27. Kristof AS, Magder S. Low systemic vascular resistance state in patients Beyersdorf F. Cardiopulmonary bypass reduction of bronchial blood
undergoing cardiopulmonary bypass. Crit Care Med 27: 1121–1127, flow: a potential mechanism for lung injury in a neonatal pig model. J
1999. doi:10.1097/00003246-199906000-00033. Thorac Cardiovasc Surg 123: 1199 –1205, 2002. doi:10.1067/mtc.2002.
28. Levy MN. The cardiac and vascular factors that determine systemic blood 121977.
flow. Circ Res 44: 739 –747, 1979. doi:10.1161/01.RES.44.6.739. 42. Starling E. The Arris and Bale Lectures on some points in the pathology
29. Maas JJ, Geerts BF, Jansen JR. Evaluation of mean systemic filling of heart disease, lecture II. Lancet 1: 652–655, 1897. doi:10.1016/S0140-
pressure from pulse contour cardiac output and central venous pressure. J 6736(00)49005-7.
Clin Monit Comput 25: 193–201, 2011. doi:10.1007/s10877-011-9294-0. 43. Taenaka Y, Tatsumi E, Nakamura H, Nakatani T, Yagura A, Sekii H,
30. Magder S. The classical Guyton view that mean systemic pressure, right Sasaki E, Akagi H, Goto M, Takano H. Physiologic reactions of awake
atrial pressure, and venous resistance govern venous return is/is not animals to an immediate switch from a pulsatile to nonpulsatile systemic
correct. J Appl Physiol 101: 1533, 2006. doi:10.1152/japplphysiol.00903. circulation. ASAIO Trans 36: M541–M544, 1990.
2006. 44. Takata M, Beloucif S, Shimada M, Robotham JL. Superior and inferior
31. Magder S. Point: the classical Guyton view that mean systemic pressure, vena caval flows during respiration: pathogenesis of Kussmaul’s sign. Am
right atrial pressure, and venous resistance govern venous return is/is not J Physiol Heart Circ Physiol 262: H763–H770, 1992.
correct. J Appl Physiol 101: 1523–1525, 2006. doi:10.1152/japplphysiol. 45. van den Berg PC, Pinsky MR. Systems approach to heart-lung interactions.
Neth J Med 57: 113–131, 2000. doi:10.1016/S0300-2977(00)00062-0.
00698.2006.
46. Versprille A, Jansen JR, Drop A, Hulsmann AR. Mean systemic filling
32. Magder S. Volume and its relationship to cardiac output and venous
pressure as a characteristic pressure for venous return. Pflugers Arch 405:
return. Crit Care 20: 271, 2016. doi:10.1186/s13054-016-1438-7. 226 –233, 1985. doi:10.1007/BF00582565.
33. Magder S, De Varennes B. Clinical death and the measurement of 47. Vieillard-Baron A, Chergui K, Augarde R, Prin S, Page B, Beauchet
stressed vascular volume. Crit Care Med 26: 1061–1064, 1998. doi:10. A, Jardin F. Cyclic changes in arterial pulse during respiratory support
1097/00003246-199806000-00028. revisited by Doppler echocardiography. Am J Respir Crit Care Med 168:
34. Mitzner W, Goldberg H, Lichtenstein S. Effect of thoracic blood 671–676, 2003. doi:10.1164/rccm.200301-135OC.
volume changes on steady state cardiac output. Circ Res 38: 255–261, 48. Voss B, Krane M, Jung C, Brockmann G, Braun S, Günther T, Lange
1976. doi:10.1161/01.RES.38.4.255. R, Bauernschmitt R. Cardiopulmonary bypass with physiological flow
35. Nanas S, Magder S. Adaptations of the peripheral circulation to PEEP. and pressure curves: pulse is unnecessary! Eur J Cardiothorac Surg 37:
Am Rev Respir Dis 146: 688 –693, 1992. doi:10.1164/ajrccm/146.3.688. 223–232, 2010. doi:10.1016/j.ejcts.2009.05.050.

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00081.2017 • www.ajpheart.org

Downloaded from journals.physiology.org/journal/ajpheart (168.196.203.210) on June 8, 2023.