Ventriculo-Arterial Coupling Analysis Predicts the

Hemodynamic Response to Norepinephrine in

Hypotensive Postoperative Patients: A Prospective
Observational Study
Pierre-Grégoire Guinot, MD1,2; Dan Longrois, MD, PhD3; Said Kamel, PhD2; Emmanuel Lorne, MD, PhD1,2;
Hervé Dupont, MD, PhD1,2

Objectives: The objectives of the present study were to evaluate, sure, higher indexed total peripheral resistance, arterial elastance,
in patients with persistent arterial hypotension in the immediate arterial elastance to end-systolic ventricular elastance ratio (2.21
postcardiac surgery period, the effects of norepinephrine infusion [1.69–2.89] vs 1.33 [1.1–1.56]; p < 0.05), and lower arterial
on ventriculo-arterial coupling, its determinants: arterial elastance compliance, indexed total peripheral resistance and cardiac index.
and end-systolic ventricular elastance, and to test the ability of Norepinephrine significantly increased arterial pressure in all sub-
arterial elastance to end-systolic ventricular elastance ratio to pre- jects. In stroke volume responders, norepinephrine increased arte-
dict stroke volume increases. rial elastance, end-systolic ventricular elastance, cardiac index,
Design: Prospective observational study. and improved arterial elastance/end-systolic ventricular elastance
Setting: Cardiac-vascular surgical ICU. coupling. The baseline arterial elastance to end-systolic ventricu-
Patients: Twenty-eight postoperative cardiac surgery patients, in lar elastance ratio predicted stroke volume responsiveness (area
whom physicians decided to administer norepinephrine infusion, under the curve [95% CI], 0.87 [0.71‒1]; p < 0.0001).
were included. Conclusions: In patients with arterial hypotension norepinephrine
Measurements and Main Results: Arterial pressure, stroke volume increased end-systolic ventricular elastance and arterial elas-
index, cardiac index, indexed total peripheral resistance, arterial tance. The effects of norepinephrine on stroke volume depend
compliance, arterial elastance, and end-systolic ventricular elas- on baseline ventriculo-arterial coupling. Although norepinephrine
tance, were measured before and after norepinephrine infusion. infusion corrects arterial hypotension in all subjects, increase of
We estimated ventriculo-arterial coupling by the arterial elastance stroke volume occurred only in subjects with altered ventriculo-
to end-systolic ventricular elastance ratio and defined stroke arterial coupling. (Crit Care Med 2017; XX:00–00)
volume responders by a stroke volume increase greater than or
Key Words: arterial hypotension; norepinephrine; surgical intensive
equal to 15%. Twenty-two of the 28 subjects had altered ven-
care; ventriculo-arterial coupling
triculo-arterial coupling (1.87 [1.57–2.51] vs 1.1 [1–1.18]). Fif-
teen of the 28 subjects (54%) were stroke volume responders.
At baseline, stroke volume responders had similar arterial pres-

B
ecause norepinephrine is the cornerstone of acute cir-
1
Department of Anaesthesiology and Critical Care, Amiens University culatory failure treatment in critically ill patients, an
Hospital, Place Victor Pauchet, Amiens, France. abundant literature has been published on this topic
2
INSERM U1088, Jules Verne University of Picardy, Amiens, France. (1, 2). Actually, most of the literature in humans has evaluated
3
Department of Anaesthesiology and Critical Care, Bichat Claude Bernard norepinephrine and its macrohemodynamic effects accord-
Hospital and INSERM1148, Paris, France.
ing to a “cardio-centric” pressure/outflow model (i.e., Guy-
Dr. Guinot contributed in data acquisition, and Drs. Guinot and Longrois
analysis and interpretation. All authors contributed in drafting of the article ton, Franck Starling) (3–6). This approach has been useful to
for important intellectual content and read and approved the article. provide a complex although not complete understanding of
The authors have disclosed that they do not have any potential conflicts the mechanisms that underlie the effects of norepinephrine,
of interest.
including cardiac surgery patients (4).
For information regarding this article, E-mail: guinotpierregregoire@gmail.
com As a complement to the “cardio-centric” pressure-flow
Copyright © 2017 by the Society of Critical Care Medicine and Wolters model, several authors have applied a model that is based
Kluwer Health, Inc. All Rights Reserved. on an integrated pressure-volume relationship of left ventri-
DOI: 10.1097/CCM.0000000000002772 cle (LV) and vascular system (7–10). This model is based on

Critical Care Medicine www.ccmjournal.org 1

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Guinot et al

measurement of the end-systolic ventricular elastance (Ees) and Inclusion criteria were patients over the age of 18 years, under
arterial elastance (EA) (7–10). Ees is an integrated measure of controlled positive-pressure ventilation and sedated, for whom
LV systolic performance derived from the complex association the physician decided to infuse norepinephrine during the first
between the inotropic efficiency and the functional, structural, 6 hours following admission to the ICU. The indications for
and geometric characteristics of the LV (11). EA describes the norepinephrine were persistent arterial hypotension (systolic
arterial load in an integrated manner that takes into account arterial pressure [SAP] < 100 mm Hg and/or mean arterial
total peripheral resistance (TPR), total arterial compliance pressure < 65 mm Hg) despite fluid therapy (17). Exclusion
(CA), characteristic impedance, and systolic and diastolic time criteria were spontaneous ventilation, permanent arrhythmia,
intervals (11). Ventriculo-arterial coupling is defined by the cardiac conduction disorders, presence of a pacemaker (or
ratio of EA/Ees (10). Ventriculo-arterial coupling model was requirement for use of temporary pacemaker), dobutamine
widely used in cardiology to describe pathophysiology and or epinephrine infusion, poor echogenicity, and right heart
understand the effect of cardiologic therapeutics (11–14). The dysfunction.
clinical relevance of this model is based on the fact that EA/
Ees predicts outcomes independently from other variables in Hemodynamic Variables
patients with cardiovascular diseases (15). Transthoracic echocardiography (CX50 Ultrasound System
At the bedside, the cardiovascular effects of norepineph- and an S5-1 Sector Array Transducer; Philips Medical System,
rine infusion are difficult to predict. Studies demonstrated a Suresnes, France) was performed by a physician blinded to the
systematic increase of arterial pressure through vasoconstric- results of the study. LV ejection fraction (LVEF), end-systolic
tive effects, regardless of the etiology (1–6). Nevertheless these volume (ESV), and end-diastolic volume (EDV) were calcu-
effects were inconsistent on cardiac function and cardiac out- lated using Simpson’s method on a four-chamber view. The
put (CO) (3–6). One study in resuscitated normotensive post- diameter of the LV outflow tract was measured on a long-axis
operative patients demonstrated that norepinephrine increase parasternal view at the time of patient inclusion. The aortic
preload, thus CO (4). In this respect, respiratory stroke volume velocity-time integral (VTIAo), pre-ejection time, and systolic
variation (SVV) was predictive of CO changes (4). time were measured with pulsed Doppler on a five-chamber
A description of the cardiovascular effects of norepineph- apical view. SV (mL) was calculated as VTIAo × Aortic area
rine may improve our understanding of the pathophysiology and was expressed as indexed SV (SVi) = SV/body surface area
of hemodynamic states and help physicians at the bedside to (mL/m2). Cardiac index (CI) (L/min/m2) was calculated as
more reliably predict the effects of norepinephrine and pos- SVi × heart rate (HR). SVV was calculated as SVV = (SVmax
sibly define new therapeutic targets that could improve patient – SVmin)/([(SVmax + SVmin)/2]) × 100. Mean echocardiographic
outcomes. In view of the multiple cardiovascular effects of variables were calculated from the average over five consecu-
norepinephrine, it would be interesting to identify which vari- tive cardiac cycles and analyzed retrospectively.
able is able to predict an increase both in arterial pressure and
stroke volume (SV). Given that the ventricular and arterial LV Ees, EA, and Ventriculo-Arterial Coupling
systems operate simultaneously, ventriculo-arterial coupling Ees was estimated at the bedside using the noninvasive single-
determines the SV and the ejection pressure (i.e., the arterial beat method described by Chen et al (18). This method does
blood pressure). Thus, we hypothesized that ventriculo-arterial not assume that the volume axis intercept of end-systolic pres-
coupling may enable better characterization of the cardiovas- sure-volume relation is at the origin of the diagram (0; 0) but
cular state, and predictability of SV changes with norepineph- can be extrapolated to intersect the volume axis at the point
rine infusion. V0; 0 (19).
The objective of this study was to demonstrate the ability of EA was estimated by using the formula EA = 0.9 × SAP/SV
EA/Ees ratio to predict an increase of SV with norepinephrine (20). SAP was measured by using invasive radial artery cathe-
infusion in hypotensive patients while intubated and sedated ter. At rest, in healthy men and women, mean EA/Ees, EA, and Ees
in the few hours following cardiac surgery. measured invasively are 1.0 ± 0.36, 2.2 ± 0.8 mm Hg/mL, and
2.3 ± 1.0 mm Hg/mL, respectively (21).
We calculated indexed TPR (TPRi) and CA as TPRi = mean
METHODS
arterial pressure (MAP) – central venous pressure (CVP)/CI
Patients (mm Hg/mL/m2) and CA = SV/arterial pulse pressure (PP).
The study objectives and procedures were approved by the local eth- Respiratory PP variation (PPV) was automatically calculated
ics committee (“Comité de Protection des Personnes Nord-Ouest by the Philips monitoring system (22) and represents the aver-
II” CHU—Place V. Pauchet, 80054 AMIENS Cedex 1). All subjects age of three successive values. Dynamic Ea (Eadyn) was defined
received written information about the study and provided their as Eadyn = PPV/SVV (17).
consent to participate. The present article was drafted in compli-
ance with the Strengthening The Reporting of OBservational Stud- Study Procedures
ies in Epidemiology (STROBE) checklist for cohort studies (16). The following clinical variables were recorded: age, gender,
This prospective, observational study was performed in the weight, ventilation variables, and main diagnosis. When subject
Amiens University Hospital cardiovascular and thoracic ICU. has inclusion criteria, baseline measurements of HR, SAP, MAP,

2 www.ccmjournal.org XXX 2017 • Volume XX • Number XXX

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Online Clinical Investigation

diastolic arterial pressure, CVP, TPRi, SVi, CI, pre-ejection time, Student paired t test, Student t test, and the Mann-Whitney U
and systolic time interval, PPV and arterial/venous blood gas test were used to assess statistical significance, as appropriate. A
levels were obtained. After 15 minutes of hemodynamic stability receiver-operating characteristic curve was established for SVV,
defined as a variation of MAP less than 10% with norepineph- TPRi, SVi, and EA/Ees ratio to predict a greater than 15% increase
rine infusion, a second set of measurements was recorded (4). in SV (4). The receiver operating characteristic (ROC) curves
All patients had been sedated by continuous infusion of propo- were compared using the DeLong test. The threshold for statisti-
fol and were fully adapted to volume-controlled mode with a cal significance was set to p value of less than 0.05. SPSS software
tidal volume set at 7–9 mL/kg of ideal body weight, and a posi- (version 24; IBM, New York, NY) was used for statistical analysis.
tive end-expiratory pressure of 5–8 cm H2O. Ventilator settings
and propofol rate were not modified during the study period.
RESULTS
Of 30 patients included in the study, two were excluded. The
Statistical Analysis
study is based on 28 subjects (Fig. 1) who had undergone car-
A sample of 30 patients was calculated to be sufficient to dem-
diac surgery (Table 1). None of the subjects were treated with
onstrate that EA/Ees ratio predict an increase of SV in response to
epinephrine or dobutamine during the study period. Subjects’
norepinephrine treatment with an area under the curve (AUC)
demographic characteristics are presented in Table 1. At base-
greater than 0.80, a power of 80%, and an alpha risk of 0.05.
line, there were no differences between SV responder and SV
The coefficient of variation (CV), precision, and least significant
nonresponder subjects. The prevalence of uncoupled subjects
change (LSC) for Ees were calculated on 10 patients. The CV is a
did not differ according to the type of surgery (p = 0.541),
statistical measure of the dispersion of Ees data points. The LSC
therefore allowing pooled analysis.
is the least amount of Ees change that can be considered to be
statistically significant, that is, the minimum percentage change
Effects of Norepinephrine Infusion on Hemodynamic
between two successive measurements that can be considered
Variables in the Overall Population (n = 28)
not to be due to random error and therefore representing a real
At baseline, median EA was 1.69 (1.39–2.24) mm Hg/mL,
change in Ees in 10 patients. The CV was 7.7% ± 0.6% and the LSC
median Ees was 1.1 (0.65–1.42) mm Hg/mL, median EA/Ees
was 10.9% ± 0.8%. The distribution of the variables was assessed
ratio was 1.69 (1.34–2.43), and 22 of the 28 subjects (79%)
using a Shapiro-Wilk test. Data are expressed as the proportion
were classified as uncoupled (i.e., EA/Ees > 1.36). Values for EA,
(%), mean ± sd, or median (interquartile range), as appropriate.
Ees, and EA/Ees ratio did not differ between men and women
Patients were classified according to the effect of norepinephrine
(p > 0.05). In the study population as a whole, norepinephrine
on SVi (4). A SV responder was defined as a greater than 15%
infusion induced increases in arterial pressure, CVP, SVi, CI,
increase in SVi (23). The nonparametric Wilcoxon rank-sum test,
TPRi, EA, and Ees (Table 2).

Effects of Norepinephrine
Infusion on Hemodynamic
Variables According to SV
Response
Fifteen of the 28 subjects (54%)
were classified as SV responders.
Baseline TPRi and EA/Ees were
higher, but SVi, CI, and LVEF
were lower in SV respond-
ers compared with SV non-
responders (Table 2). In SV
responders, norepinephrine
infusion induced increases in
arterial pressure, SVi, CI, TPRi,
EA, Ees, and decreases in CA, EA/
Ees ratio (Table 2 and Fig. 2).
In SV nonresponders, nor-
epinephrine infusion induced
increases in arterial pressure,
TPRi, EA, Ees, and decreases in
CA (Table 2 and Fig. 2).
Changes in EA and TPRi were
greater in SV nonresponders
than in SV responders (41%
Figure 1. Study flow chart.
[22–61%] vs 15% [1–21%];

Critical Care Medicine www.ccmjournal.org 3

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Guinot et al

TABLE 1. Characteristics of the Study Participants on Inclusion

Variables SV Responders (n = 15) SV Nonresponders (n = 13) p

Age (yr), mean ± sd 69 ± 10 69 ± 9 NS

Gender (female/male) 4/11 3/10 NS
Body mass index (Kg/m2), mean ± sd 31.4 ± 6 30.3 ± 5 NS
New Simplified Acute Physiology Score, mean ± sd 48 ± 13 40 ± 13 NS
Disease, n (%)
  Arterial hypertension 13 (87) 10 (77) NS
  Aortic stenosis 7 (47) 7 (54) NS
  Diabetes mellitus 5 (33) 3 (23) NS
 Dyslipidemia 5 (33) 7 (54) NS
Heart surgery, n (%)
  Valve replacement 9 (54) 9 (69)
 CABG 2 (13) 3 (23)
  Mixed (valve replacement and CABG) 2 (13) 0 NS
  Other (atrial myxoma, ascending aorta) 2 (13) 1 (8)
Left ventricular ejection fraction (%), mean ± sd 51 ± 9 56 ± 15 NS
Cardiopulmonary bypass time (min), mean ± sd 108 ± 49 98 ± 35 NS
ICU stays (d), median (25–75le) 3 (2–6) 3 (2–5) NS
Respiratory variables, mean ± sd
  Tidal volume (mL/kg of predicted body weight) 7.7 ± 0.5 7.7 ± 0.7 NS
  Total positive end-expiratory pressure (cm H2O) 6 ± 1 6 ± 2 NS
  Propofol (mg/kg/hr) 1.5 ± 0.5 1.6 ± 0.4 NS
Cumulated fluid infusion before norepinephrine treatment (mL), 717 ± 297 827 ± 188 NS
mean ± sd
Norepinephrine dose (µg/kg/min), median (25–75le) 0.10 (0.064–0.2) 0.089 (0.069–0.11) NS
Central venous saturation (%), mean ± sd 59 ± 12 62 ± 7 NS
Arterial lactates (mmol/L), mean ± sd 1.62 ± 0.52 1.76 ± 0.66 NS
Death, n (%) 1 (7) 1 (8) NS
CABG = coronary artery bypass graft, NS = not significant, SV = stroke volume.
p value refers to comparison between stroke volume (SV) responders and SV nonresponders patients.

p = 0.0001 and 41% [18–60%] vs 8% [0–17%]; p = 0.0001). With an AUC of 0.72 (95% CI, 0.54–0.9; p = 0.040), SVi
Changes in Ees did not differ between the two groups of patients fairly predicted SV responsiveness. With an AUC of 0.73 (95%
(29% [13–60%] vs 25% [12–61%]; p = 0.821). Changes in EA/ CI, 0.55–0.93; p = 0.032), TPRi fairly predicted SV responsive-
Ees ratio differed between SV responders and SV nonresponders ness. With an AUC of 0.66 (95% CI, 0.46–0.83; p = 0.135), SVV
(–16% [–23 to 0%] vs 1% [0–22%]; p = 0.010). did not predict SV responsiveness. The AUC of EA/Ees ratio was
greater than those of SVi and TPRi (p < 0.05).
Predictability of EA/Ees Ratio, TPRi, SVi, and SVV
to Discriminate SV Increases With Norepinephrine
Infusion DISCUSSION
Baseline EA/Ees ratio was predictive of SV responsiveness with The main result of the study is that norepinephrine improve-
an AUC of 0.87 (95% CI, 0.72–0.98; p = 0.001) (Fig. 3). The ment in ventriculo-arterial coupling was associated to SV
cutoff was 1.5 with a sensitivity of 94% (68–100%), a speci- response in cardiac postoperative hypotensive subjects. Both
ficity of 77% (46–95%), a positive likelihood ratio of 4.07, a Ees and EA are abnormal in many subjects with persistent arte-
negative likelihood ratio of 0.087, a positive predictive value of rial hypotension despite volume expansion. Most of the sub-
82%, and a negative predictive value of 91%. jects presented abnormal EA/Ees ratio, mainly due to high EA

4 www.ccmjournal.org XXX 2017 • Volume XX • Number XXX

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Online Clinical Investigation

TABLE 2. Comparison of Hemodynamic Variables in Stroke Volume Responders and

Nonresponders
Hemodynamic Variables SV Responders (n = 15) SV Nonresponders (n = 13) p

Indexed stroke volume (mL/m2), mean ± sd

  Pre norepinephrine 19 ± 8 27 ± 10 0.038
  Post norepinephrine 25 ± 10 a
27 ± 8 NS
Left ventricular end-systolic volume (mL), mean ± sd
  Pre norepinephrine 94 ± 59 54 ± 16 0.027
  Post norepinephrine 86 ± 55 62 ± 15 NS
Left ventricular end-diastolic volume (mL), mean ± sd
  Pre norepinephrine 133 ± 63 106 ± 33 NS
  Post norepinephrine 138 ± 57 114 ± 22 NS
Cardiac index (mL/min/m ), mean ± sd
2

  Pre norepinephrine 1.56 ± 0.6 1.95 ± 0.5 0.033

  Post norepinephrine 1.99 ± 0.6 a
2.05 ± 0.5 NS
Heart rate (beats/min), mean ± sd
  Pre norepinephrine 81 ± 19 76 ± 16 NS
  Post norepinephrine 80 ± 19 79 ± 18 NS
Systolic arterial pressure (mm Hg), mean ± sd
  Pre norepinephrine 84 ± 11 85 ± 10 NS
  Post norepinephrine 119 ± 8 a 119 ± 13 a NS
Diastolic arterial pressure (mm Hg), mean ± sd
  Pre norepinephrine 46 ± 8 46 ± 8 NS
  Post norepinephrine 60 ± 9 a
61 ± 13 a
NS
Mean arterial pressure (mm Hg), mean ± sd
  Pre norepinephrine 59 ± 11 59 ± 8 NS
  Post norepinephrine 80 ± 7 a
80 ± 11 a
NS
PP (mm Hg), mean ± sd
  Pre norepinephrine 37 ± 10 39 ± 9 NS
  Post norepinephrine 58 ± 10 a
58 ± 12 a
NS
Central venous pressure (mm Hg), mean ± sd
  Pre norepinephrine 7 ± 4 8 ± 4 NS
  Post norepinephrine 9 ± 4 a
10 ± 4 a
NS
Respiratory variation of stroke volume (%), mean ± sd
  Pre norepinephrine 25 ± 113 19 ± 10 NS
  Post norepinephrine 22 ± 9 23 ± 15 NS
Respiratory variation of PP (%), mean ± sd
  Pre norepinephrine 15 ± 7 13 ± 7 NS
  Post norepinephrine 11 ± 6 a 11 ± 8 NS

(Continued)

Critical Care Medicine www.ccmjournal.org 5

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Guinot et al

TABLE 2. (Continued). Comparison of Hemodynamic Variables in Stroke Volume

Responders and Nonresponders
Hemodynamic Variables SV Responders (n = 15) SV Nonresponders (n = 13) p

Dynamic arterial elastance (units), median (25–75le)

  Pre norepinephrine 0.7 (0.38–1) 0.78 (0.45–0.89) NS
  Post norepinephrine 0.58 (0.3–0.79) 0.44 (0.27–1.26) NS
Inferior vena cava diameter (cm), mean ± sd
  Pre norepinephrine 1.69 ± 0.3 1.79 ± 0.3 NS
  Post norepinephrine 1.66 ± 0.7 1.64 ± 0.4 NS
Left ventricular ejection fraction (%), mean ± sd
  Pre norepinephrine 37 ± 12 48 ± 11 0.011
  Post norepinephrine 44 ± 19 a
47 ± 11 NS
Total arterial compliance (mL/mm Hg), mean ± sd
  Pre norepinephrine 1.06 ± 0.44 1.4 ± 0.58 NS
  Post norepinephrine 0.88 ± 0.37 a
0.89 ± 0.24 a
NS
Indexed total peripheral resistance (mm Hg/mL/m ), 2

  median (25–75le)
  Pre norepinephrine 34 (28–51) 26 (23–36) 0.033
  Post norepinephrine 37 (30–57) 35 (30–45) a
NS
EA (mm Hg/mL), median (25–75le)
  Pre norepinephrine 1.8 (1.4–3.3) 1.6 (1.1–2) NS
  Post norepinephrine 1.9 (1.6–3.3) a
2.1 (1.8–2.6) a
NS
Ees (mm Hg/mL), median (25–75le)
  Pre norepinephrine 0.83 (0.57–1.34) 1.15 (0.72–1.72) NS
  Post norepinephrine 1.14 (0.84–1.56) a
1.58 (1.14–1.93) a
NS
EA/Ees ratio, median (25–75le)
  Pre norepinephrine 2 (1.7–2.9) 1.34 (1.1–1.61) 0.0001
  Post norepinephrine 1.8 (1.4–2.4) a
1.36 (1.23–1.69) 0.029
Ea = arterial elastance, Ees = end-systolic ventricular elastance, NS = not significant, PP = pulse pressure, SV = stroke volume.
p < 0.05 refers to comparison within groups, post versus pre norepinephrine infusion.
a

p value refers to comparison between stroke volume (SV) responders and SV nonresponders.

(low CA, high TPRi) and low Ees. Baseline EA/Ees has the highest cardiovascular system and its response to treatment? There
AUC ROC to predict SV increase. are several answers. The first one is that there is strong evi-
The results of our study confirm previous ones (6, 13, 14) dence from the cardiology area that ventriculo-arterial cou-
but they add a new perspective on the cardiovascular effects pling independently predicts outcomes and the response to
of norepinephrine. The predominant effect of norepineph- therapy (24, 25). When applied to the context of critical care
rine on myocardial and vascular functions may depend on the and cardiac surgery patients, it is first an extension of these
cardiovascular baseline state. Cardiovascular state was better validated concepts. Second, in a patient who is hypotensive
described by ventriculo-arterial coupling than by flow vari- despite volume expansion and who receives norepinephrine,
ables (CO, TPRi) analyzed independently from each other. the efficacy of this treatment can be classically estimated on
Norepinephrine may have mainly inotropic effects in case of three independent levels: 1) increase in arterial blood pres-
LV dysfunction and mainly vasoconstrictive effects in case of sure and SV; none of these variables differentiated in this
arterial dysfunction. Correction of either of these variables study between SV responders and nonresponders; 2) when
results in improvement of ventriculo-arterial coupling. echocardiography variables are measured, there are differ-
One clinically relevant question may be why should we con- ences that reveal which SV responders and nonresponders
sider ventriculo-arterial coupling as a variable to describe the patients are different before and after norepinephrine infusion;

6 www.ccmjournal.org XXX 2017 • Volume XX • Number XXX

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Online Clinical Investigation

Figure 2. A, Ventricular pressure-volume diagram in stroke volume (SV) responder patients (A) and in SV nonresponder patients (B). EA = arterial
elastance, Ees = end-systolic ventricular elastance, EDV = left ventricular end-diastolic volume, ESV = left ventricular end-systolic volume, LV = left
ventricle.

contractility, CA, TPRi, and their interaction explains

the mechanisms of the SV responsiveness. It must be
stressed that Ees is the only variable that takes into
consideration LV contractility in a (nearly) load-inde-
pendent manner (7).
Norepinephrine infusion increased EA and Ees but
not to the same extent between groups, resulting in
ventriculo-arterial coupling improvement and SV
increase in the SV responders group. We demonstrated
an increase in Ees in relation to several known phar-
macologic effects of norepinephrine (activation of
alpha- and beta-adrenergic receptors, improvement
in coronary perfusion pressure, etc.) (26–28). On the
contrary, the effects of norepinephrine on EA may have
been different between the two groups because the
hemodynamic response to norepinephrine depends
on the baseline vasomotor tone (29). In SV nonre-
sponders, the increase in EA may be mainly due to vaso-
constrictive effects on resistive component (TPRi). SV
did not decrease because the increase of arterial load
was compensated by the increase of Ees. In SV respond-
ers, EA changes may be mostly due to inotropic effects.
Norepinephrine did not change TPRi because the val-
ues were already high due to sympathetic activation
Figure 3. Receiving operating curves of arterial elastance/end-systolic ventricular probably secondary to more altered LV function before
elastance, respiratory variation of stroke volume (SVV), indexed total peripheral
resistance (TPRi), and indexed stroke volume (VESi) to discriminate response of SV
infusion of norepinephrine (Table 2) (29). Thus, the
to norepinephrine infusion. V-A = ventriculo-arterial. Ees increase was greater than that of EA restoring ven-
triculo-arterial coupling, increasing LVEF and SV (9).
nevertheless, “cardio-centric” variables (SV) do not allow a The effects of norepinephrine infusion in SV responders may
thorough understanding of the mechanisms that explain the not be related to preload changes because preload variables
differences between SV responders and nonresponders; and (EDV, inferior vena cava diameter, and SVV) did not change
3) VA-coupling description that integrates HR, SV, ventricular with norepinephrine infusion.

Critical Care Medicine www.ccmjournal.org 7

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Guinot et al

Comparison of the predictive value of the various have been developed that include arterial input impedance
hemodynamic variables shows that the EA/Ees ratio had the and wave reflection, but such measurements would be dif-
highest ROC AUC. SVV was not predictive of SV increase ficult to perform at the bedside. Despite these limitations,
because SV increase was not associated to preload changes noninvasive evaluation of E es and E A was validated against
(4). This validates our inclusion criteria: patients with per- the gold standard method and has been widely used to
sistent arterial hypotension despite volume expansion. The evaluate the effects of therapeutic interventions (12, 15).
higher predictive value of EA/Ees ratio is probably due to the
fact that this variable reflects HR, LV inotropism, SV, TPRi,
CONCLUSIONS
arterial pressure, and their interactions, reflecting the inte-
In conclusion, analysis of ventriculo-arterial coupling
grated physiology of the cardiovascular system, in which the
allowed better prediction of the SV response to norepineph-
heart and arterial circulation are coupled. Flow variables
rine. The response of the cardiovascular system to norepi-
analyzed independently from each other (SV, TPRi,…) were
nephrine depends on LV contractility, the vascular system,
less predictive because they do not reflect cardiac and vas-
and their interactions. These results could be an incentive to
cular interactions.
assess the cardiovascular effects of norepinephrine in cardiac
Our results provide a strong incentive to use combined flow
surgery patients by taking into account not only CO and/or
and pressure variables to evaluate the cardiovascular effects of
blood pressure but also ventriculo-arterial coupling.
catecholamines. This approach is an extension of ventriculo-
arterial coupling from the cardiology literature to the postcar-
diac surgery period. The results obtained in this study confirm REFERENCES
1. De Backer D, Biston P, Devriendt J, et al; SOAP II Investigators: Com-
the importance of an integrated analysis of the cardiac and vas- parison of dopamine and norepinephrine in the treatment of shock. N
cular systems (30, 31). Engl J Med 2010; 362:779–789
Several limitations of our study should be considered. 2. Gamper G, Havel C, Arrich J, et al: Vasopressors for hypotensive
Because we included subjects operated of different car- shock. Cochrane Database Syst Rev 2016; 2:CD003709
diac surgery, reliability of our results can be questioned. 3. Persichini R, Silva S, Teboul JL, et al: Effects of norepinephrine on
mean systemic pressure and venous return in human septic shock.
We did not demonstrate any difference of baseline vari- Crit Care Med 2012; 40:3146–3153
ables (demographic, ICU) between SV responders and 4. Maas JJ, Pinsky MR, de Wilde RB, et al: Cardiac output response to
SV nonresponders. E es is a variable that reflects the func- norepinephrine in postoperative cardiac surgery patients: Interpreta-
tion with venous return and cardiac function curves. Crit Care Med
tional, structural, and geometric characteristics of the LV 2013; 41:143–150
(32). Ventriculo-arterial coupling represents a variable of 5. Perez P, Kimmoun A, Blime V, et al: Increasing mean arterial pres-
integrated cardiovascular physiology that is an alteration sure in cardiogenic shock secondary to myocardial infarction:
perse, probably independent from the underlying type of Effects on hemodynamics and tissue oxygenation. Shock 2014;
41:269–274
heart disease (7, 32). The methods used to calculate E es
6. Levy B, Perez P, Perny J, et al: Comparison of norepinephrine-dobu-
and EA can be questioned. We measured E es by a nonin- tamine to epinephrine for hemodynamics, lactate metabolism, and
vasive single-beat method that is based on a linear end- organ function variables in cardiogenic shock. A prospective, ran-
systolic pressure-volume relation and a constant volume domized pilot study. Crit Care Med 2011; 39:450–455
7. Sagawa K, Suga H, Shoukas AA, et al: End-systolic pressure/vol-
axis intercept (V0) of the end-systolic pressure-volume
ume ratio: A new index of ventricular contractility. Am J Cardiol 1977;
relation (18). E es calculation assumes that the end-systolic 40:748–753
pressure-volume relationship is independent of load and 8. Sagawa K: The ventricular pressure-volume diagram revisited. Circ
that its slope is linear. V 0 is assumed to be negligible com- Res 1978; 43:677–687
pared with ESV and to be insensitive change of loading or 9. Sunagawa K, Maughan WL, Burkhoff D, et al: Left ventricular inter-
action with arterial load studied in isolated canine ventricle. Am J
inotropic conditions (7). Changes in LV contractility can Physiol 1983; 245:H773–H780
be due to changes in E es or V 0. Although it can be argued 10. Sunagawa K, Sagawa K, Maughan WL: Ventricular interaction with
that the estimation of end-systolic pressure from the radial the loading system. Ann Biomed Eng 1984; 12:163–189
artery is not fully validated, this would only affect the pre- 11. Chantler PD: Arterial ventricular uncoupling with age and disease and
recoupling with exercise. Exerc Sport Sci Rev 2017; 45:70–79
cision of absolute values of E A and E es but not of the EA/
12. Maurer MS, Sackner-Bernstein JD, El-Khoury Rumbarger L, et al:
E es ratio because the error in end-systolic pressure would Mechanisms underlying improvements in ejection fraction with
be similar. In this regard, the predictive value of the E A/E es carvedilol in heart failure. Circ Heart Fail 2009; 2:189–196
for increases in SV is valid. We measured E A from a radial 13. How OJ, Røsner A, Kildal AB, et al: Dobutamine-norepinephrine, but
artery catheter that certainly differs from aortic pressure, not vasopressin, restores the ventriculoarterial matching in experimen-
tal cardiogenic shock. Transl Res 2010; 156:273–281
especially in younger patients, but probably less so in 14. Beurton A, Ducrocq N, Auchet T, et al: Beneficial effects of norepi-
elderly patients, who constituted the majority of subjects nephrine alone on cardiovascular function and tissue oxygenation in a
in our study. Nevertheless, radial artery pressure has been pig model of cardiogenic shock. Shock 2016; 46:214–218
demonstrated to provide a good estimation of end-systolic 15. Ky B, French B, May Khan A, et al: Ventricular-arterial coupling,
remodeling, and prognosis in chronic heart failure. J Am Coll Cardiol
LV pressure (33). Arterial load assessment was based on 2013; 62:1165–1172
the two-element Windkessel model and integrative simpli- 16. von Elm E, Altman DG, Egger M, et al; STROBE Initiative: The
fication, defined by the effective E A. More precise models Strengthening the Reporting of Observational Studies in Epidemi-

8 www.ccmjournal.org XXX 2017 • Volume XX • Number XXX

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Online Clinical Investigation

ology (STROBE) statement: Guidelines for reporting observational 25. Matsumoto K, Tanaka H, Ooka J, et al: Significant prognostic impact
studies. Ann Intern Med 2007; 147:573–577 of improvement in ventriculo-arterial coupling induced by dobutamine
17. Guinot PG, Abou-Arab O, Guilbart M, et al: Monitoring dynamic stress on cardiovascular outcome for patients with dilated cardiomy-
arterial elastance as a means of decreasing the duration of norepi- opathy. Eur Heart J Cardiovasc Imaging 2016; 17:1296–1304
nephrine treatment in vasoplegic syndrome following cardiac sur- 26. Hattori Y, Takeda Y, Nakaya H, et al: Induction by endogenous noradren-
gery: A prospective, randomized trial. Intensive Care Med 2017; aline of an alpha 1-adrenoceptor-mediated positive inotropic effect in
43:643–651 rabbit papillary muscles. Br J Pharmacol 1993; 109:1232–1238
18. Chen CH, Fetics B, Nevo E, et al: Noninvasive single-beat determina- 27. Skomedal T, Borthne K, Aass H, et al: Comparison between alpha-1
tion of left ventricular end-systolic elastance in humans. J Am Coll adrenoceptor-mediated and beta adrenoceptor-mediated inotropic
Cardiol 2001; 38:2028–2034 components elicited by norepinephrine in failing human ventricular
19. Garcia D, Barenbrug PJ, Pibarot P, et al: A ventricular-vascular cou- muscle. J Pharmacol Exp Ther 1997; 280:721–729
pling model in presence of aortic stenosis. Am J Physiol Heart Circ 28. ZANETTI ME, OPDYKE DF: Similarity of acute hemodynamic
Physiol 2005; 288:H1874–H1884 response to 1-epinephrine and l-arterenol. J Pharmacol Exp Ther
20. Kelly RP, Ting CT, Yang TM, et al: Effective arterial elastance as 1953; 109:107–115
index of arterial vascular load in humans. Circulation 1992; 29. Wilson RA, Di Mario C, Krams R, et al: In vivo measurement of regional
86:513–521 large artery compliance by intravascular ultrasound under pentobarbi-
21. Chen CH, Nakayama M, Nevo E, et al: Coupled systolic-ventricular tal anesthesia. Angiology 1995; 46:481–488
and vascular stiffening with age: Implications for pressure regula- 30. Merillon JP, Fontenier GJ, Lerallut JF, et al: Aortic input impedance in
tion and cardiac reserve in the elderly. J Am Coll Cardiol 1998; normal man and arterial hypertension: Its modification during changes
32:1221–1227 in aortic pressure. Cardiovasc Res 1982; 16:646–656
22. Mahjoub Y, Lorne E, Micaux Y, et al: Accuracy of automated con- 31. Segers P, Steendijk P, Stergiopulos N, et al: Predicting systolic and
tinuous calculation of pulse pressure variation in critically ill patients. diastolic aortic blood pressure and stroke volume in the intact sheep.
Intensive Care Med 2011; 37:360–361 J Biomech 2001; 34:41–50
23. Guinot PG, Urbina B, de Broca B, et al: Predictability of the respira- 32. Senzaki H, Chen CH, Kass DA: Single-beat estimation of end-systolic
tory variation of stroke volume varies according to the definition of pressure-volume relation in humans. A new method with the potential
fluid responsiveness. Br J Anaesth 2014; 112:580–581 for noninvasive application. Circulation 1996; 94:2497–2506
24. Borlaug BA, Kass DA: Ventricular-vascular interaction in heart failure. 33. Pauca AL, Wallenhaupt SL, Kon ND, et al: Does radial artery pressure
Heart Fail Clin 2008; 4:23–36 accurately reflect aortic pressure? Chest 1992; 102:1193–1198

Critical Care Medicine www.ccmjournal.org 9

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.