Bongard Frederic , Darryl Sue. Diagnosis and Treatment Critical Care

Подождите немного. Документ загружается.



INTENSIVE CARE MONITORING

197

demonstrate when the catheter has assumed this position.

Indicators of proper tip placement include (1) a decline in

pressure as the catheter moves from the pulmonary artery

into the “wedged” position, (2) ability to aspirate blood from

the distal port (eliminating the possibility of overwedging),

and (3) a decline in end-tidal CO

concentration with infla-

tion of the balloon (produced by a rise in alveolar dead

space). In a patient receiving positive end-expiratory pres-

sure (PEEP), another indicator of correct positioning is an

increase pulmonary capillary wedge pressure less than 50%

of any increase in PEEP. This is so because normal lung and

chest wall compliances are approximately equal at end expi-

ration. Therefore, intrathoracic pressure will increase by 50%

of PEEP, and pulmonary capillary wedge pressure also will

increase by at most 50%. Lung disease (decreased lung com-

pliance generally) will distort this relationship, but almost

always to less than 50%. If the pulmonary capillary wedge

pressure rises more than 50% of the PEEP, repositioning

must be considered.

The pulmonary capillary wedge (pulmonary capillary

occlusion) pressure (PCWP) estimates left ventricular end-

diastolic pressure and thus serves as an estimate of left ven-

tricular preload. Because the pulmonary vasculature is a

low-resistance circuit, the pulmonary artery end-diastolic

pressure in normals is usually only 1–3 mm Hg higher than

the mean pulmonary capillary wedge pressure and has been

used to estimate left ventricular pressure when the pul-

monary capillary wedge pressure is not available—but this is

unreliable and inconsistent in the case of lung disease, pul-

monary hypertension, or tachycardia. Normal values for pul-

monary artery catheter pressures are shown in Figure 8–8.

Pulmonary capillary filtration pressure (P

cap

) is a measure

of the potential difference that drives fluid from the pulmonary

vasculature into the perivascular interstitial and alveolar spaces.

The contribution of hydrostatic and osmotic pressure differ-

ences to fluid filtration is described by Starling’s law. The equa-

tion relating mean PA pressure, PCWP, and P

cap

= PCWP + 0.4 × (PA – PCWP)

Adult respiratory distress syndrome (ARDS) widens the

PA-to-PCWP gradient and increases P

cap

, contributing to

pulmonary edema.

Clinical Applications

A. Pressure Measurements—In most instances, PCWP is

an accurate indicator of left ventricular end-diastolic pres-

sure. Correlation between CVP and PCWP may be poor in

critically ill patients with cardiopulmonary disease because



Figure 8–9. The effect of airway pressure on the pulmonary vasculature is divided into three zones. A pulmonary

artery flotation catheter should wedge in zone III, where there is a continuous column of blood between the pul-

monary capillary and the left atrium. (Reproduced, with permission, from West JB, Dollery CT, Naimark A: Distribution of

blood flow in isolated lung: Relation to vascular and alveolar pressures. J Appl Physiol 1964;19:713.)



CHAPTER 8

198

of differences between right and left ventricular function. In

this group, both absolute values and relative changes in CVP

are unreliable because alterations in the pulmonary vascular

bed affecting the right side of the heart do not equally affect

the left ventricle. This is of particular importance following

pulmonary embolization, which increases right ventricular

afterload without affecting left ventricular end-diastolic

pressure. CVP and pulmonary systolic and diastolic pres-

sures are all elevated, whereas PCWP is decreased because of

the decline in forward flow.

PCWP correlates best with left atrial pressure (LAP) when

the latter is less than 25 mm Hg. However, PCWP will be

lower than LAP in hypovolemic patients, whose pulmonary

vasculature collapses during peak inspiration. When LAP

increases to more than 25 mm Hg—which may occur after

acute myocardial infarction with decreased left ventricular

compliance—PCWP tends to underestimate left-ventricular

end-diastolic pressure (LVEDP). As left ventricular function

deteriorates, the contribution that atrial contraction makes to

left ventricular filling is increased, and LVEDP can be signifi-

cantly higher than PCWP. Several conditions affect the accu-

racy of PCWP as an indicator of LVEDP. In mitral stenosis,

left atrial pressure at end diastole may be significantly higher

than left ventricular pressure. This is diagnosed by the pres-

ence of large v waves on a PCWP tracing. Large left atrial

myxomas also elevate PCWP. Aortic regurgitation produces

an underestimation of LVEDP by PCWP because the mitral

valve closes early despite increasing left ventricular pressure.

Mitral regurgitation results in accentuation of LVEDP

because of backward flow during systole. Pericardial tampon-

ade restricts filling of all four cardiac chambers and results in

an equalization of CVP and PCWP because all such pressures

are under the restrictive influence of the tamponade. Positive

end-expiratory pressure (PEEP) adversely affects the ability of

PCWP to monitor left ventricular preload. High positive air-

way pressures (PEEP >15 mm Hg) can result in pulmonary

vascular collapse, causing PCWP to reflect airway pressure

instead of left atrial pressure (conversion from zone III to

zone I). As with CVP, an esophageal pressure transducer per-

mits calculation of transmural rather than transthoracic pres-

sures. However, because pulmonary compliance is not

disturbed uniformly, the pressure obtained through the

esophageal probe may not correctly reflect the pressure that

surrounds the pericardium.

While PCWP often closely estimates LVEDP, these values

may not accurately reflect true LV preload, which is a func-

tion of LV end-diastolic volume and stretch of myocardium.

Patients with LV hypertrophy, diastolic heart failure, and LV

ischemia have distorted relationships between LVEDP and

ventricular preload. Consequently, a single PCWP measure-

ment may be less helpful than therapeutic trials of volume

loading or diuretics with serial measurements.

B. Mixed Venous Oxygen Saturation—Mixed venous oxy-

gen saturation (S

–

) is obtained from blood from the pul-

monary artery drawn out of the distal port of the catheter.

Pulmonary artery blood should be withdrawn slowly to

avoid inadvertently pulling “pulmonary capillarized” blood

with a misleadingly higher O

saturation.

Mixed venous oxygen saturation is an indicator of sys-

temic oxygen utilization. Normally, peripheral oxygen con-

sumption (

) is independent of oxygen delivery (D

Therefore, as cardiac output and oxygen delivery decline,

peripheral oxygen extraction increases to keep consumption

constant. This results in decreased mixed venous oxygen sat-

uration. Conversely, sepsis may cause a reduction in periph-

eral oxygen consumption, thereby increasing mixed venous

oxygen saturation.

The partial pressure of oxygen in mixed venous blood is

normally about 40 mm Hg, resulting in a hemoglobin satu-

ration of 75%. Oxygen content can be calculated for both

arterial and venous hemoglobin saturations (%Sat Hb) using

the following formula:

= 1.34 × Hb × %Sat + (0.0031 × P

)

If hemoglobin concentration is in grams per deciliter,

oxygen content is expressed in milliliters per deciliter.

Dissolved oxygen (0.0031

× P

) contributes minimally to

oxygen content but may become significant in patients who

are profoundly anemic. The normal arteriovenous oxygen

content difference—C(a–v)

—is 5 mL/dL. Hypovolemia and

cardiogenic shock both increase the difference (>7 mL/dL),

whereas sepsis decreases it (<3 mL/dL). Left-to-right intrac-

ardiac shunts produce a significant step-up in hemoglobin

saturation in the right ventricle and therefore decrease the

C(a–v)

gradient.

Mixed venous saturation can be obtained continuously

from pulmonary artery catheters with integral fiberoptic

oximetry capabilities. Dual oximetry combines mixed

venous and arterial pulse oximetry (Spa

) to provide con-

tinuous estimates of oxygen extraction and intrapulmonary

shunting. From continuous oximetry data, the ventilation-

perfusion index (

QI) can be calculated by the following

equation:

where P

is alveolar oxygen tension, calculated from the

alveolar gas equation.

QI correlates well with shunt (

Qs/

Qt)

over a wide range of parameters and clinical conditions.

C. Complications—Complications of pulmonary arterial

catheterization may occur both on insertion and subse-

quently. The incidence of a pneumothorax with either the

subclavian or internal jugular approach is 2–3%. Catheter

knotting is related to the size of the catheter and the insertion

length. Smaller catheters knot more frequently, as do those

with excessive redundancy in the ventricle. The incidence of

catheter-induced transient right bundle branch block is



Hb Sp P

aO AO

+×− + ×

1 32 1 0 0031

132

.()(. )

bbS PvO AO×− + ×()(. )1 0 0031



INTENSIVE CARE MONITORING

199

between 0.1% and 0.6% and is thought to be caused by direct

trauma to the bundle of His. The incidence increases to as

high as 23% in patients with preexisting left bundle branch

block. Ventricular arrhythmias also may occur, although they

are usually transient and do not require treatment. Other

complications that may occur during insertion include tra-

cheal laceration, innominate artery injury, and bleeding.

Pulmonary artery rupture may occur at the time of place-

ment, as a result of laceration by the catheter tip, or subse-

quently, from overinflation of the balloon in the distal

pulmonary artery. The overall incidence of rupture is sub-

stantially less than 1%. Contributory factors include distal

position of the catheter, decreased vessel diameter (primary

pulmonary hypertension), systemic anticoagulation, and

prolonged balloon inflation. Hemoptysis is usually the pre-

senting sign. The need for complete removal of the catheter

is debatable because the requirements for monitoring are

compounded by the complication. The catheter should be

withdrawn to a more proximal site and the patient posi-

tioned with the affected side down to optimize ventilation-

perfusion relationships. Emergency thoracotomy is required

in rare cases when uncontrolled bleeding occurs.

Air embolism occurs most commonly with tubing

changes and transducer calibrations. Approximately 20 mL/s

of air is required in adults before symptoms appear, with

75 mL/s required to produce hemodynamic collapse and

death. The precipitating cause is mechanical obstruction of

right ventricular outflow by the air embolus. Patients should

be placed in the left decubitus and steep Trendelenburg posi-

tion. This puts the outflow tract in a dependent location and

allows the air to rise in the ventricle. Aspiration of air through

the pulmonary artery catheter has been reported with varying

results. Obstruction of the pulmonary vasculature by the

embolus results in hypoxemia, increased pulmonary artery

pressure, and right ventricular dysfunction. Passage of air

through a patent foramen ovale may cause cerebral emboliza-

tion and stroke.

Thromboemboli may originate from the tip or body of

the catheter and can result in pulmonary emboli. Catheters

left in place for long periods may cause subclavian or jugular

venous thrombosis. Other complications include infective

endocarditis, sepsis, aseptic thrombotic endocarditis, and

rupture of the chordae tendineae.

The avoidance of sepsis from pulmonary arterial

catheters is identical to the protocols for central venous

catheters. Daily sterile catheter care, dressing changes, and

regular rotation of insertion sites are critical to minimize

catheter-related infections.

Binanay C et al: Evaluation study of congestive heart failure and

pulmonary artery catheterization effectiveness: The ESCAPE

trial. JAMA 2005;294:1625–33. [PMID: 16204666]

Harvey S et al: Assessment of the clinical effectiveness of pul-

monary artery catheters in management of patients in intensive

care (PAC-Man): A randomised, controlled trial. Lancet

2005;366:472–7. [PMID: 16084255]

National Heart, Lung, and Blood Institute Acute Respiratory

Distress Syndrome (ARDS) Clinical Trials Network:

Pulmonary-artery versus central venous catheter to guide treat-

ment of acute lung injury. N Engl J Med 2006;354:2213–24.

[PMID: 16714768]

Pinsky MR, Vincent JL: Let us use the pulmonary artery catheter

correctly and only when we need it. Crit Care Med

2005;33:1119–22. [PMID: 15891346]

Sakr Y et al: Use of the pulmonary artery catheter is not associated

with worse outcome in the ICU. Chest 2005;128:2722–31.

[PMID: 16236948]

Shah MR et al: Impact of the pulmonary artery catheter in criti-

cally ill patients: Meta-analysis of randomized clinical trials.

JAMA 2005;294:1664–70. [PMID: 16204666]

Summerhill EM, Baram M: Principles of pulmonary artery

catheterization in the critically ill. Lung 2005;183:209–19.

[PMID: 16078042]



Cardiac Output

The bedside technique of measuring cardiac output by ther-

modilution added a new dimension to ICU monitoring.

When a known quantity of hot (or cold) solution (indicator)

is injected into the circulation, a time-temperature curve

may be produced that allows calculation of flow. The area

under the time-temperature curve is inversely proportional

to cardiac output. This is calculated using the Stewart-

Hamilton indicator dilution equation:

where V

is the volume of injectate (mL) and T

and C

are temperature, specific gravity, and specific heat of

blood (B) and indicator (I), respectively.

Either iced or room-temperature thermal boluses can be

used, although the use of iced fluid slightly improves the

signal-to-noise ratio. Room-temperature boluses generally

are adequate except when cardiac output is very low. For best

results, the difference between the blood’s temperature and

that of the injectate should be at least 12°C, which is easily

achieved with room-temperature injectate. Bolus injection

speed and warming of the indicator as it passes through the

catheter have only minimal effects (±3%). When technique is

optimal, measurement repeatability is within 10%. Severe

cardiac arrhythmias may reduce repeatability and yield

results that may not accurately reflect average cardiac output.

Timing of injection with a particular phase of respiration

(end expiration) improves consistency of measurements.

Excessive patient movement also may produce erratic results.

for D W

= 108.

Cardiac output

V(T T)SC

IBI II

× − ×××



CHAPTER 8

200

Continuous thermodilution cardiac output measure-

ments can be obtained using special pulmonary artery

catheters. In one type, the right ventricular segment of the

catheter warms the blood by a small amount above body

temperature. A sensitive downstream thermistor records the

change in temperature. In another type, blood flow is esti-

mated by how much electric current is needed to maintain a

segment of the catheter at a temperature slightly above body

temperature. Blood flow is directly related, but in a complex

manner, to the amount of energy needed for a constant tem-

perature. These methods correlate well with conventional

bolus thermodilution but may differ in systematic or nonsys-

tematic fashion.

Other Methods to Measure Cardiac Output

A. Indicator Dilution Cardiac Output—This technique

relies on dilution of a colored dye. A bolus of dye is injected

intravenously through a central venous catheter while

peripheral arterial blood is withdrawn. The arterial sample is

continuously passed through a densitometer. The area under

the dye curve is calculated, and a modification of the

Stewart-Hamilton equation is applied. This techniques is

rarely used in the ICU.

B. Doppler Ultrasound—Doppler devices measure ascend-

ing aortic flow and calculate cardiac output. A continuous-

wave Doppler probe is placed in the sternal notch to measure

the velocity of aortic blood flow. A separate A-mode pulsed

Doppler probe is centered in the third or fourth anterior

intercostal space to measure the cross-sectional diameter of

the aortic root. The stroke volume is the product of the

cross-sectional area and the average blood velocity. Cardiac

output is calculated by multiplying the heart rate and the

stroke volume. Potential sources of error include (1) mis-

alignment of the Doppler beam, which produces errors in

measurement of blood velocity, (2) the assumption that the

aorta is circular, and (3) the assumption that aortic blood

flow is laminar. Each of these factors accounts for a cardiac

output error that approaches 15% when compared with

determinations obtained by other means. The difference

between suprasternal Doppler ultrasound cardiac output

and standard thermodilution has been reported to range

from –4.9 to +5.8 L/min. An esophageal probe is now avail-

able that measures descending aortic flow. An insertion

depth of about 30 cm is required to reach the esophageal

“window.” The aortic root is sized using an A-mode pulsed

Doppler, and a single measurement of ascending aortic flow

is performed with a continuous-mode suprasternal probe.

The esophageal probe is then calibrated against the cardiac

output obtained by the suprasternal technique. This method

yields results as good as those obtained with the suprasternal

technique and offers the added advantage of providing data

continuously. A recently developed transtracheal probe uses

a pulsed Doppler probe attached to the distal end of an endo-

tracheal tube. Controlled studies have shown good accuracy,

but extensive clinical experience is wanting.

C. Thoracic Bioimpedance—This noninvasive technique

measures stroke volume by passing a small alternating cur-

rent (2.5–4.0 mA) through the chest at radiofrequency

70–100 kHz. Four pairs of electrodes (one transmitter and

one sensor) are required. Two pairs are placed at the base of

the neck and two at the level of the xiphoid process in the

midcoronal plane. The change in thoracic impedance is due

to blood flow, ventilation, and body movement. Respiratory

variations occur much more slowly than those associated

with blood flow and can be eliminated by the computer algo-

rithm. Similarly, motion artifacts can be rejected by special

circuitry. The majority of systolic blood flow is due to pul-

satile blood flow in the descending thoracic aorta. Stroke vol-

ume is obtained by analyzing the impedance change over a

cardiac cycle. Heart rate is determined at the same time and

multiplied by stroke volume to yield cardiac output. Other

parameters obtained include ejection velocity index, thoracic

fluid index, and ventricular ejection time. The thoracic fluid

index is thought to correlate with extravascular lung water,

whereas ejection time has been used as a parameter of car-

diac function.

Because bioimpedance cardiac output measurements are

noninvasive, they can be repeated frequently. The volume of

electrically participating tissue (VEPT) is critical in determi-

nation of the stroke volume. Although changes in body habi-

tus are included in the nomogram used to calculate VEPT,

small changes can produce significant error. Similarly, elec-

trode placement is important. A 2-cm change in the distance

between the sensing electrodes will produce a 20% variation

in recorded cardiac output. Overall correlation with ther-

modilution in adults has been only fair. The method is fur-

ther limited by inaccuracies caused by dysrhythmias.

Furthermore, readings are difficult during patient move-

ment, including shivering. Cardiac output is overestimated

when preload is reduced; in low-flow states, when inotropes

are required; and with aortic insufficiency. Underestimation

is produced by hyperdynamic sepsis, hypertension, and

intracardiac shunts.

D. Fick Method—Cardiac output may be calculated by relat-

ing oxygen consumption to arterial and mixed venous oxy-

gen saturation using Fick’s equation:

Calculation of cardiac output using Fick’s equation is the

reference with which all other techniques are compared. The

Cardiac output

Ca vO

−×



10()

Cardiac output

indicator dose(mg)

average

×60

cconcentration time×



INTENSIVE CARE MONITORING

201

arteriovenous oxygen content difference requires that a pul-

monary artery catheter be placed to obtain mixed venous

blood. Oxygen consumption is calculated by measuring the

oxygen content difference between inspired and exhaled gas.

A modification of this technique substitutes mixed venous

(rebreathing), arterial P

, and expired volume of CO

for the corresponding oxygen values in the Fick equation.

E. Pulse Waveform Analysis—A noninvasive method that

shows some promise despite varying reported accuracy uses

algorithmic analysis of the arterial pulse waveform. The

waveform bears a certain relationship with stroke volume

but is modified a great deal by the capacitance, impedance,

and other characteristics of upstream and downstream arte-

rial beds. The aortic impedance plays a major role in this

relationship and is likely to differ between patients. Thus

“calibration” of a noninvasive system (or one using an arte-

rial catheter) using the arterial waveform with thermodilu-

tion cardiac output has been necessary. Recently, several

commercially available systems have been reported to have

adequate agreement with conventional measurements.

Sources of Error

Use of correct temperatures and volumes is the most impor-

tant factor contributing to accurate thermodilution cardiac

output results. If the amount of indicator injected is less than

the amount used in calculation, the fall in indicator temper-

ature will be less than anticipated, and the cardiac output

will be falsely elevated. Cardiac output also will be falsely

elevated if the injectate is warmer than that used in the cal-

culation. The latter problem has been largely overcome with

the introduction of new cardiac output computers that

measure the injectate’s temperature and automatically enter

the value into the calculation.

Right-to-left intracardiac shunt will result in loss of the

indicator, causing a falsely elevated cardiac output. Left-to-

right shunts permit recirculation of indicator that has

already passed through the lungs. This produces multiple

peaks in the time-temperature curve that cannot be inter-

preted by the cardiac output computer, resulting in a bad

curve alert. When tricuspid regurgitation occurs, blood and

indicator mix, resulting in prolongation of transit time. The

curve produced has a slow upstroke and decay, thereby

increasing the area underneath it. This causes sporadic read-

ings and underestimation of the true cardiac output.

Derived Parameters

Cardiac output measurements may be combined with sys-

temic arterial, venous, and pulmonary artery pressure deter-

minations to calculate a number of hemodynamic variables

useful in assessing the overall hemodynamic status of the

patient (Table 8–1). Oxygen transport parameters also may

be calculated (see below).

Cholley BP, Payen D: Noninvasive techniques for measurements of

cardiac output. Curr Opin Crit Care 2005;11:424–9. [PMID:

16175028]



Pulse Oximetry

Pulse oximetry affords a noninvasive estimate of arterial oxy-

gen saturation using the change in light absorption across a

Table 8–1. Hemodynamic calculations and normal ranges.

Formula Normal Range

Stroke volume, mL 60–90 ml

Stroke volume index, mL/m

30–65 ml

Cardiac index, L/min/m

2.8–4.2 L/min/m

Systemic vascular resistance, dyne-s/cm

1200–1500 dyn-s/cm

Pulmonary vascular resistance, dyne-s/cm

100–300 dyn-s/cm

CO = cardiac output; BSA = body surface area; BP = systemic blood pressure; CVP = central venous pressure; PAP = pul-

monary artery pressure; PAWP = pulmonary artery wedge pressure.

1000 × CO (L/min)

HR (beats/min)

Stroke volume (mL/beat)

BSA (m

)

CO (L/min)

BSA (m

)

[Mean BP (mm Hg) – CVP (mm Hg)]

× 80

CO (L/min)

[Mean PAP (mm Hg) – PAWP (mm Hg)]

× 80

CO (L/min)



CHAPTER 8

202

vascular bed during the arterial pulse. In the ICU, pulse

oximetry has important uses and has become a standard of

care in many institutions. There are, however, a number of

issues that should be understood and considered with this

monitoring technique. In particular, the reliability of this

method may be limited in patients with severe hypoxemia,

abnormal arterial pulsations, and hypoperfusion of the site

of measurement.

When light of a particular wavelength is transmitted

through a clear solvent containing solute that absorbs light at

that wavelength, the amount of light absorbed is the product

of solute concentration, path length, and the extinction coef-

ficient (determined by the solute and the wavelength). For a

hemoglobin solution, the relative concentrations of oxy- and

deoxyhemoglobin can be determined in a spectrophotome-

ter because the extinction coefficients are different for these

two hemoglobin species at certain wavelengths.

Pulse oximetry uses the beat-to-beat changes in light

absorption through a vascular bed to estimate arterial O

sat-

uration, discarding any nonvariable light absorption by con-

sidering only the difference between peak and nadir light

intensities. The method determines O

saturation by a com-

plex calculation that includes several important assump-

tions. With the use of two light-emitting diodes (LEDs)

producing light in the red and infrared ranges, pulse oxime-

try is able to estimate oxyhemoglobin as a proportion of the

sum of oxyhemoglobin plus deoxyhemoglobin—the so-

called functional oxyhemoglobin saturation. Pulse oximeters

are “calibrated” by comparison of arterial blood oxygen sat-

uration in volunteers to calculated values; the devices use a

“look-up table” to translate the measured proportion to the

displayed saturation. Pulse oximetry is subject to artifact-

caused errors. Movement of the oximeter probe, extraneous

incident light (especially if pulsatile), variations in arterial

pulsation, dependent position, venous pulsations, and other

factors may result in incorrect O

saturation readings. Pulse

oximeters are most commonly transmission pulse oximeters,

in which light is passed through tissue (ear or fingertip) to a

sensor on the opposite side, but they may be reflectance pulse

oximeters, in which light passes through tissue but is

reflected back to a sensor on the same side as the light source.

Validity

The accuracy of pulse oximetry is generally considered good

in the range of normoxia to mild hypoxemia. However, accu-

racy may be suspect during more severe hypoxemia, such as

when arterial O

saturation is below 75%. In this range, dif-

ferences between measured O

saturation and pulse oximetry

saturation range from 5–12%.

A. Patient Factors—Patients in the ICU frequently have

hypotension, poor distal extremity perfusion, and impaired

oxygen delivery—or are being given pharmacologic vasopres-

sors or vasodilators. These factors affect blood flow to the site

of pulse oximetry and vary the contour and intensity of the

beat-to-beat pulse used to calculate O

saturation. Most

devices are programmed to avoid reporting O

saturation

when low perfusion or a poor pulse signal is being measured.

In some of the few studies considering these issues in the ICU,

failure of the pulse oximeter to measure O

saturation was not

infrequent in patients with hemodynamic instability

(12–15%). However, other studies have demonstrated that

some pulse oximeters continue to measure and report O

sat-

uration despite very poor blood flow and severe hypotension.

These results may not be reliable, and there is concern that

pulse oximeter O

saturation under these conditions may be

misleadingly high. Pulse oximeter technology continues to

evolve. The latest-generation devices have improved resist-

ance to motion artifact and low perfusion. These are expected

to be more reliable and accurate in the ICU setting.

B. Abnormal Hemoglobins—The pulse oximeter cannot

measure carboxyhemoglobin nor accurately measure oxyhe-

moglobin in the presence of carboxyhemoglobin. The oxy-

gen saturation displayed is essentially equal to the difference

between total hemoglobin and deoxygenated hemoglobin

(100% – the percentage of deoxyhemoglobin), but the relative

concentrations of oxy- and carboxyhemoglobin are

unknown. Other substances in the blood may or may not

affect pulse oximetry. Bilirubin has little effect on pulse

oximetry; methemoglobin, generated in the presence of oxi-

dizing agents such as nitrites and sulfonamides, usually

increases the difference between functional O

saturation and

oxyhemoglobin, but a sufficiently high methemoglobin con-

centration also may have the peculiar effect of causing the

pulse oximeter to read 85% regardless of other conditions. A

number of dyes such as indocyanine green and methylene

blue also have effects on the accuracy of measurement.

Clinical Applications

Pulse oximetry has widespread usefulness in the ICU, espe-

cially in adjusting inspired oxygen, during weaning from

mechanical ventilation, and in testing different levels of

PEEP, inverse I:E ratio, or other mechanical ventilator adjust-

ments. Other uses include monitoring during procedures

such as bronchoscopy, gastrointestinal endoscopy, cardiover-

sion, hemodialysis, and radiography. Pulse oximetry is par-

ticularly accurate in following O

saturation in patients who

have mild to moderate hypoxemia (O

saturation >75%) but

without severe hypoperfusion or hypotension. It cannot be

regarded as a complete substitute for arterial blood gas deter-

minations partly because of the lack of P

and pH determi-

nations but also because of the relationship between P

and

saturation when the latter is above 90–95%. Results of

pulse oximetry should be interpreted cautiously in patients

with carboxyhemoglobinemia or methemoglobinemia.

Keogh BF, Kopotic RJ: Recent findings in the use of reflectance

oximetry: A critical review. Curr Opin Anaesthesiol

2005;18:649–54. [PMID: 16534307]

McMorrow RC, Mythen MG: Pulse oximetry. Curr Opin Crit Care

2006;12:269–71. [PMID: 16672788]



INTENSIVE CARE MONITORING

203



Airway CO

Monitoring

A disposable colorimetric device that detects CO

can con-

firm endotracheal tube placement and position. When the

device tests positively for CO

, this confirms that the endo-

tracheal tube is in the trachea. However, a negative result is

not as reliable, and alternative means for checking tube

placement must be used.

Continuous airway CO

monitoring uses a rapidly

responding infrared CO

analyzer. Capnography is a continu-

ous display or recording of CO

concentration during each

breath. Other devices may display the end-tidal CO

fraction

or partial pressure (P

ETCO

The infrared analyzer uses an appropriate wavelength of

infrared light for which the CO

concentration is propor-

tionate to the absorption of the light. It has the advantage of

relatively low cost, real-time sampling, reliability, ease of cal-

ibration, and acceptable response time.

Capnography shows a continuous display of expired and, if

desired, inspired CO

concentration or partial pressure. The

expired CO

waveform can give a qualitative assessment of the

degree of ventilation-perfusion mismatching. For example,

the steepness of the slope of the “alveolar plateau” indicates

more severe

Q mismatching because it demonstrates empty-

ing of progressively less well ventilated lung units compared

with a waveform showing a flatter alveolar plateau in a patient

with less severe

Q mismatching. The inspiratory segment

also should be inspected to confirm that the inspired gas is free

of CO

as a result of malfunction of the ventilator’s expiratory

valve or some other component.

End-Tidal and Mixed Expired P

In normal subjects at rest and breathing at a normal tidal vol-

ume and respiratory rate, P

ETCO

is close numerically to arte-

rial P

, with the usual difference between Pa

and

ETCO

0–4 mm Hg (P[a–

]

). In patients with respira-

tory failure, contribution to expired gas from dead space and

high

Q lung units decreases CO

concentration during

expiration and at end expiration. The P(a–

)

becomes

increasingly large, with a strong correlation between

P(a–

)

and the dead space–tidal volume ratio. The

ETCO

should not be used as a substitute for Pa

patients with lung disease. Furthermore, P(a–

)

cannot

be assumed to remain constant in the face of lung disease

and mechanical ventilation.

The mixed expired CO

fraction or partial pressure (P

–

ECO

)

is usually determined from collection of expired gas for several

minutes. This should be distinguished from P

ETCO

sampled at

the end of a single breath. The mixed expired CO

fraction can

be used with Pa

to calculate the dead space–tidal volume

ratio (V

) using the modified Bohr equation:

In addition, if minute ventilation is measured,

where P

is barometric pressure.

Volumetric Capnography

Capnography records CO

concentration against time dur-

ing expiration. If CO

is plotted against expired volume, then

a semiquantative estimate of dead space:tidal volume ratio is

obtained. Equipment for doing volumetric capnography can

be built into mechanical ventilators. Volumetric capnogra-

phy has been used to help diagnose pulmonary embolism

and theoretically should be applicable to other situations

where dead space:tidal volume ratio is useful (eg, weaning,

ARDS, and asthma).

Validity

As described earlier, end-tidal P

should not be used as an

accurate estimate of Pa

. Patients with either worsening of

gas exchange function (increased Pa

) or improvement in

function (decreased Pa

) can have a fall in P

ETCO

.The

former occurs because of an increase in P(a–

)

; the lat-

ter represents a parallel fall in both P

ETCO

and Pa

However, among the relatively few reports involving ICU

patients—most of whom were receiving mechanical ventila-

tion—some indicate that Pa

and P

ETCO

track together rel-

atively well, with mean differences less than 5 mm Hg and no

change in difference during weaning or extubation. In COPD

patients, the difference was considerably higher (as much as 9

mm Hg) but, again, relatively constant. In contrast, other stud-

ies have found that P(a–

)

varies considerably, and while

there was correlation with V

, there was a lack of a con-

stant value for P(a–

)

that would allow “tracking” of

from P

ETCO

alone. In particular, one study demon-

strated that both increases and decreases in P(a–

)

may

result from mechanical ventilator adjustments.

Clinical Applications

Airway CO

monitoring has the advantage of being nonin-

vasive, and studies are available that indicate a decrease in

the number of arterial blood gases obtained when this

modality is used. However, it is clear that the critically ill

patient with respiratory failure will have the largest and

most unpredictable difference between Pa

and P

ETCO

;

in these patients, P

ETCO

is an unreliable estimate of Pa

On the other hand, the difference between Pa

and

ETCO

can be used as a measure of dead space:tidal volume

ratio and therefore as a measure of the severity of gas

exchange derangement.

Although studies are lacking on the benefit of routine mon-

itoring of airway CO

, capnography and P

ETCO

monitoring



V L /min STPD V L/min,BTPS

CO E

0 826(,). ( )=× ×

EECO

Pa P

CO ECO

−



CHAPTER 8

204

have been used in several clinical situations. First, airway CO

monitoring can provide rapid noninvasive assurance of correct

endotracheal tube placement. The capnogram should show

increasing CO

concentration during expiration, and P

ETCO

should be a plausible value. Second, P

ETCO

has been used dur-

ing cardiopulmonary resuscitation as a measure of the effec-

tiveness of artificial circulatory assistance; a very low P

ETCO

suggests that venous blood is not adequately returning to the

central circulation. Third, the combination of arterial and end-

tidal P

provides an estimate of the inefficiency of ventila-

tion (V

). Some researchers have suggested that

P(a–

)

can be used to titrate the optimal amount of PEEP.

While the smallest difference in P(a–

)

has correlated with

the highest degree of tissue oxygen delivery, this measurement

has not proved ideal in all studies. Finally, it has been suggested

that capnography can help in weaning patients from mechani-

cal ventilation, but the predictive value of airway CO

monitor-

ing in this clinical situation is unclear.

Belpomme V et al: Correlation of arterial P

and P

ETCO

in pre-

hospital controlled ventilation. Am J Emerg Med 2005;

23:852–9. [PMID: 16291440]

Kallet RH et al: Accuracy of physiologic dead space measure-

ments in patients with acute respiratory distress syndrome

using volumetric capnography: Comparison with the meta-

bolic monitor method. Respir Care 2005;50:462–7. [PMID:

15807908]

Moon SW et al: Arterial minus end-tidal CO

as a prognostic fac-

tor of hospital survival in patients resuscitated from cardiac

arrest. Resuscitation 2007;72:219–25. [PMID: 17101205]

Verschuren F et al: Volumetric capnography as a screening test for

pulmonary embolism in the emergency department. Chest

2004;125:841–50. [PMID: 16117730]



Transcutaneous Blood Gases

Using transcutaneous blood gas monitors, partial pressures

of oxygen and carbon dioxide may be measured in the tissue

beneath heated skin electrodes. This monitoring technique

has value because it reflects tissue levels, but it cannot yet be

employed as a substitute for blood gas monitoring.

Principle

The Clark electrode, similar to that used in blood gas analyz-

ers, has been modified to be used on the skin surface. The

skin in the area of the electrode is heated to 43–45°C. This

heating is necessary to make the skin permeable to oxygen,

but it has the additional effect of increasing perfusion in the

tissues beneath the probe. Since transcutaneous P

(Ptc

)

reflects the oxygen tension level of the tissue beneath the

probe, values may be affected either by arterial oxygenation

or by systemic and regional perfusion. At relatively normal

cardiac output and with normal regional blood flow, Ptc

values reflect arterial P

values. In adults, the ratio of

transcutaneous-to-arterial P

is normally about 0.8,

whereas in children it tends to be higher. (Values in neonates

may be very close to equal.) However, when either cardiac

output or regional perfusion is decreased, the ratio of

transcutaneous-to-arterial P

is decreased in proportion to

the level of decreased perfusion. Hence transcutaneous oxy-

gen monitoring may be used as a monitor of both oxygena-

tion and perfusion. A low Ptc

value is an indicator that the

patient is either hypoxemic or in a low-flow state (or has

reduced regional perfusion).

Transcutaneous P

(Ptc

) has been measured using

a modified P

electrode attached to the skin surface. In

contrast to Ptc

,however,CO

is more soluble than O

,so

tissue stores of CO

act as a buffer, reducing the dependence

of Ptc

on blood flow and metabolism. In theory, Ptc

should mirror Pa

more closely than Ptc

reflects Pa

and there is no need to heat the skin at the monitoring site.

Newly designed sensors with heaters have been studied

recently. Results have been variable.

Both Ptc

and Ptc

devices should be calibrated

against known P

and P

. Because of the heating of the

Ptc

electrode site, the location must be changed every 4–6

hours to minimize the risk of thermal injury.

Clinical Applications

Many neonatal ICUs routinely employ Ptc

monitoring and

have found good correlation with arterial blood gases except in

patients with severe cardiac compromise. In adults, Ptc

measurement is best used as a measure of tissue hypoperfusion.

A reduction of Ptc

may be an early indicator of low flow, par-

ticularly if pulse oximetry does not indicate severe hypoxemia.

Kagawa S, Severinghaus JW: Errors in monitoring transcutaneous

on the ear. Crit Care Med 2005;33:2414–5. [PMID:

16215403]

Kagawa S et al: Initial transcutaneous P

overshoot with ear

probe at 42

C. J Clin Monit Comput 2004;18:343–5. [PMID:

15957625]



Respiratory Mechanics

Measured parameters are tidal volume, vital capacity, airway

pressure, and intrathoracic pressure. From these, respiratory

system and lung compliance, airway resistance, and work of

breathing can be estimated. Modern mechanical ventilators

often are equipped to measure airway pressure, tidal volume,

inspiratory flow, and other derived values. They may be able

to display in real time flow-volume or pressure-volume

loops. A discussion of respiratory compliance and resistance

is found in Chapter 12.

Tidal Volume

In the ICU, tidal volume is measured most commonly in

patients who have endotracheal tubes and require mechanical

ventilation. Respiratory inductive plethysmography provides



INTENSIVE CARE MONITORING

205

a noninvasive estimate of tidal volume. Tidal volume should

be monitored frequently in patients receiving mechanical

ventilation. When volume-preset modes are used, a difference

in expired volume compared with preset volume indicates

that there is a leak in the ventilator circuit, that inspiratory

flow demand is extremely high, or that inspiratory peak pres-

sure exceeds the preset limit. In pressure-controlled ventila-

tion, tidal volume is used to adjust the level of set airway

pressure, and any change in expired tidal volume indicates a

change in lung or chest wall compliance or airway resistance.

During spontaneous respiration, tidal volume monitoring

using noninvasive measurement can be employed to help

identify patients with obstructive sleep apnea or abnormal

breathing patterns (Cheyne-Stokes respiration).

Maximum Inspiratory and Expiratory Airway

Pressure

Inspiratory and expiratory maximum pressures are deter-

mined by a manometer connected either to a mouthpiece or

to tubing adapted to fit onto the endotracheal tube. These

pressures are measured correctly starting at functional resid-

ual volume so that lung and chest wall elastic recoils are

neutralized, and the pressures reflect only respiratory mus-

cle strength. In practice, this detail is often omitted. Normal

maximum inspiratory pressure is more than –80 to –100 cm

O; maximum expiratory pressure in normal individuals

exceeds 120–150 cm H

O. Maximum negative inspiratory

and positive expiratory pressures are useful in assessment of

respiratory failure in patients with neuromuscular disor-

ders. Studies support the use either of the average of maxi-

mum inspiratory and expiratory pressures or of vital

capacity in roughly predicting the onset of hypercapnia in

these patients when pressures fall by about 70% or when

vital capacity is less than about 55% of predicted. These

direct measures are notably better than extrapolations of

respiratory muscle strength from measurements of the

strength of the extremities.

Intrathoracic (Intraesophageal) Pressure

Intrathoracic pressure requires a pressure sensor within the

chest—almost always a balloon placed in the lower third of the

esophagus and connected to a suitable manometer. The bal-

loon must be filled carefully with a small amount of air so that

the changes in intrathoracic pressure are faithfully recorded

without the confounding effects of the balloon’s compliance.

Care must be taken to position the balloon within the chest

and not in the stomach. Systems for measuring esophageal

pressure are available commercially. Esophageal pressure is

used most often to determine lung or chest wall compliance or

work of breathing, but it also can be helpful in identifying

auto-PEEP. A potential use is to “correct” pulmonary artery or

pulmonary artery wedge pressures for large swings in

intrathoracic pressure during the respiratory cycle.

Lung and Chest Wall Compliance

Calculation of compliance of the respiratory system (chest

wall and lungs together) is reviewed in Chapter 12. The com-

ponents of respiratory system compliance can be subdivided

into chest wall and lung compliances. Lung compliance (C

)

is calculated as the ratio of change in volume (ΔV) to change

in pressure (ΔP), where ΔV is usually the tidal volume and

ΔP is the difference between end-inspiratory and end-

expiratory transpulmonary pressure. Transpulmonary pres-

sure is the pressure difference between the pressure in the

airway and the esophageal pressure. Normal lung compliance

is about 200 mL/cm H

O at end-expiratory volume. If chest

wall compliance (C

) is desired, the formula 1/C

= 1/C

+ 1/C

can be used, where

is the respiratory system).

Decreased lung compliance has been used as a criterion of

ARDS in some clinical studies but is not usually required for

its clinical diagnosis. Low lung or chest wall compliance sug-

gests increased work of breathing and could suggest that

weaning would be difficult or inappropriate. On occasion,

abnormal chest wall compliance as a cause of respiratory fail-

ure is not identified unless measured. If low chest wall com-

pliance is found to contribute to respiratory failure, a very

different approach to treatment may be warranted.

Work of Breathing

The work required for breathing is derived from the for-

mula for mechanical work: W = force × distance. For

breathing in which the force is applied over a predeter-

mined volume, substitute pressure for work, and for dis-

tance, substitute change in volume. Thus work of

breathing is the product of ΔP ×ΔV. Work of breathing can

be calculated for the patient-ventilator system, with the

result indicating the work being done by the mechanical

ventilator for a particular tidal volume. The same work, of

course, would be done by the patient breathing sponta-

neously, but this is usually not relevant because the patient

is unlikely to breathe at the same rate, tidal volume, and I:E

ratio. For spontaneous breathing, tidal volume should be

measured nonintrusively, and the pressure difference

between end expiration and end inspiration is measured

inside the thorax (esophageal pressure). Thus a setup for

measuring mechanical work of breathing generally meas-

ures tidal volume (noninvasive or through the endotra-

cheal tube) and esophageal pressure.

Work of breathing calculated from tidal volume and

esophageal pressure may underestimate actual work being

performed by a significant amount. Some of the difference

arises from nonuniform expansion of the lung or chest wall

or compression of intrathoracic gas, and the calculation does

not consider the expiratory work of breathing. In practice,

work of breathing is rarely measured for clinical purposes.

Theoretically, increased work of breathing should be a good

predictor of the success of weaning from mechanical ventila-

tion and could provide a guide to maximizing respiratory



CHAPTER 8

206

muscle strength, maximizing lung and chest wall compliance,

and minimizing airway resistance. In the area of research in

critically ill patients, work of breathing has been used to

characterize an excessive burden on the respiratory muscles

from high resistance in the ventilator circuit, poorly func-

tioning expiratory and inspiratory valves, development of

auto-PEEP, and other factors.

Bigatello LM, Davignon KR, Stelfox HT: Respiratory mechanics

and ventilator waveforms in the patient with acute lung injury.

Respir Care 2005;50:235–45. [PMID: 15691393]

Durbin CG Jr: Applied respiratory physiology: Use of ventilator

waveforms and mechanics in the management of critically ill

patients. Respir Care 2005;50:287–93. [PMID: 15691397]

Gattinoni L, Eleonora C, Caironi P: Monitoring of pulmonary

mechanics in acute respiratory distress syndrome to titrate ther-

apy. Curr Opin Crit Care 2005;11:252–8. [PMID: 15928475]

Haitsma JJ: Physiology of mechanical ventilation. Crit Care Clin

2007;23:117–34. [PMID: 17368160]

Jonson B: Elastic pressure-volume curves in acute lung injury and

acute respiratory distress syndrome. Intensive Care Med

2005;31:205–12. [PMID: 15605228]

Stahl CA et al: Dynamic versus static respiratory mechanics in

acute lung injury and acute respiratory distress syndrome. Crit

Care Med 2006;34:2090–8. [PMID: 16755254]



Respired Gas Analysis

Measurement of respired gases includes determination of

output (

) and O

uptake (

). In the steady state,

these are considered equivalent to CO

production and O

consumption. The ratio

is the respiratory gas

exchange ratio, which is equal to the respiratory quotient

(RQ) in the steady state.

and

are measured by comparing inspired and

expired gas concentrations of O

and CO

and knowing the

inspired or expired minute ventilation. Gas concentrations

are measured using O

and CO

analyzers. Oxygen analyzers

measure P

using an electrochemical method to generate

current proportionate to P

. Infrared CO

analyzers are

highly reliable and accurate. Both inspired and expired gas

concentrations are needed.

The product of mixed expired CO

fraction (F

–

ECO

) and

expired minute ventilation is

(L/min) expressed at stan-

dard temperature and pressure, dry (STPD). For

(L/min), the calculation is more complex, having to take into

account the inspired oxygen fraction and the small but sig-

nificant difference between inspired and expired minute ven-

tilation owing to the gas exchange ratio.

Automated instruments allow for continuous measure-

ment of

and

and often report these results in tables

or graphs. Indirect calorimetry uses

and an estimate of

the substrate mix being used for energy production (RQ) to

estimate the energy expenditure or caloric requirement of

the patient:

and

also can be measured using the Fick equa-

tion with knowledge of arterial and mixed venous O

and

contents (mL/dL or mL/L of blood) and cardiac output

(thermodilution or other technique). For critically ill

patients, this requires a pulmonary artery catheter, and most

commonly it has been an intermittent calculation. However,

with continuous cardiac output technology, continuous

mixed venous oximetry, and arterial pulse oximetry, a con-

tinuously calculated oxygen consumption is now possible.

Validity

Carbon dioxide output (

) determination is generally

satisfactory in the ICU, including patients receiving mechan-

ical ventilation. With high inspired O

,CO

analyzers require

adjustment according to the manufacturer’s instructions, but

this does not detract from accuracy.

On the other hand,

is calculated most often using the

difference between inspired and expired fraction of O

and

an adjustment that accounts for the small difference between

inspired and expired minute ventilation. This formula is very

sensitive to inspired O

concentration, especially when F

0.50. In fact, some investigators are suspicious of

deter-

minations from expired gas when F

exceeds 0.21, and

many commercial systems are not reliable when F

greater than 0.30–0.40.

Calculation of oxygen consumption by the Fick equation

also has significant inherent inexactness. Intermittent ther-

modilution measurements of cardiac output have signifi-

cant variance, as do measurements of arterial and venous

saturation, as well as hemoglobin. Any errors in these

measurements are increased geometrically during calcula-

tion of oxygen consumption—so-called mathematical cou-

pling. However, the importance of this limitation to the

accuracy of this methodology in clinical practice is a matter

of controversy.



Clinical Applications

Monitoring of oxygen delivery and oxygen consumption

allows titration of ICU interventions to specific physiologic

end points for individual patients. However, it is not clear

that providing “supranormal” oxygen delivery improves out-

come for critically ill patients with shock, sepsis, trauma, or

hemorrhage. One important exception is during the early

phase of sepsis and septic shock, in which attention to oxy-

gen delivery (assessed using central venous oxygen satura-

tion) increases survival.

Oxygen uptake by indirect calorimetry in the ICU also

may be used to determine metabolic requirements so that

appropriate caloric requirements as well as substrate

kcal/day V (L/min) 1440 min/dayO

=×



×+×(3.82 1.23 RQ) kcal/L V