Central Venous Pressure a Useful but Not So.30

8/11/2019 Central Venous Pressure a Useful but Not So.30

1/4

R. Phillip Dellinger, MD, FCCM, Section EditorConcise Definitive Review

Central venous pressure: A useful but not so simple measurement

Sheldon Magder, MD

Central venous pressure mea-surements are frequently usedfor the assessment of cardiacpreload and volume status (1).

This is not surprising, considering theready availability of central venous pres-sure measurements for any patient whohas a central venous line. Central venouspressure can even be estimated in mostpeople by examining the distention of jug-ular veins (2). However, the use of the cen-tral venous pressure is much criticizedbecause central venous pressure poorlypredicts cardiac preload and volume status(35). I argue that the reason for the lack ofappreciation of the usefulness of the central

venous pressure is the failure to considerthe physiologic determinants of the central

venous pressure and potential errors inmeasurement (6, 7).

PRINCIPLES OF MEASUREMENT

Before we assess the physiologic mean-ing of the central venous pressure, somebasic principles of measurement need to

be considered. An important point that isoften not respected is that hydrostaticpressures are relative to an arbitrary ref-erence level, and changes in the reference

level change the measured pressure. The

effect of leveling on the measurement ofcentral venous pressure is particularly

important because small changes in cen-tral venous pressure have large hemody-

namic effects. For example, the normalgradient for venous return is in the range

of 4 mm Hg to 6 mm Hg (8), and thenormal cardiac function curve starts at 0

and plateaus in most people by 10 mmHg. The commonly accepted referencelevel for vascular measurements is the

midpoint of the right atrium, for this iswhere the blood returning to the heart

interacts with cardiac function. As rou-tinely taught to medical students, thiscan be identified on physical examination

at a vertical distance 5 cm below the sternalangle, which is where the second rib meets

the sternum (2). This is true whether thesubject is supine or sitting up at a 60-degree angle because the right atrium is

anterior in the chest and the atrium has arelatively round shape. Thus, a 5-cm verti-cal line from the sternal angle remains in

the approximate center of the atrium evenwhen the person is sitting upright at a

60-degree angle. This means that patientsdo not have to be supine for measurements

when this reference level is used.

More commonly, the mid-thoracic posi-tion at the level of the fifth rib is used in

intensive care units. This is easier to teachbut should be used only for measurementsin the supine position, because this refer-

ence position changes in relation to themid-right atrium with changes in posture.

The greater simplicity of the mid-thoracic

position also likely results in less rigor inproper leveling. Values measured relative to

the mid-thoracic reference level are on av-erage 3 mm Hg greater than those based on

the reference level 5 cm below the sternalangle (9).

A second important principle of mea-surement is that the value of central ve-nous pressure that determines cardiac pre-

load is the central venous pressure relativeto the pressure surrounding the heart, or

what is called the transmural pressure. Thistoo is the source of a lot of measurement

errors (10). The heart is surrounded bypleural pressure, and pleural pressure var-ies relative to atmospheric pressure during

the respiratory cycle, whereas measuringdevices are zeroed relative to constant at-

mospheric pressure. At end-expiration, pleu-ral pressure is only slightly negative rela-tive to atmospheric pressure, and thus

the central venous pressure measuredrelative to atmosphere at this part of thecycle is close to the transmural pressure,

whether the person is breathing sponta-neously or with positive-pressure ventila-

tion. However, in patients breathing withpositive end-expiratory pressure (PEEP),transmural central venous pressure rela-

tive to atmosphere will always overesti-mate the transmural pressure, and thereis no simple way to correct for this prob-

lem. At low levels of PEEP, however, es-pecially in patients with decreased lung

compliance, the effect is small. Further-more, as discussed below, it is really

From McGill University Health Centre, Montreal,Quebec, Canada.

The author has not disclosed any potential con-flicts of interest.

Copyright 2006 by the Society of Critical CareMedicine and Lippincott Williams & Wilkins

DOI: 10.1097/01.CCM.0000227646.98423.98

Objective:To review the clinical use of central venous pressure

measurements.

Data Sources:The Medline database, biographies of selected

articles, and the authors personal database.

Data Synthesis: Four basic principles must be considered.

Pressure measurements with fluid-filled systems are made rela-

tive to an arbitrary reference point. The pressure that is important

for preload of the heart is the transmural pressure, whereas the

pressure relative to atmosphere still affects other vascular beds

outside the thorax. The central venous pressure is dependent

upon the interaction of cardiac function and return function. There

is a plateau to the cardiac function curve, and once it is reached,

further volume loading will not increase cardiac output.

Conclusions: If careful attention is paid to proper measure-

ment techniques, central venous pressure can be very useful

clinically. However, the physiologic or pathophysiological sig-

nificance of the central venous pressure should be considered

only with a corresponding measurement of cardiac output or at

least a surrogate measure of cardiac output. (Crit Care Med

2006; 34:22242227)

KEY WORDS: right atrial pressure; fluid administration; cardiac

output; resuscitation

2224 Crit Care Med 2006 Vol. 34, No. 8


2/4

the hemodynamic response to a changein central venous pressure that is impor-

tant clinically.Although expiration is normally pas-

sive, active expiration is very common incritically ill patients. When expiration isactive, contraction of abdominal and tho-racic muscles increases pleural pressureduring expiration, and there may not beany phase during the respiratory cycle in

which pressure measured from a trans-ducer referenced relative to atmosphericpressure gives a close approximation ofatrial transmural pressure (Fig. 1). Theonly thing that then can be done in this

situation is to examine multiple cyclesand make the measurement in a cycle

where there is minimal forced expiratoryeffort. Sometimes, there is no value thatis satisfactory, and a measurement earlyin the expiratory phase may be a betterestimate than the value at end-expiration,but it is still a guess.

Another important consideration forthe measurement of central venous pres-sure is where to make the measurementin relation to the normal a, c, and v

waves. The a and v waves can often bein the range of 810 mm Hg, which

means that there is a large difference inthe value at the top, middle, or bottom(Fig. 2). The choice is arbitrary and eachpart of the cycle has physiologic signifi-cance. However, for the estimate of car-diac preload, which is the most commonclinical question, the pressure at the baseof the c wave is most appropriate be-cause this is the last atrial pressure before

ventricular contraction and therefore thebest estimate of cardiac preload (11). Ifthe c wave cannot be identified, thebase of the a wave gives a good approx-

imation. Alternatively, if the monitor hasthe capacity, a vertical line drawn throughthe Q wave of the electrocardiogram willhelp identify this position. On the otherhand, if there is a tall a or v, the peakof these waves still has hemodynamicconsequences for upstream organs suchas the liver and kidney. Furthermore, thecentral venous pressure in most dependentparts of the body in the supine position is810 mm Hg higher than that measuredon the basis of 5 cm below the sternal angle

measurement, and this is the pressure thatdrives the local capillary filtration.

The central venous pressure can be es-timated on physical examination by mea-suring the distention of the jugular veinsrelative to the sternal angle. One then adds5 cm H2O to the measured distention toobtain the central venous pressure (12). Toconvert the value of central venous pres-sure in cm H2O to mm Hg, one needs todivide the value in cm H2O by 13.6, whichis the density of mercury compared to thatof water, and multiply by 10 to convert cmto mm Hg (or simply divide by 1.36). It is

worthwhile doing this before inserting cen-

tral lines, for the pressure estimate will tellyou that the value obtained with the trans-ducer is in the appropriate expected range.It also improves ones skills in using the

jugular venous distention to assess centralvenous pressure noninvasively.

DETERMINANTS OF THE

CENTRAL VENOUS PRESSURE

Central venous pressure is determinedby the interaction of two functions: car-

Figure 1.Example of a central venous pressure (CVP) tracing for a patient with forced expiration. Inspand the lines mark inspiration. The pressure rises

throughout the expiratory phase because of transmission of pleural pressure to cardiac structures. Making the measurement an end-expiration will greatly

overestimate the true central venous pressure. The digital value on the monitor will also likely be an overestimate. A reasonable guess is a measurement

early in expiration, before the patient begins to push (arrow).

Figure 2.Example of a central venous pressure (CVP) tracing with prominent aand vwaves. There is a small cwave after the awave, followed by

the xdescent. The appropriate point for measurement is the base of the cwave (or the awave when the cwave cannot be seen). In this example,

the difference between the bottom (the correct position) and the top is 8 mm Hg.

2225Crit Care Med 2006 Vol. 34, No. 8


3/4

diac function, which represents the clas-sic Starling length-tension relationship,and a return function, which defines thereturn of blood from the vascular reser-

voir to the heart (13). Thus the centralvenous pressure by itself has little mean-ing. The central venous pressure in anormal person in the upright posture isusually less than zero (atmospheric pres-sure) with a normal volume and normal

cardiac function (14). However, a low cen-tral venous pressure also can indicate hy-povolemia or can be present in someone

who is hypervolemic (i.e., with increasedreturn function) but has a very dynamicheart. On the other hand, a high central

venous pressure can be present in someonewith a high volume and normal cardiac func-tion as well as in someone with normal

volume and decreased cardiac function.Thus, a central venous pressure measure-ment must be interpreted in the light of ameasure of cardiac output or at least asurrogate of cardiac output, such as ve-nous oxygen saturation or pulse pressure

variations. The situation is similar to theanalysis of PCO2; to properly interpret theclinical meaning of PCO2, one needs toknow the pH.

USE OF THE CENTRAL

VENOUS PRESSURE

Central venous pressure is commonlyused to optimize cardiac preload. How-ever, an essential point is that the cardiacfunction curve has a plateau and when

that plateau is reached, further volumeloading and increasing the central venouspressure will not alter cardiac output.The increase in central venous pressure

will only contribute to peripheral edemaand congestion of organs. The plateau isdue to restriction by the pericardium, orin the absence of the pericardium, thecardiac cytoskeleton. A difficult problemfor managing the care of patients is thatthe central venous pressure at which car-diac filling is limited is highly variable (3,15, 16). It can occur at a central venous

pressure as low as 2 mm Hg (measuredrelative to 5 cm below the sternal angle)but also as high as 18 to 20 mm Hg.However, as a working number, the car-diac function curve will plateau in mostpeople by a central venous pressure of10 mm Hg (1214 mm Hg when themid-thoracic reference level is used) (9).

When the central venous pressure ishigher than 10 mm Hg and there is aquestion of the potential for a volumeload to increase cardiac output, one

should first consider possible reasons forwhy the central venous pressure is higherthan normal. Explanations include chronicpulmonary hypertension, high positiveend-expiratory pressure (whether exter-nal or internal), and some other restric-tive processes.

The gold standard for determinationof whether or not cardiac function is vol-ume-limited is to perform a fluid chal-

lenge and determine whether an increasein central venous pressure results in anincrease in cardiac output. For this pur-pose I recommend that there be an in-crease in central venous pressure of atleast 2 mm Hg, for that magnitude ofchange can be recognized on most mon-itors. For the test to be positive thereshould be an increase in cardiac output of300 mL/min, a value in the range of re-producibility of thermodilution cardiacoutput devices (17). In reality, evensmaller changes in central venous pres-sure should increase cardiac output insomeone whose heart is on the ascendingpart of the cardiac function curve. Con-sider someone in whom the plateau of thecardiac function curve occurs at a central

venous pressure of 10 mm Hg and thecardiac output at the plateau is 5 L/min.The slope of the line connecting the pla-teau to the zero intercepts indicates thatcardiac output should increase by 500mL/min for every 1-mm Hg increase incentral venous pressure, and that is stillan underestimate of the steep part of thefunction curve. Furthermore, the increase

in cardiac output should occur as soon asthe central venous pressure is increased,for on the basis of Starlings law, an in-crease in end-diastolic volume will affectthe stroke volume of the next beat.

If the clinical question is simply todetermine whether the person is volume-responsive at a given central venous pres-sure, the type of fluid used for the fluidchallenge is not important. What is im-portant is to run the fluid in as fast aspossible; the faster the fluid is given, thelesser has to be given. When I am con-

cerned about giving too much volumeunnecessarily, I sometimes use a pres-sure bag to increase the speed of theinfusion, and as soon as the central ve-nous pressure increases by 2 mm Hg, Imeasure the cardiac output.

An interesting approach to a volumechallenge that can avoid extrinsic volumeinfusion is to elevate the patients leg toprovide an autotransfusion and observethe cardiac response (18). Another possi-ble test is to perform a hepatojugular

reflux (12). In this test the abdomen iscompressed and the effect on jugular ve-nous distention is observed. It has beenshown that if jugular venous distentionpersists for more than 10 secs, it is indic-ative of right-heart dysfunction, and al-though this has not been directly studied,it would likely mean that the patient willnot respond to volume.

The important clinical question with re-

gard to fluid responsiveness in most pa-tients should be phrased in the negative: Isit unlikely that this patient will respond tofluids? To this end, examination of thepattern of respiratory variations in the cen-tral venous pressure is useful to predict alack of fluid responsiveness in patients whohave spontaneous inspiratory efforts (15).This examination was also shown to be ef-fective for patients who are mechanically

ventilated but have at least some triggeredefforts. The first step is to determine whetherthere is an adequate inspiratory effort. If thepatient has a pulmonary artery catheterin place, respiratory fluctuations in pul-monary artery pressure give an indicationof the adequacy of the inspiratory effort.If there is no pulmonary artery catheter,simple observation of the patient is oftenadequate. If the central venous pressureas measured at the base of the a wavefalls by 1 mm Hg during inspirationand this is not due to the relaxation ofexpiratory muscles, usually the patient

will respond to fluids, although some pa-tients may not. However, the test is moreimportant in the negative sense. If there

is no inspiratory fall in the central venouspressure and a fall in pulmonary arteryocclusion pressure of at least 2 mm Hg, itis very unlikely that cardiac output willincrease in response to fluids.

The magnitude of the y descent inthe central venous pressure tracing pro-

vides another potential predictor of a lackof fluid responsiveness. In a small series,

we found that no patient with a y de-scent of 4 mm Hg, including the ydescent that occurs during spontaneousinspiration, had an increase in cardiac

output in response to fluids (19). How-ever, some patients with a y descent 4mm Hg also did not respond to fluids;thus, once again, a prominent y descentindicates that the heart is operating onthe plateau of its function curve and theoutput will not increase in response tofluids, but a value less than this does notrule out volume limitation.

Besides the assessment of volume sta-tus, the pattern of change in central ve-nous pressure in relation to a change in

2226 Crit Care Med 2006 Vol. 34, No. 8


4/4

cardiac output can be very useful (as longas there is no major change in pleural orabdominal pressures). If a fall in cardiacoutput is observed, the next question toask is what happened to the central ve-nous pressure, because this allows an as-sessment of the interaction of cardiac andreturn functions. If cardiac output falls

with a fall in central venous pressure, theprimary problem is a decrease in the re-

turn function, which most often is due toa loss of stressed vascular volume; vol-ume infusion is likely the best therapeu-tic approach. If the cardiac output falls

with a rise in central venous pressure, theprimary problem is a decrease in pumpfunction, and therapy should be aimed atimproving pump function.

Note that in all the discussion aboveon fluid challenges I have referred to thecentral venous pressure and not the pul-monary artery occlusion pressure for themanagement of cardiac preload. That is

because the central venous pressure in-dicates where the heart interacts with thereturning blood. Whether cardiac limita-tion is due to a right-heart problem or aleft-heart problem, the right atrium isthe place where cardiac function inter-acts with the return function (6). Fur-thermore, the right and left hearts are inseries, and once the right-heart functioncurve reaches a plateau, changes in left-heart function will no longer affect flow,except if the change in function alters theload on the right heart and thereby altersthe plateau. The expression is no left-sided success without right-sided suc-cess. It is for this reason that I argue thatthe pulmonary artery occlusion pressureshould never be used to optimize cardiacpreload. Similarly, measurements of left

ventricular size by echocardiography alsoshould not be used to assess cardiac pre-load.

A very important distinction that mustbe made is the difference between cardiac

output being volume-responsive and apatientsneed for volume. All the discus-sion so far has considered how to identify

volume responsiveness. The need forfluid is based on clinical parameters suchas the presence of hypotension, the cur-rent use of vasopressors, and even justthe need to establish volume reserves.There is a paucity of data in the literatureto provide a basis for appropriate guide-

lines for the use of fluids for these pur-poses, and empirical studies are neededto provide answers.

CONCLUSIONS

The central venous pressure is thereto be used by the thoughtful clinician,and as long as respect is paid to basicphysiologic principles as well as princi-ples of measurement, in my opinion itcan provide a useful guide to assessmentof cardiac preload, volume status, and thecause of a change in cardiac output and

blood pressure.

REFERENCES

1. Boldt J, Lenz M, Kumle B, et al: Volume

replacement strategies on intensive care

units: Results from a postal survey.Intensive

Care Med1998; 24:147151

2. Bates Guide to Physical Examination and

History Taking. Seventh Edition. Philadel-

phia, Lippincott, 1999

3. Michard F, Teboul JL: Predicting fluid respon-

siveness in ICU patients: A critical analysis of the

evidence.Chest2002; 121:20002008

4. Kumar A, Anel R, Bunnell E, et al: Pulmo-

nary artery occlusion pressure and central

venous pressure fail to predict ventricular

filling volume, cardiac performance, or the

response to volume infusion in normal sub-

jects.Crit Care Med2004; 32:691699

5. Shippy CR, Appel PL, Shoemaker WC: Reli-

ability of clinical monitoring to assess blood

volume in critically ill patients. Crit Care

Med1984; 12:107112

6. Magder S: More respect for the CVP. Inten-

sive Care Med1998; 24:651653

7. Magder S: How to use central venous pres-

sure measurements. Curr Opin Crit Care

2005; 11:264270

8. Nanas S, Magder S: Adaptations of the pe-

ripheral circulation to PEEP. Am Rev Respir

Dis 1992; 146:688693

9. Bafaqeeh F, Magder S: CVP and volume re-

sponsiveness of cardiac output. Am J Respir

Crit Care Med2004; 169:A344

10. Magder S. Diagnostic information from the

respiratory variations in central hemody-

namics pressures. In: Respiratory-Circula-tory Interactions in Health and Disease.

Scharf SM, Pinsky MR, Magder S (Eds). New

York, Marcel Dekker, 2001, pp 861 882

11. Lodato RF: Use of the pulmonary artery cath-

eter.Semin Respir Crit Care Med 1999; 20:

2942

12. Ducas J, Magder SA, McGregor M: Validity of

the hepato-jugular reflux as a clinical test for

congestive heart failure. Am J Cardiol1983;

52:12991304

13. Magder S, Scharf SM: Venous return.In: Res-

piratory-Circulatory Interactions in Health

and Disease. Scharf SM, Pinsky MR, Magder S

(Eds). New York, Marcel Dekker, 2001, pp

93112

14. Notarius CF, Levy RD, Tully A, et al: Cardiac

vs. non-cardiac limits to exercise following

heart transplantation.Am Heart J1998; 135:

339348

15. Magder SA, Georgiadis G, Cheong T: Respi-

ratory variations in right atrial pressure pre-

dict response to fluid challenge. J Crit Care

1992; 7:7685

16. Magder S, Lagonidis D: Effectiveness of albu-

min versus normal saline to test volume re-

sponsiveness in post-cardiac surgery pa-

tients.J Crit Care 1999; 14:164171

17. Magder S: Cardiac output measurement. In:

Principles and Practice of Intensive Care Mon-

itoring. Tobin, MJ (Ed). Chicago, McGraw-Hill,

1997, pp 797810

18. Boulain T, Achard JM, Teboul JL, et al:

Changes in BP induced by passive leg raising

predict response to fluid loading in critically

ill patients. Chest 2002; 121:12451252

19. Magder S, Erice F, Lagonidis D: Determi-

nants of the y descent and its usefulness as

a predictor of ventricular filling. J Intensive

Care Med2000; 15:262269

2227Crit Care Med 2006 Vol. 34, No. 8

Documents

Central Venous Pressure a Useful but Not So.30