d

v

c

o

fi

o

m

a

i

f

o

l

c

P

n

d

Mtvwevtrdtqfip

pablmcchdrhftl

est

J

ARTICLE IN PRESS

Evaluation of Left Ventricular Function in Anesthetized Patients UsingFemoral Artery dP/dtmax

Stefan G. De Hert, MD, PhD,* Dominique Robert, MD,* Stefanie Cromheecke, MD,*

Frédéric Michard, MD, PhD,† Jan Nijs, MD,‡ and Inez E. Rodrigus, MD, PhD‡

t

m

d

i

d

�s

v

m

c

v

i

i

a

d

c

©

K

Objective: The purpose of this study was to compare dP/

tmax estimated from a femoral artery pressure tracing to left

entricular (LV) dP/dtmax during various alterations in myo-

ardial loading and contractile function.

Participants: Seventy patients scheduled for elective cor-

nary artery bypass surgery.

Methods: All patients were instrumented with a high-

delity LV catheter, a pulmonary artery catheter, and a fem-

ral arterial catheter. In 40 patients, hemodynamic measure-

ents were performed before and after passive leg raising

nd before and after calcium administration (5 mg/kg); and

n 30 other patients, hemodynamic measurements were per-

ormed before and after dobutamine infusion (5 �g/kg/min

ver 10 minutes).

Results: LV and femoral dP/dtmax were significantly corre-

ated (r � 0.82, p < 0.001), but femoral dP/dtmax systemati-

ally underestimated LV dP/dtmax (bias � �361 � 96 mmHg/s).

assive leg raising induced significant increases in central ve-

ous pressure and LV end-diastolic pressure, but femoral

ee (University Hospital Antwerp, Edegem, Belgium), and written

iapc

oieGeifildclpsalpidvto

UCF

mB

ournal of Cardiothoracic and Vascular Anesthesia, Vol xx, No x (Month),

ered. Calcium administration induced significant and

arked increases in LV dP/dtmax (23% � 9%) and femoral

P/dtmax (37% � 14%) associated with a significant increase

n stroke volume (9% � 2%). Dobutamine infusion also in-

uced significant and marked increases in LV dP/dtmax (25%

8%) and femoral dP/dtmax (35% � 12%) associated with a

ignificant increase in stroke volume (14% � 3%). Overall, a

ery close linear relationship (r � 0.93) and a good agree-

ent (bias � �5 � 17 mmHg/s) were found between

hanges in LV dP/dtmax and changes in femoral dP/dtmax. A

ery close relationship was also observed between changes

n LV dP/dtmax and changes in femoral dP/dtmax during each

ntervention (leg raising, calcium administration, and dobut-

mine infusion).

Conclusion: Femoral dP/dtmax underestimated LV dP/

tmax, but changes in femoral dP/dtmax accurately reflected

hanges in LV dP/dtmax during various interventions.

2006 Elsevier Inc. All rights reserved.

EY WORDS: coronary surgery, left ventricular function,

P/dtmax, stroke volume, and LV dP/dtmax remained unal- dP/dtmax

OST VARIABLES DESCRIBING myocardial contrac-tile function are influenced by cardiac loading condi-

ions.1,2 A relative exception to this is the peak rate of leftentricular (LV) pressure development (dP/dtmax), which variesith preload but is relatively independent of afterload.3 Both in

xperimental and clinical research, LV dP/dtmax is used as aariable for assessment of LV contractile function. In addition,his variable can be used to evaluate the individual functionaleserve capacity of the heart, by analysis of changes in LVP/dtmax in response to alterations in cardiac loading condi-ions.4-9 Unfortunately, the measurement of LV dP/dtmax re-uires the catheterization of the left ventricle with a high-delity pressure catheter and hence is not common in clinicalractice.Most critically ill patients with hemodynamic instability and

atients undergoing high-risk surgery are instrumented with anrterial catheter for accurate arterial pressure monitoring andlood sampling. In this context, it can be considered to calcu-ate the dP/dtmax derived from the arterial pressure curve (the

aximal ascending slope of the peripheral arterial pressureurve). However, whether this latter parameter reflects LVontractile function remains to be determined. The authorsypothesized that the use of a peripheral artery–derived dP/tmax would provide similar information to LV dP/dtmax withegard to the assessment of functional reserve capacity of theeart. To test this hypothesis, the changes in dP/dtmax derivedrom the femoral artery pressure tracing were compared withhe changes in LV dP/dtmax during various alterations in cardiacoading conditions and contractile function.

METHODS

The study was performed in 70 patients (54 men, 16 women) with anjection fraction �45% scheduled for elective coronary artery bypassurgery. The study was approved by the Institutional Ethical Commit-

nformed consent was obtained. Patients undergoing repeat coronaryrtery surgery, concurrent valve repair, or aneurysm resection andatients with unstable angina or with valve insufficiency were ex-luded.

All preoperative cardiac medication was continued until the morningf surgery. In the operating room, patients received routine monitoringncluding 5-lead electrocardiogram, radial and pulmonary artery cath-ters with continuous cardiac output (CO) measurement (Swananz CCO/VIP; Edwards Lifesciences LLC, Irvine, CA), pulse oxim-

try, capnography, and blood and urine bladder temperature monitor-ng. All patients were also instrumented with a thermistor-tipped fluidlled femoral arterial catheter connected to a transpulmonary thermodi-

ution monitor (PiCCOplus; Pulsion, Munich, Germany). This is aevice that quantifies several parameters, including continuous (pulseontour) CO and derived parameters, cardiac preload, systemic vascu-ar resistance, and extravascular lung water. For this technique, theatient requires a central venous catheter in the internal jugular orubclavian vein (in this study, the right atrial port of the pulmonaryrtery catheter) and an arterial catheter with a thermistor in 1 of thearger arteries of the body (in this study, the femoral artery). Therinciple is that a known volume of a thermal indicator (20 mL ofce-cold saline) is injected into the central vein. The injectate rapidlyisperses within the pulmonary and cardiac volumes (intrathoracicolume). When the thermal signal reaches the arterial thermister, aemperature difference is detected and a dissipation curve is generatedn which CO is calculated.

From the Departments of *Anesthesiology and ‡Cardiac Surgery,niversity Hospital Antwerp, Edegem, Belgium; and †Department ofritical Care, Marie Lannelongue Hospital, University Paris XI, Paris,rance.Address reprint requests to Stefan G. De Hert, MD, PhD, Depart-

ent of Anesthesiology, University Hospital Antwerp, Wilrijkstraat 10,-2650 Edegem, Belgium. E-mail: stefan.dehert@ua.ac.be© 2006 Elsevier Inc. All rights reserved.1053-0770/06/xx0x-0001$32.00/0

doi:10.1053/j.jvca.2005.11.006

12006: pp xxx

tMAp

(ltwLPTrida(d2tost

pvbitravmta

tdac

spb�Tb2dw

ptptamGDt

bssmci�iida3

0uO

C

2 DE HERT ET AL

ARTICLE IN PRESS

Anesthesia was based on a continuous infusion of remifentanil at 0.2o 0.4 �g/kg/min and sevoflurane, 0.5% to 1% end-tidal concentration.

uscle paralysis was obtained with pancuronium bromide, 0.1 mg/kg.fter intubation, standard median sternotomy and pericardiotomy wereerformed, and the aortic cannula was secured in place.In each patient, a sterilized, prezeroed electronic tipmanometer

MTCP3Fc catheter; Dräger Medical Electronics, Best, The Nether-ands; frequency response � 100 KHz) was inserted in the LV cavityhrough the apical dimple. Zero and gain setting of the tipmanometersere also checked against a high-fidelity pressure gauge (Druck Ltd,eicester, UK) after removal. The catheter was connected to a Hewlettackard monitor (HP78342A; Hewlett Packard, Brussels, Belgium).he output signal of the pressure transducer system was digitally

ecorded at 1-millisecond intervals (Codas; DataQ, Akron, OH), allow-ng the calculation of LV dP/dtmax as previously described.5,7 FemoralP/dtmax was obtained automatically from the analysis of the femoralrterial pressure curve by the available thermodilution monitorPiCCOplus). This device automatically calculates and displays dP/tmax. The analog pressure signal is captured with a sampling rate of50 Hz and then smoothed to a sampling rate of 125 Hz in order to findhe maximum difference between 2 following samples of the rising partf the arterial pressure curve. The dP/dtmax value displayed on thecreen represents the floating average of the heart beats sampled duringhe last 12 seconds and is updated on a beat-by-beat basis.

Heart rate was kept constant by epicardial atrioventricular sequentialacing at a rate of 90 beats/min. All measurements were obtained withentilation suspended at end-expiration. Measurements were obtainedefore and after each intervention. In protocol 1 (n � 40), the firstntervention consisted of a leg elevation by raising the caudal part ofhe surgical table by 45°. After the measurements, the table waseturned to its normal position and hemodynamic parameters werellowed to recover for at least 10 minutes. After return to baselinealues, a new control measurement was obtained and a bolus of 5g/kg of calcium chloride was administered. In protocol 2 (n � 30),

he intervention consisted of a continuous administration of dobut-mine, 5 �g/kg/min.

All data were controlled for normal distribution. The sample size ofhe study was calculated based on the changes in dP/dtmax with theifferent interventions as the primary outcome variable. For protocol 1,minimum detected difference of 40 mmHg/s with leg elevation was

Table 1. Hemodynamic Effects of Leg R

Control

Left ventricular dataEDP (mmHg) 9 � 3maximal LVP (mmHg) 94 � 7dP/dtmax (mmHg/s) 954 � 126�dP/dtmax (mmHg/s) 33 � 66

PAC dataCVP (mmHg) 8 � 3MPAP (mmHg) 17 � 6SV (mL) 61 � 10

Femoral artery pressurederived datapeak pressure (mmHg) 100 � 10dP/dtmax (mmHg/s) 590 � 117�dP/dtmax (mmHg/s) 27 � 67SV (mL) 64 � 9

Abbreviations: EDP, end-diastolic pressure; LVP, left ventricular presVP, central venous pressure; MPAP, mean pulmonary artery pressu*Statistically significant difference (p � 0.01) versus control.

onsidered clinically significant. For a power of 0.9 and � � 0.01, a a

ample size of 36 patients was calculated to be appropriate. Forrotocol 2, a minimum detected difference of 100 mmHg/s with do-utamine was considered clinically significant. For a power of 0.9 and� 0.01, a sample size of 28 patients was calculated to be appropriate.he minimal detected differences considered clinically significant wereased on the data obtained in previous studies (protocol 1,5,6 protocol7). To find a correlation of 0.7 between femoral dP/dtmax and LVP/dtmax with a power of 0.9 and � � 0.01, a sample size of 23 patientsas necessary.Effects before and after the intervention were compared using a

aired t test. Linear regression analysis assessed the relationship be-ween absolute values and changes in dP/dtmax calculated from theressure tracings in the left ventricle and in the femoral artery. Statis-ical analysis was performed using the SigmaStat 2.03 software pack-ge (SPSS, Leuven, Belgium). Agreement between the 2 methods ofeasurement was assessed with the Bland-Altman analysis, using theraphPad software (version Prism 4) (GraphPad Software Inc, Saniego, CA). Data are expressed as mean � standard deviation. Statis-

ical significance was accepted at p � 0.01.

RESULTS

The hemodynamic effects of leg raising, calcium, and do-utamine adminstration are presented in Tables 1 and 2. Pas-ive leg raising induced significant increases in systemic pres-ure, central venous pressure, LV end-diastolic pressure, andean pulmonary arterial pressure. None of the other variables

hanged with leg raising (Table 1). Calcium administrationnduced significant and marked increases in LV dP/dtmax (23%

9%) and femoral dP/dtmax (37% � 14%) associated with anncrease in stroke volume (9% � 2%) (Table 1). Dobutaminenfusion also induced significant and marked increases in LVP/dtmax (25% � 8%) and femoral dP/dtmax (35% � 12%)ssociated with a significant increase in stroke volume (14% �%) (Table 2).LV and femoral dP/dtmax were significantly correlated (r �

.82, p � 0.001) (Fig 1), but femoral dP/dtmax systematicallynderestimated LV dP/dtmax (bias � �361 96 mmHg/s) (Fig 2).verall, a very close linear relationship (r � 0.93, p � 0.001)

g and Calcium Administration (n � 40)

g Raising Control Calcium

3 � 4* 9 � 3 11 � 38 � 7* 92 � 6 110 � 8*2 � 128 967 � 122 1285 � 168*

239 � 78

2 � 4* 9 � 3 10 � 32 � 5* 18 � 4 19 � 54 � 11 60 � 9 66 � 8*

4 � 11* 98 � 11 120 � 13*5 � 124 574 � 102 805 � 134*

222 � 757 � 12 64 � 11 70 � 10*

dP/dt, rate of pressure development; PAC, pulmonary artery catheter;, stroke volume; SVV, stroke volume variation.

aisin

Le

11098

126

1161

6

sure;re; SV

nd a good agreement (bias � �5 � 17 mmHg/s) were found

bdsdi

adcda

c

ueiaccbmplrrw(wwipwfb

pct

f

o

3EVALUATION OF LEFT VENTRICULAR FUNCTION

ARTICLE IN PRESS

etween changes in LV dP/dtmax and changes in femoral dP/tmax (Figs 3 and 4). A very close relationship was also ob-erved between changes in LV dP/dtmax and changes in femoralP/dtmax during each intervention (leg raising, calcium admin-stration, and dobutamine infusion) (Fig 5).

DISCUSSION

The findings of the present study indicate that in coronaryrtery surgery patients the femoral artery pressure–derivedP/dtmax systematically underestimated LV dP/dtmax but thathanges in femoral dP/dtmax accurately reflected changes in LVP/dtmax during various changes in cardiac loading conditionsnd function.

Assessment of LV contractile function remains a majorhallenge in the perioperative care of patients. The methods

Table 2. Hemodynamic Effects of Dobutamine Infusion (n � 30)

Control Dobutamine

Left ventricular dataEDP (mmHg) 8 � 3 9 � 4LVP (mmHg) 96 � 8 120 � 9*dP/dtmax (mmHg/s) 926 � 110 1290 � 165*�dP/dtmax (mmHg/s) 239 � 78

PAC dataCVP (mmHg) 8 � 3 8 � 4MPAP (mmHg) 18 � 6 17 � 5SV (mL) 59 � 10 68 � 11*

Femoral artery pressurederived data

peak pressure(mmHg)

103 � 10 119 � 9*

dP/dtmax (mmHg/s) 612 � 122 865 � 136*�dP/dtmax (mmHg/s) 235 � 55SV (mL) 63 � 8 72 � 11*

Abbreviations: EDP, end-diastolic pressure; LVP, left ventricularressure; dP/dt, rate of pressure development; PAC, pulmonary arteryatheter; CVP, central venous pressure; MPAP, mean pulmonary ar-ery pressure; SV, stroke volume; SVV, stroke volume variation.

*Statistically significant difference (p � 0.01) versus control.

Fig 1. Relationship between LV (left ventricular) dP/dtmax and

emoral dP/dtmax. c

sed can be divided into 2 groups based on the analysis ofither the ejection phases (such as ejection fraction) or thesovolumetric (pre-ejection) phases of cardiac contraction (suchs LV dP/dtmax). Even more challenging in perioperative patientare is the individual assessment of the functional reserveapacity of the heart. For this purpose, several approaches haveeen proposed such as the assessment of preload-adjustedaximal power10 or the measurement of the systolic arterial

ressure variation.11 Another possibility is to assess the relativeoad at which each individual ventricle is working.4,8,12 Briefly,elative load of the ventricle is indicative of the contractileeserve of the ventricle. Low relative load (�70%) is associatedith normal systolic function, whereas high relative load

�80%) is indicative of cardiac dysfunction. In a ventricleorking at low relative load, an increase in loading conditionsill result in a moderate elevation of ventricular pressure with

mprovement of systolic function and a faster rate of LVressure fall, whereas a similar intervention in a ventricleorking at high relative load induces impairment of systolic

unction and slowing of LV pressure fall.4,12 This concept haseen applied in coronary artery surgical patients, where it was

Fig 2. Bland and Altman representation of LV dP/dtmax and fem-

ral dP/dtmax.

Fig 3. Overall relationship between changes in LV dP/dtmax and

hanges in femoral dP/dtmax.

slpatvhidccrpmwau

bigdtmt

sdfod6NddcLclLc

advpddFpppotci

a

4 DE HERT ET AL

ARTICLE IN PRESS

hown that the hemodynamic response to an increase in cardiacoading conditions (ie, leg elevation ) allowed identification ofatients with impaired LV function, in whom the heart is notble to sustain an additional load of any kind. These patientsypically responded to leg elevation with decreases in strokeolume and dP/dtmax, delayed myocardial relaxation with en-anced load dependence of LV pressure fall, and a markedncrease in LV end-diastolic pressure.5-7,9,13-16 Analysis of dP/tmax and changes in dP/dtmax may be helpful to assess myo-ardial function and functional reserve of the heart in thelinical setting. Unfortunately, the measurement of LV dP/dtmax

equires the presence of an intraventricular pressure catheter,rohibiting its routine use in clinical practice. If similar infor-ation could be obtained from a less invasive method, thisould provide an additional tool not only in the hemodynamic

ssessment of the patient but also in the evaluation of individ-al perioperative myocardial function.The present study indicated that the calculation of dP/dtmax

ased on the femoral artery pressure tracing might provide thisnformation. Because most patients scheduled for major sur-ery and/or patients expected to develop perioperative hemo-ynamic instability are instrumented with an arterial catheter,he dP/dtmax data obtained from a peripheral artery become

ore readily available. These data may constitute an additionalool in perioperative hemodynamic monitoring and patient care.

However, there are a number of methodologic issues thathould be taken into account. First, physiologically, LVP/dtmax and femoral artery– derived dP/dtmax represent dif-erent phases in the cardiac cycle. LV dP/dtmax is a measuref isovolemic contraction, whereas femoral artery– derivedP/dtmax actually occurs during the LV ejection phase (Fig). This explains the observed difference in absolute values.evertheless, the changes with the interventions in femoralP/dtmax closely resembled the changes in LV dP/dtmax, in-icating that, despite this difference in phase of the cardiacycle, analysis of femoral dP/dtmax data reflected changes inV dP/dtmax. Secondly, although LV dP/dtmax reflects myo-ardial contractility, this variable is also influenced by pre-oad and heart rate. Whereas the different determinants ofV dP/dtmax have been well explored, the influence of

Fig 4. Bland and Altman representation of changes in LV dP/dtmax

nd changes in femoral dP/dtmax.

hanges in loading conditions and heart rate on peripherali

d

rtery– derived dP/dtmax remain to be identified. The presentata were obtained at a fixed heart rate regulated by atrio-entricular pacing so this factor did not interfere with theresent observations. In addition, peripheral artery– derivedP/dtmax is an ejection phase index that depends not only oneterminants of LV function but also on arterial compliance.or a given LV dP/dtmax value, the higher the arterial com-liance, the lower will be the ascending slope of the arterialressure curve (ie, femoral dP/dtmax) during the ejectionhase.17 Therefore, further studies, which are the subject ofngoing research, will have to evaluate how peripheral ar-ery– derived dP/dtmax is modified by changes in heart rate,ardiac loading conditions, and arterial waveform patternsncluding the effects of reflected waves. The present study

Fig 5. Relationships between changes in LV dP/dtmax and changes

n femoral dP/dtmax during leg elevation, calcium administration, and

obutamine infusion.

wiLtbpcdorarvwastdotdmida

ccfcdgdfvvovc

REN

r

m

r

pc

ro

dp

effects of �-adrenergic stimulation on the length-dependent regulation

o8

d

pm7

al

t

mp

o

v

c

p

m

s

e

a max

phase variable.

5EVALUATION OF LEFT VENTRICULAR FUNCTION

ARTICLE IN PRESS

as performed in a homogenous group of patients undergo-ng elective coronary artery bypass surgery with preservedV systolic function (as assessed by preoperative LV ejec-

ion fraction). Therefore, these results cannot unequivocallye extrapolated to other types of patients (ie, intensive careatients with sepsis). All factors that may affect arterialompliance (eg, vasoactive agents) may also affect femoralP/dtmax independently of LV contractile function. More-ver, a given change in LV contractile function might beeflected by a greater change in femoral dP/dtmax whenrterial compliance is low (elderly patient with atheroscle-osis) than when it is high (young patient without anyascular disease). In addition, in the present study, patientsith valvular pathology were also excluded. Because both

ortic regurgitation and stenosis also significantly affect thelope of the arterial pressure curve, it cannot be excludedhat in these instances changes in peripheral artery– derivedP/dtmax may behave differently. Finally, in this study, fem-ral dP/dtmax was assessed with the commercially availableranspulmonary thermodilution monitor (PiCCOplus). Thisevice uses a fluid-filled arterial system. Fluid-filled systemsay be less reliable than microtip transducers, although this

s mainly related to the timing of the signals (because of theelay of the signal transmission with a fluid-filled system)nd not to the actual values of the arterial pressures.

Based on the observations from the present study, it isoncluded that in anesthetized patients scheduled for electiveoronary artery bypass surgery with preserved LV function,emoral dP/dtmax underestimated LV dP/dtmax. However,hanges in femoral dP/dtmax accurately reflected changes in LVP/dtmax during various interventions. These observations sug-est that data obtained from peripheral arterial–derived dP/tmax may be helpful in the perioperative assessment of cardiacunction and in the analysis of the effects of therapeutic inter-entions. However, further studies are required to confirm thealue of femoral dP/dtmax to track changes in LV function inther clinical situations and to compare the actual value of thisariable with other ways of assessing myocardial function in

linical practice.

REFE

1. Robotham JL, Takata M, Berman M, et al: Ejection fractionevisited. Anesthesiology 74:172-183, 1991

2. Schiller NB: Ejection fraction by echocardiography: The fullonty or just a peep show? Am Heart J 146:380-382, 2003

3. Little WC: The left ventricular dP/dtmax–end-diastolic volumeelation in closed-chest dogs. Circ Res 56:808-815, 1987

4. Gillebert TC, Leite-Moreira AF, De Hert SG: Relaxation-systolicressure relation: A load-independent assessment of left ventricularontractility. Circulation 95:745-752, 1997

5. De Hert SG, Gillebert TC, Ten Broecke PW, et al: Contraction-elaxation coupling and impaired left ventricular performance in cor-nary surgery patients. Anesthesiology 90:748-757, 1999

6. De Hert SG, Gillebert TC, Ten Broecke PW, et al: Length-ependent regulation of left ventricular function in coronary surgeryatients. Anesthesiology 91:379-387, 1999

7. De Hert SG, Van der Linden PJ, ten Broecke PW, et al: The

CES

f myocardial function in coronary surgery patients. Anesth Analg9:835-842, 19998. Gillebert TC, Leite-Moreira AF, De Hert SG: Load-dependent

istolic dysfunction in heart failure. Heart Fail Rev 5:345-355, 20009. De Hert SG, ten Broecke PW, Mertens E, et al: Effects of

hosphodiesterase III inhibition on length-dependent regulation ofyocardial function in coronary surgery patients. Br J Anaesth 88:779-

84, 200210. Schmidt C, Roossens C, Struys M, et al: Contractility in humans

fter coronary artery surgery: Echocardiographic assessment with pre-oad-adjusted maximal power. Anesthesiology 91:58-70, 1999

11. Michard F: Changes inarterial pressure during mechanical ven-ilation. Anesthesiology 103:419-428, 2005

12. Gillebert TC, Leite-Moreira AF, De Hert SG: The hemodynamicanifestation of normal myocardial relaxation: A framework for ex-

erimental and clinical evaluation. Acta Cardiol 52:223-246, 199713. De Hert SG, Van der Linden PJ, ten Broecke PW, et al: Effects

Fig 6. Simultaneous recording of electrocardiogram (ECG), left

entricular (black curve) and femoral arterial (gray curve) pressure

urves (middle panel), and of respective dP/dtmax tracings (bottom

anel). For the figure, both pressure curves were obtained using a

icrotip transducer in order to prevent differences in timing of the

ignal transduction. This figure clearly shows the physiologic differ-

nces between the LV dP/dtmax, which is a pre-ejection phase vari-

ble, and the femoral dP/dt , which represents in fact an ejection

f nicardipine and urapidil on length-dependent regulation of myocar-

dA

mns

t

fA

df

P

6 DE HERT ET AL

ARTICLE IN PRESS

ial function in coronary artery surgery patients. J Cardiothorac Vascnesth 12:677-683, 199914. De Hert SG, Van der Linden PJ, ten Broecke PW, et al: Assess-ent of length-dependent regulation of myocardial function in coro-

ary surgery patients using transmitral flow velocity patterns. Anesthe-iology 93:374-381, 2000

15. De Hert SG, ten Broecke PW, Rodrigus IE, et al: The effects of

he pericardium on length-dependent regulation of left ventricular M

unction in coronary artery surgery patients. J Cardiothorac Vascnesth 15:300-305, 200116. De Hert SG, Van der Linden PJ, ten Broecke PW, et al: Effects of

esflurane and sevoflurane on length-dependent regulation of myocardialunction in coronary surgery patients. Anesthesiology 95:357-363, 2001

17. Lodato RF: Arterial pressure monitoring, in Tobin MJ (ed):rinciples and Practice of Intensive Care Monitoring. New York, NY,

cGraw-Hill, 1998, pp 733-750