CLIN. CHEM. 36/8(B), 1585-1593 (1990)
Through a Glass Darkly: the Lung as a Window to Monitor Oxygen Consumption, Energy
Metabolism, and Severity of Critical Illness
John H. SIe9eI
Medical diagnosis and therapeutic monitoring for critical
illness require adaptation of laboratory analyses to the bedside. These are greatly helped by the modification of physiological and biochemical data-acquisition techniques to increase the number and accuracy of noninvasive variables
that can be obtained from the patient. This paper addresses
the choice of noninvasive measurements and is largely
directed at the assessment of oxygen debt as a measure of
the severity of ischemic and septic metabolic processes.
Especially considered are those noninvasive measures of
cardiorespiratory adequacy, key variables that need to be
considered together with the metabolic function to adequately reflect the patient’s state of accommodation to critical
illness or injury. I describe a noninvasive sensor system
linked to a computer work-station that functions in a pattern
recognition mode to permit classification of patients as to the
type and severity of their physiological adaptation. This
system can serve as a sophisticated bedside monitor of the
severity of the patient’s condition, as a guide to therapy.
AdditIonal Keyphrases: oxygen debt
hypovolemia
drome
shock
sepsis
.
metabolic acidosis
multiple organ failure syn-
In health, oxygen consumption is a closely regulated
phenomenon, oxygen being the critical carbon acceptor in
the generation of energy from metabolic fuel substrates.
However, in conditions of altered metabolic states such as
are seen in patients after injury or during severe sepsis,
oxygen consumption increases, especially when these conditions are associated with an increase in body temperature. Conversely, in hypovolemic ischemic, cardiogenic, or
vasodilatory shock states, where oxygen delivery is restricted by low flow, oxygen consumption decreases, falling
below the oxidative requirements
of the various organ
metabolic processes. The oxygen debt produced results in a
severe metabolic acidosis. The concept of oxygen debt is
illustrated in Figure 1.
Quantifying Severity of Low Flow States by Oxygen Debt
Normal resting oxygen consumption (V02)1 in young
men has been shown to average 140 mL/min per square
meter of body surface area, which is equivalent to approx-
Departments of Surgery, University of Maryland School of
Medicine and Johns Hopkins University Schoolof Medicine; Maryland Institute for Emergency Medical Services Systems. Address
for correspondence: Department of Surgery, University of Maryland School of Medicine, Room T1R68, 22 S. Greene St., Baltimore,
MD 21201.
‘Nonstandard abbreviations: ‘02, oxygen consumption;DV,
pulmonarydispersivebloodvolume;EFx, cardiacejectionfraction;
CI, cardiacindex;TPR, total peripheral resistance; RQ, respiratory
quotient; MOFS, multiple organ failure syndrome; and ARDS,
adult respiratory distress syndrome.
Received January 26, 1990; acceptedJune 7, 1990.
C
so
0 E’
#{149}
‘lii
#{149}vcesslve
02 defIcit
producea
lethal cell lalury
wIth non-recovery
Fig. 1. Diagrammatic representation of the accumulation and repayment of oxygen debt
This represents the decrease in oxygenconsumption occurnng after loss of
perfusionto metabolizing tissues and resulting in the development of an
oxygen deficit. This deficit is defined as the accumulating difference over time
between the oxygendemand(equalto the stable V02 at baseline) and the
actualV02.The deficit therefore is the integral of the decreasein VO, below
the demand over a given period of time. From Siegel et al. (1); reprinted with
permissionfrom Churchill Livingstone, New
York, NY
imately 3.5 mljmin per kilogram of body weight, or 250
mlJmin in the 70-kg textbook man (1). The normal rate of
V02 is reduced by the restriction in tissue perfusion occurring in hypovolemic shock. This decrease in V02 below the
obligatory oxidative needs produces an oxygen deficit (oxygen debt) over time. If resuscitation occurs early in the
process, before a critical level of ischemia has been reached,
there is a rapid repayment of the oxygen debt, with an
overshoot, in which the unmetabolized metabolic acids
(reflected by increases in lactate in plasma and in the base
deficit) are oxidized with full recovery from the ischernic
insult. If the oxygen debt remains unrepaid or is increased
for a longer period of time, critical cellular injuries occur,
with intracellular
organelle disruption and death of the
most vulnerable cells, leaving the individual with significant organ damage-primarily
in the organs with high
oxidative requirements: brain, liver, kidney, myocardium,
and immunologic tissues. These cellular injuries induced
by acute ischemia predispose to later complications
related
to failures of organ function. The specific effects of these
failures of cell and organ function are revealed clinically as
an altered host defense, which predisposes to sepsis, or as a
decompensating organ injury, as is seen in the post-ischemic brain syndrome. Finally, if the oxygen debt is
unrepaid for a long enough period, the debt accumulates to
a level at which sufficient lethal cell injury is present to
prevent recovery, and death may follow, often associated
with the so-called multiple organ failure syndrome (MOFS)
(2).
The relationship of this oxygen debt to the likelihood of
death has been quantified by Crowell and Smith (3) in a
canine model, where they found an LD50 at an oxygen debt
of 120 mLfmin per kilogram. More recently, a precise
quantitative study of the relationship of oxygen debt to the
CLINICAL CHEMISTRY, Vol. 36, No. 8(B), 1990
1585
probability of death has been carried out by Dunham et al.
(4), using a hemorrhagic shock model that more accurately
reproduces the circumstances of patient hypovolemia and
resuscitation. Their data (Figure 2) demonstrate the LDso
to be at an oxygen debt of 113.5 mL/kg, approximately the
same as that by Crowell and Smith (3).
In addition, the Dunham study (4) showed that not only
can an accurate prediction of the probability of mortality
(Pdoath)
be made from the accumulated oxygen debt, but
also if the oxygen debt rate can be ascertained, then the
time remaining from any given oxygen debt until the
estimated LDsc is reached also can be estimated as a guide
to the urgency of resuscitation.
The consequence of the ischemic restriction of oxygen
availability
is to prevent the oxidative metabolism of
energy fuel substrates, most prominently glucose, as well
as various amino acids and lipids. As a result, the magnitude of the oxygen debt is directly proportional to the
degree of the metabolic acidosis, reflected by the magnitude
of the total base deficit (negative base excess), and by the
increase in plasma lactate (4). The relationship between
Pdeath
and base deficit in hemorrhagic hypovolemia also
has clinical importance. Similarly, my coworkers and 1(5),
studying 408 patients with blunt multiple trauma, also
have shown that in humans the probability of death can be
predicted from the magnitude of the metabolic acidosis
(Figure 3).
Moreover, our studies (5) have shown that the severity of
the ischemic shock process (quantified by the base deficit)
interacts with the physiological
severity of any direct
contusive organ injury to the brain (as reflected by a
decrease in the patient’s Glasgow Coma Scale). These two
factors appear to be the primary determinants of outcome
in injured patients. From the extracellular base excess and
Glasgow Coma Scale data obtained at admission, a more
comprehensive linear logistic model also has been developed by which the probability of death can be estimated in
post-trauma patients (5).
Table 1. Oxygen Consumption In Post-Trauma
Patlentsa
Vos
Age range,
years
13-30
31-49
50-65
miJmln per m2
vo2, mL/mmnb
4.28
299
166
3.76
263
98
3.36
235
>65
98
3.19
223
a Based on 645 observations in 165 patients.Data from TacchinoAM,
Siegel JH, EmanueleT, et al., citedby Siegel et al. (1).
b Estimated
for a 70-kg man with 1.73 m2 body surfacearea.
n
283
mL/kg per mm
173
152
136
129
lii.
95.
LINEARLOGISTIC1DEL
95.
‘S-
ee.
M()
*
75.
75.
A
‘S.
‘a.
j
-e.al(BAsE ExcEss)
-
2.S
C 8.8001
ss.
‘a.
LD55
4*.
e
45.
95.
as.
a..
BASE EXCESS.-12.
Is.
“.
h...I....,....a_.L...I....i....I....,....I....i...d
I
‘a.
a.
-a.
-14.
-U.
ADMISSION BASE EXCESS (trlVL)
Fig. 3. Prediction of death (mortality: M, %) from the admission
extracellularbase deficit(negativeexcess, mmol/L)in patientswith
blunttrauma
From Siegelet al. (5). Copyright 1990, American Medical Association; used
with permission
Evaluation of the Post-Injury Hyperdynamic Response from
Measurement of Oxygen Consumption
pDEATH
-
-.0196’Oa_OEBT #{149}.BO816*O2_t(BT2
‘.6657
M’-57 R2.954
F254’565.7
pC.OeOl
In the post-injury
period, trauma induces an altered
metabolic state characterized by a hyperdynainic circulation. The main effect of this response appears to be an
increase in the delivery of oxygen to tissues whose demand
for oxygen consumption has been increased by the sequence of eicosanoid, cytokine, and hormonal
mediator
interactions initiated by the trauma, and its shock-induced
ischemic response (6, 7). The mean magnitude of oxygen
consumption in post-trauma patients is increased (Table 1),
but the limit of the post-injury stress response is age
dependent (1). To a large extent, this age dependence is
undoubtedly related to the decrease in lean body mass that
also occurs as a function of age, as demonstrated by Moore
et al. (8) in their classic studies of the body cell mass.
However, regardless of age, the adequacy of the post-injury
response to a period of traumatic ischemic shock is reflected
in the ability of the hyperdynamic
cardiovascular response
to increase oxygen consumption above the pre-injury baseline to compensate for the altered metabolic response to
injury.
02 OffV(mL’kgT
Fig.2. Predictionof death from oxygen debt in experimental canine
model: D = death, S = survival
From Dunham et a!. (4)
1586 CLINICAL CHEMISTRY, Vol. 36, No. 8(B), 1990
Because the primary function of the cardiac output is the
adequate delivery of oxygen to meet the oxygen-consumption needs of the body, one can monitor the degree of
severity and the rate of recovery from hypovolemic or
low-perfusion shock states by evaluating the oxygen consumption (V02). This can be done indirectly
through the
direct measurement of cardiac output by indicator dilution
techniques, computing the oxygen consumption by multiplying the flow by the arteriovenous oxygen content difference (Ca_2),
adjusted for body surface area or body mass
(Table 2).
Oxygen consumption
also can be measured
directly
through quantifying
the oxygen mass balance obtained
from pulmonary gas exchange. This can be done either on
a breath-by-breath
basis, taking into account the dynamic
flow and compositional changes in the inspired and expired
gas, or through an oxygen-balance methodolog’ involving
use of a mixing chamber to provide a mean V02 over a
period of several minutes. The latter method has an advantage for use in patients who are not intubated or ventilated,
because a flow-through head tent can be used. However,
either of these noninvasive methodologies also can be used
in association with a ventilator at an increased fraction of
inspired oxygen (F102) in patients who are severely ill or
injured, when mechanical ventilation is required for control of respiration.
Relation of Oxygen Delivery to Oxygen Consumption
Under normal conditions, nearly all of the oxygen-carrying capacity of the blood is provided by the hemoglobin.
Thus, when hemoglobin concentration is decreased, the
cardiac output must be increased to permit oxygen delivery
great enough to allow the mandated oxygen consumption
(Table 2). Therefore, the maintenance of an adequate
concentration of hemoglobin becomes a critical factor in
optimizing oxygen delivery at an efficient magnitude of
cardiac output, i.e., one that does not require excessive
cardiac work.
Finally, I emphasize that oxygen consumption is related
not only to the rate of flow throughout the body, which is in
turn a function of the quality of cardiac ejection and the
hemoglobin concentration, but also to the adequacy of the
total blood volume, so that a maximum volume perfusion of
the metabolizing capillary beds with an adequate oxygen
Table 2. Effect of HemoglobIn Reduction on
Circulatory Oxygen Delivery
At a hemoglobin concentration of 100 g/L:
Arterial: 100 g/L x 90% satn. x 1.39a
Venous: 100 g/L x 40% satn. x 1.39
=
=
delivery can be achieved. Studies of the regulation of
oxygen consumption in 165 post-trauma patients of all ages
(1) (Figure 4) showed that the V02 is a function not only of
the cardiac ejection fraction, which sets the magnitude of
cardiac output, but also of the magnitude of the blood
volume, as reflected in the pulmonary reservoir blood
volume (DV). An adequate blood volume not only ensures a
maximization of pulmonary capillary blood flow distribution to improve oxygen exchange, but also allows for
distention and perfusion of the body capillary beds. Figure
4 also demonstrates that a significant negative factor
reducing oxygen consumption is the patient’s age, which,
as noted earlier, has its effect primarily through the decrease of lean body mass. Intercept
(INT) is the unexplained effect.
Quantifying the Effectiveness of Therapy for Shock by
Measuring Oxygen Consumption
As noted earlier, the main function of the cardiac output
is to achieve a magnitude of oxygen consumption at which
all metabolic needs are compensated. Consequently, determination of an adequate stable volume of oxygen consumption can be substituted for quantification of optimum total
blood flow. In various types of pathophysiological
conditions (cardiogenic shock, traumatic shock, or severe sepsis,
with or without liver disease), the patient attempts to
increase flow until a steady-state maximum effective volume of oxygen consumption is achieved (Figure 5) (1).
Thus, in the iatrogenic resuscitation
from one of these
shock states, volume infusion and cardiac inotropic support
are increased until the patient reaches his or her steadystate maximum
oxygen consumption, i.e., a condition that
cannot be further increased by additional support measures. This therapeutic technique also allows the physician
to use the noninvasively
obtained V02 to monitor the
effectiveness of the patient’s hyperdynamic
cardiovascular
response to achieve optimum steady-state physiological
compensation after a major low-flow shock state (Figure 5).
An example of the interaction of the cardiac output and
blood volume in regulating oxygen consumption during
and after human traumatic hypovolemic shock is shown in
Figure 6. This case demonstrates that the use of oxygen
consumption is a better single variable for quantifying the
adequacy of perfusion than is blood pressure or heart rate,
125 mL/L #{176}2
56 milL 02
C2
difference =
69 mL/L 02
For a cardiac output of 4.00 L/min:
V02 + 69 mLJL 02 x 4.00 LJmin = 276 mL/min 02
At a hemoglobin concentration of 70 g/L:
Arterial: 70 g/L x 90% satn. x 1.39
=
88mL/L02
Venous:
70 g/L x 40% satn. X 1.39
=
39mL/L02
Ca_vcj difference = 49 mL/L 02
For a cardiac output of 4.00 L/min:
V02 = 49 mL/L 02 x 4.00 11mm= 196 mL/min 02
Therefore to achieve the same oxygen consumption as with Hb at
100 g/L, the cardiac output must increase:
V02 = 49 mL/L 02 x 5.63 L/min = 276 mL’min #{176}2
#{149}
1.39 mL of 02 carried pergramof hemoglobin at 100% saturation.
Throughout, dissolved
omitted.
Source: Siegel JH. Patternand processIn the evolutionof and recoveryfrom
shock. In: Siege!JH, ChodoffPD,eds. The aged and high risk surgical patient.
New York: Grune and Stratton, 1976:381; used with permission.
a
DY/N’
77.3$
‘NT
a
a,’.
4
I.E..
*0U
+
+
MEANS
VO2/M
AG.,
lBSmI/m’
U.S
DV/M’a
221
EVa a 7S.S%
V02/m’
YRS.
,nU/m’
‘
C..ftlaI.,.
#{149}
.ee #{149}
10’ (EPa’)
naSS4 r’.35
+
.35(DV/m’)
F...’135
AU. AGES
pI,rn,tI.n..a.ty
.Is,.III..,,t
a,
-.7 (AGE) +
SI.2
p.OOOI
ad,.no a ,a.th.1
(p.OOOII
Fig. 4. Relativeeffect,in unitsof I0Sm2, of ejection fraction (EFx),
pulmonary dispersive blood volume (DV/m2), and age on oxygen
consumption per body surface area (V02/m2), as derived from
regression equation of 654 results for 165 post-trauma patients
From data of Tacchino AM, Siegel JH, EmanueleT, eta!., as reportedin Siegel
et al. (1); used with permissionfrom Churchill Uvingstone, New York, NY
CLINICALCHEMISTRY, Vol. 36, No. 8(B), 1990 1587
250
S
200
S
2
0
150
HI
2
0
U
2
N
0,
5.
100
‘4,
‘4
0
54 I,,CI).
50
n
s’
150
TRAUMA +20 InC)
20
F#{149}, 255
13 InICI)
CARDIAC.
p
131n4C11
51Hk$+10,
SEPSIS.
Gaol
0
0
I
2
3
4
CI (CANOIAC
3
5
INDEX
1
8
S
10
11
I/minim’)
Fig. 5. Oxygen consumption:flow relationships in the critically ill:
regression lines derived from 758 studies of patients with cardiogenic shock, post-trauma, or with major chronic sepsis or cirrhotic
hyperdynamic liver disease
Source: same as in Fig. 4; used with permission from Churchill Uvingstone,
New York, NY
cardiovascular
measures used to evaluate
the adequacy of circulation
in ordinary clinical practice.
Figure 6 shows two phases in the physiological pattern of
adaptation to hemorrhage
shock and resuscitation
in a
the conventional
24-year-old man, body surface area = 1.79 m2, who was
involved in a motor vehicle accident. The patient sustained
a ruptured spleen, a fractured pelvis, hemorrhagic hypovolemia, and chest contusion with myocardial depression.
His physiological response pattern, obtained by the use of
invasive measurement techniques, is shown before (Figure
flit) and after (Figure 6B) the institution of adequate
volume resuscitation and cardiac inotropic support to help
him achieve an adequate hyperdynamic response (1).
Figure flit illustrates that the cardiac output was markedly depressed (1.9 L/min) because of a marked decrease in
blood volume. The pulmonary dispersive blood volume (DV
=
223 mL) was lower than the 358 mL (DV = 200 mL/m2)
normally expected for this patient. His hemoglobin concentration was essentially normal at 116 g/L. As a result of his
low pulmonary blood volume, which was compounded by
severe myocardial
depression (cardiac ejection fraction of
only 44%), the normalized
body flow was reduced to a
cardiac index of 1.1 L/min per square meter. Even though
his body tissues showed a maximum attempt to extract
oxygen, an arteriovenous oxygen content difference of 90
mLIL, the oxygen consumption was only 98 mL’min per
square meter and there was a significant base deficit of
-2.9 mmollL. Despite this severe hypovolemic ischemia,
the use of the blood pressure measurement as an index of
severity would have been totally misleading in this patient,
because his blood pressure was 107/85 mmHg, with a mean
arterial pressure of 93 mmHg, owing to a markedly
increased arterial vasoconstrictor
response, as evidenced by
an increased total peripheral resistance (TPR) of 38.11
mN cm - 5_5
-
Pw,SIOtjGlC DI6LTSTI4
MIo#{128}S4,I1tPsrn’
byr-orumOelcIAC CIUTPUIOETP1Zi08Tt0( I STATE CLASS
Dt
8’ 8’84
Ti..
1613$
1Mru.tti
No. I
P#{248}tI.ntA
Mt8
$ hEIST-Si
*q.
24
S.N
Ad..U.d:
V 7.14
8tt.ntn9
PPqatct,n:
UILES
HolpIt
68.6 in
tiVt
66.8kg 859 1.79 a2
Pot C1s 0 Stats SlIt la 9 6.3 Dlspaslue
Vol 223. .1
PN’ISIOLJAIC cQLflV3a1MWSS4,IitEIStTY
D’18-OWJ1IiII CIAC
OUTPUTO(T1tMltOI
I STATECLASS
D.ta
V 5.14
Ii..
II: 9
1i.tu.itta,
Mo.
2
S
NWES * .551975
Ag.
94 Wa
Sa:ii
No.iLtad:
5’ 7/24
Attandt,i
isjaIct.i,
UTLES
11.1gM. U.h In
aI*
05.9kg 192 1.79 .2
Pot C*
A Utats blat To I
8.4 0Ia,astus
Not 459.
.1
PT 0001ES59EIAPN
STWSS SESPGNSE
Pr. 050195055559W
_________________
tUM&1c ocas.
510553
O2
A Stotp:
7.4
o,$ t.#{149}, U.,,
C9 XOGSIILC
.4.91
B Stota
5454
#{149},
2
Ck
4.7
1.59
t,..
05SF oscar
:r
8St.&.
I 09r52
3.3
C Stul..:
is
2.2
9.4
S St.t..
inj.SitePit19rits
.
8.9
C St.t.
lnj.
t.
iais
Day
eiPL
.V.tugs
I
t.._.
JL!JIPB.tCWIJE
ritaP.F
Ibde: 00) (IHII.
rnttr#{248}i
flo2 49.8 P6P, S lD:
N :1OM PPresi.:
M: 15 Io&oIRca: It
Pi,t
Day
Iuqs
.1).
33.9
t
7 I
9ri.
SiteRXial
Ptta,i
(sL130PR7S3C.5
ro’Pt
7.64 #{234}Iu: 7.55 Co-ue: 9.8
aCD2 19.2 PuCII9:27.1 LUD 175
2: 1.914: 6.9 OP:107/*5119: 128 HoCD3:29.7 IOjCO3:
23.5 I.L)2:
I:
1.1 La :11.9 IsP: 93
1PQ31Il PaD?.145.9 PuD? 17.9R
116
Li: 223. SV: 16.2 OfP:17.0
VT!: 1.SS#{248}O2
: 97.9 SaD?: 45.1 Ri
:
U: 359. Ejl:.26
PlP:16.9
EFx: 44 l4gl . Il.
Sill;
664T ; 7
U: 2.51W: 36.9I:22.9
Os Lx: -2.9 tO/VT:
Site
-
I
ktorSE
4
tbde: 00) Tool. inth-nJ
53j
7
1O2
68.6P!P: 6 1Dt 33.9
lOtO P05KPress.: 5#{128}
K) :
boLe:
II
Iota!
Satc: II
: 7.49Co-t2: 3.4
39.6 P+C02:42.6 .52 255
31.2 I*j(03: 32.8 t.521,r3: 55
7.50 4u
0: 9.9Id : 2.9 OP:119’53119 131 ItD3:
1: 4.St.a:3.4 IW:69
179:622 P902.Sl.9
i:4I9.$U:6?.5R:1S.0
B
Ii: 497. Ejl:.22 PlP:12.0
U: 8.2 1(5: 92.3I:27.a
V1!:4.9SdJ2:95.1
(ix:
62 11gb
.
Os Lx:
11.9
1.6
M2: D6.3 AuD? 334
5,02:73.9111
:6.5
68’GT; 0.4
a’Ui: 0.6
IS.OT
34
Fig. 6. Physiological consultation report (dye-dilution output determination) for a 24-year-old man who sustained a motor vehicle accident with
ruptured spleen, fractured pelvis, hypovolemia, chest contusion, and myocardial depression (cardiogenic D state): (A) before initiation01
cardiac inotropic support, and ( at follow-upone day later, after initiation of volume replacement and cardiac inotropic support with dopamine
(13 gig/kg per minute)and dobutamine(11 A9/kg per minute), with therapeutic shift to normal stress response (A state)
From Siegel et a!. (1); used with permissionfrom Churchill Livingstone,NewYork, NY
1588
CLINICAL CHEMISTRY, Vol.36, No. 8(B), 1990
After treatment
was instituted (Figure 6B), the blood
volume increased to a DV of 409 mL, the cardiac contractility was improved by the addition of continuous intravenous inotropic support (a combination of dopamine and
dobutamine), and as a result the cardiac ejection fraction
was increased to 82%. As a consequence of both the increased blood volume and the improved myocardial function, the cardiac output rose to 8.8 L/min (cardiac index =
4.9 L/min per square meter). The increased body perfusion
resulted in an increased oxygen consumption to 165 mL/
mm per square meter, which is close to the age-corrected
post-stress mean value expected for this patient (Table 1).
This was achieved at a relatively normal arteriovenous
oxygen content difference (34 mL/L). The combination of
more adequate perfusion with an improved oxygen consumption corrected the base deficit. Of considerable clinical
interest with regard to the use of blood pressure as a
measure of hypovolemic shock, his high diastolic blood
pressure despite the ischemic shock state was actually
lowered to 118/53 mmHg and the mean arterial pressure
was reduced to 69 mmHg, with the decrease in TPR to 6.28
mN - cm s. This decrease in peripheral vascular resistance reflects the abolishment of the previous pathophysiologic vasoconstrictor response to trauma as flow and blood
volume were increased.
Because oxygen consumption can be adequately measured noninvasively,
this type of therapeutically induced
major transition
from
hypovolemic ischemic shock to a
state of physiological adequacy clearly can be monitored
equally effectively by the indirect invasive measure of
cardiac output and arteriovenous oxygen content difference
used here, or by the direct noninvasive measurement of
oxygen consumption by using the inspired-expired
oxygen
mass balance obtained from the pulmonary gas exchange.
Use of V02 and VCO2 to Monitor Altered Metabolic Fuel
Utilization in Sepsis
The physiological response in the post-injury period is
manifested
by a hyperdynarnic
cardiovascular
state, which
becomes greater in patients who develop severe sepsis or
MOFS (9, 10). As part of this normal post-stress response,
oxygen consumption rises and remains increased, or increases further
as sepsis develops (9). However, when
sepsis occurs, the metabolic fuel control is apparently
altered, so that there is a relative inhibition
of glucose
Table 3. MetabolIc
vax,
Group
n
mL/mln per m2
A: In sepsis, all cases, regardless of full mixture
Nonseptic
128
133 ± 27
Septic
246
144 ± 19#{176}
B: Effect of lipids and sepsis
Nonseptic (no lipid infusion)
94
136±28
Septic (no lipid infusion)
132
142±19
oxidation
by peripheral
tissues (11, 12), even though the
body’s total glucose turnover and uptake may be greatly
increased (12-14). Under septic conditions, lactate concen-
trations in skeletal muscle and plasma increase even in the
presence of adequate oxygen delivery and high oxygen
consumption. This increase is secondary to a post-insulinreceptor inhibition of glucose oxidation at the point of
control of pyruvate decarboxylation by pyruvate dehydrogenase (14, 15). This makes the septic increase in plasma
lactate different from that seen in low-flow perfusion states,
where decreased oxygen delivery rather than altered metabolic fuel control is the limiting factor in oxygen consumption (16).
At the same time as this alteration in glucose oxidation
is induced, the down-regulation of pyruvate dehydrogenase
activity shifts the pattern of metabolic fuel utilization, so
that lipid fuels and branched-chain keto acids are oxidized
in preference to glucose, especially in skeletal muscle and
heart tissue (9, 11, 17). During the adaptation to sepsis, the
ratio between oxygen consumption and carbon dioxide
production is altered so that the respiratory quotient (RQ)
decreases (11). This alteration in the metabolic utilization
of oxygen (V02) and in CO2 production (VCO2) in septic
patients is shown in Table 3A. Septic patients have significantly greater oxygen consumption than do nonseptic
patients, but with no significant change in carbon dioxide
production. Thus the RQ decreases.
Analyzing the change in the RQ over the entire range of
caloric intake (Figure 7) indicates that, although there is a
wide spread in the data, the mean RQ rises as total calories
are increased. However, for any given amount of calorie
intake, the septic patients had a lower RQ than the nonseptic patients, this difference in the VCO2/V02 relationship being highly significant (11). Table 3B summarizes the
preferential utilization of lipid fuels in sepsis. Compared
with nonseptic patients, in whom lipid infusion produces
little or no significant V02 effect, in septic patients the
preferential utilization of administered lipid fuels increases
oxygen consumption as well as carbon dioxide production,
while maintaining
the lower VCO2/V02 ratio. As a result,
in sepsis the RQ generally remains <0.9, unless extremely
large quantities of calories are given, in excess of metabolic
needs. Under those conditions, a net lipogenesis occurs that
is in excess of the increased lipid oxidation capacity of the
peripheral muscle and heart (11). The mass balance of C02,
Gas Exchange
vcoz,
per m2
mIJmIn
34
114
124±18
151 ± 14
0.98 ± 0.12
0.87 ± 0.10#{176}
129 ± 29
126
±
19
130 ± 29
123 ± 19
(fj9)b
Nonseptic (lipid infusion)
Septic (lipidinfusion)
RQ
0.97 ± 0.12
0.87 ± 0.10
(0.08)
125 ± 22
134 ± 19
(0.001)
1.01 ± 0.12
0.89 ± 0.09
(0.0001)
(0.001)
nonsepticgroup (P’zO.OOOl; Student’s t-testof differences betweenmeans).
#{176}Significantly
different from nonseptic group: P value(as indicatedby Scheff#{233}
S method of simultaneous inferencefor all contrasts in the analysis of variance)
a Significantly
differentfrom
given in parentheses.
Adaptedfrom Nanni et al. (11).
CLINICAL CHEMISTRY, Vol. 36, No. 8(B), 1990
1589
RO= O.000141cAL/$so) -0.09
No.380
r’o..304
(sosis)
+ 0.878
dynamic septic MOFS. This abnormality in V02 is accompanied by a metabolic acidosis with an excessive increase
in plasma lactate (to >5 mmolJL) and in the metabolic base
deficit (to exceed -6 mmoIJL), despite a high cardiac
output. These features can also serve as quantitative
indi-
0
F3,357o.78 p.0.0001
0
#{149}
8
0
00
I
SEPTIC
ON0NSEPTIC
TOTAL
CALOfllESjfn’BSA
tauc000 *
Fig. 7. Effect of increase in daily calorie input on respiratory
quotient (RQ) in septic and nonseptic patients
Note that for any given calorieamount, the ROis lowerin patients with sepsis.
Covarianceregressionequation is also shown. BSA, bodysurfacearea.From
Nanni ot al. (II).
Copyright 1984, Williams & Wilkins; used with permission
as a measure of VCO2, can also be obtained noninvasively
from the pulmonary
gas exchange and the RQ can be
computed as a means of quantifying shifts in metabolic fuel
utilization
and caloric needs compatible with the posttraumatic hyperdynamic state or with the transition to a
septic response.
Altered
Oxygen
Consumption
in the Septic Multiple Organ
Failure Syndrome
Finally, in severe end-stage sepsis, there is evidence of
impairment
of a wide range of oxidative
processes in
association with the development of MOFS (9,10). The
end-stage metabolic failure associated with MOFS appears
to involve marked abnormalities in peripheral muscle
metabolism, with greatly accelerated proteolysis (septic
autocannibalism) (18) as well as severe impairment in
skeletal muscle utilization of glucose as an oxidative fuel
and profound alteration in hepatic gluconeogenesis (15),
ketogenesis (19), and lipogenesis (20). Conjugative detoxification processes in the liver are inhibited, and there is a
marked reprioritization of hepatic acute-phase protein synthetic function (21). Under these circumstances of septic
MOFS, actual oxygen consumption may decrease, despite
the maintenance of a persistent exaggerated cardiovascular hyperdynamic state. This deterioration in the V02 in
the presence of a pathophysiologic high-flow/low-resistance
state (the septic B state) is the characteristic feature of
end-stage septic MOFS (9, 10).
This change also can be monitored noninvasively
by
demonstrating a decrease in oxygen consumption when the
nomnvasive measures of cardiac output are showing an
excessive increase in flow and an associated widened pressure pulse dynamic, together with the characteristic physical diagnostic features of hyperdynamic skin perfusion.
The likelihood of a transition.towards
septic MOFS can also
be suggested by a progressive increase in lactate and
triglycerides in plasma, accompanied by a decrease in base
excess to negative values while the VO2 remains high or
normal and the RQ is <1.0. High probability of a fatal
septic deterioration is suggested when oxygen consumption
becomes insufficient to meet the metabolic needs in hyper1590 CLINICAL CHEMISTRY, Vol. 36, No. 8(B), 1990
cators of the severity of the organ failure process and, in
some cases, as predictors of outcome. Thus, the oxygen
consumption and carbon dioxide production, as well as the
relationship between them, are noninvasive information of
considerable clinical value. Measurements of these variables can be used to assess the presence of a hyperdynamic
septic shock state as well as the likelihood of an altered
nutritional fuel utilization characteristic of sepsis and the
pathophysiological deterioration of the septic MOFS. Also,
by using a modification of Weir’s methodology (22) with
these determinations, the metabolic utilization of carbohydrates, fats, and proteins for oxidation can be determined.
Consequently, by utilizing the measurement of oxygen
consumption and the change in RQ, one can follow the
degree of severity of the septic process and quantify
the
metabolic fuel needs, as guides to the magnitude, quality,
and adequacy of nutritional support measures.
Abnormalities of Oxygen Exchange in the Adult Respiratory
Distress Syndrome
In cases of severe respiratory insufficiency following
injury or sepsis, the severity of the post-traumatic lung
syndrome known as adult respiratory distress syndrome
(ARDS) can be quantified by computing the alveolar to
arterial
oxygen gradient, normalized
by the arterial oxygen tension (A - a/pao2), at a known fraction of inspired
oxygen. This so-called respiratory index (RI), when compared with the percentage
of pulmonary
veno-arterial
mixture
(sp/t),
provides an index of the availability of
ventilatory alveolar oxygen delivery to the perfusing blood
whereby the magnitude
of the respiratory
exchange dysfunction can be quantified (23,24). Figure 8 shows that not
only the respiratory index but also the slope of the respiratory index topo
shunt relationship [RJJ(sp/t)]
are higher in those patients who developed severe, eventually fatal ARDS, compared with those patients who had
only septic or post-traumatic hyperdynamic
states without
ARDS, or those with minimal ARDS who survived (24).
The respiratory index also can be estimated noninvasively
from transcutaneous
Pa measurements (Tc02), adjusted
by an empirically derived regression that relates the Tc
to the Pa02 at a known fraction of inspired oxygen (25). The
respiratory index relationship
can be used not only to
quantify the severity of the ARDS process, but also to
predict the likelihood
of death and severity of ARDS in
post-trauma and septic patients (24).
Use of Pattern-Recognition
Quantifications
Techniques in Noninvasive
of Human Pathophysiologic
States
The use of pattern recognition techniques to facilitate the
interpretation
of noninvasive measurements is shown in
Figures 9 and 10. Figure 9 shows the graphic output of an
on-line computer-based system connected to a set of totally
noninvasive sensors used to develop a physiological pattern
that is characteristic of various cardiogenic states.
Panel A illustrates a pattern for either hypovolemic or
cardiogenic low-flow states (cardiogenic D state). This pattern can be compared with that manifested by a patient
with a compensated high-V02 post-traumatic or compensated septic state (septic A state, panel B) or that charac-
RI
.1065
-
11
-
756
82
CI
-1.898 (ARDS-0) #{149}.1211(Qsp’Qt)(ARDS-D)
-.923
(Dsp’Ui.)
.744
F32
-
594.0
p < .0801
5
X
0
FVR
-0.68
ARDSDIED
NON-REDSAND
EROS SURUI(D
xx
X
LOAD
36.2
x
x xl
C
S
WA
2285
I(ANARDS-D
S
JLFI)NARV
SHUNT(Osp’Qt)
Fig. 8. Respiratory index/pulmonary shunt relationships in posttrauma and septicARDS, based on 756 samples from 240 patients
The slope of the physiological relationship betweenthe RI and the sp/t
in
EVA
0.64
ARDS deaths (ARDS-D) is significantly different from the combined slope of
non-ARDS and ARDS survivors (ARDS-S). Regression equation as shown.
From Laghi et al. (24). Copyright 1989, Williams & Wilkins; used with
1.080
78.4
permission
TPR
919
teristic of a decompensated hyperdynam.ic septic MOFS
(septic B state, panel C). In Figure 10 the pattern seen in
severe ARDS respiratory insufficiency (C2 state) is shown
together with the individual sensor measures over time, as
the prototype changes successively from the D to A to B
state patterns shown in Figure 9.
In this noninvasive system, oxygen consumption is
measured by a direct oxygen-consumption monitoring
device (Delta Track; SensorMedics, Yorba Linda, CA); the
transcutaneous Pa07 (Tc0) and PaCO, (Tcc02) are also
measured noninvasively (ensorMedics), and these skin
measurements are corrected by use of the algorithms previously described, so that a noninvasive RI can be computed (25). Measures of blood pressure are obtained noninvasively from an oscillometric cuff device, and heart rate is
derived from an electrocardiogram. Cardiac output and
cardiac ejection velocity are estimated from impedence
cardiography (NCCOM), which is corrected by an empirically derived correlation between the noninvasive impedance measures and invasive determinations of cardiac
output with use of cardiogreen dye dilution (Siegel et al.,
unpublished data). In intubated patients, one can also
obtain the ventilatory flow dynamics and airway pressures
(26).
The physiological picture of a specific patient’s adaptive
pattern is compared with the prototype states (shown in
Figures 9 and 10), which have been derived from a previous
analysis of multivariable data from a large number of
critically
ill patients classified by their disease pattern (10,
27, 28). The likelihood or similarity of the studied patient
to each of these prototype states is compared by using a
covariance matrix derived from a bayesian statistical analysis (29). Because the physiological pattern presentation is
in terms of standard deviations from a normal resting state
(R state), one patient can be compared with another, or
with his or her own previous pattern as a state change
occurs, and the likelihood of the patient being in a partic-
EQ
0.80
rvp
0.48
LOAD
83.6
IPR
663
EQ
0.75
IVE
VCO2I
0.81
ZFIO2:
*
50
PaC02
16.00
39.0
1
state
diagrams, obtained from noninvasive
sensors, of (A) prototypic cardiogenic D state, (B) prototypic normal
Fig. 9.
Computer-generated
stress response A state seen after injury and in compensated sepsis,
and (C) prototypic decompensated septic metabolic failure response
B state seen in septic MOFS
All variables are normalized by the mean and standard deviationof that
variable in the resting normal control patient state (A state).The mean (0
standard deviations) is the fourth circle in from the outermost circle, so that
increases, or decreases, from the resting control state can be clearly seen in
standarddeviationsfrom A state. A: In the 0 state,low cardiac index (Cl), V02,
and VCO2 are associated with increases in heart rate (HA) and in TPR, and
with decrease in cardiac ejection velocity (dzidt) and cardiac stroke load
(LoAD). & In the A state there is a high cardiac index, HR. V02, and VCO2;
reducedTPR; and increased cardiac ejection velocity and LOAD.In sepsis the
relation between V02 andVCO2shows a faNin RO. C In the B state there is
a high cardiac index and HR, but decreased V02 with pathophyslologically
low
TPR, and an increase in RI accompanied by a decrease in pulmonary
compliance (coMm.). Cardiac ejection velocity is hlgh because of reduced
peripheral resistance (TPR). The relation between V02 and VCO2 showsa
marked septic decrease in RO
CLINICALCHEMISTRY,Vol.36, No.8(B),1990 1591
ID I:
34553
AGE: 42
PILE 98.0 Kg BSA2. 16
1/15/1990
16:46: 4
.i_
saa
160.
------
DSP
EVA
0.26
LOAD
66.6
TPR
558
RO
1.00
tz/dt
/FIO2:
VCO2I
V#{128} PaCO2 1’lS
C2 state
10.90 55.0
70
.1..2.
II
VCO2
-,-\--
iii::::: :111
lyE
#{163}8
#{149}
1.-.--38
49
-
PaCO2(reg)
10
15
25
25
CallouSe)
5
19
15
29
25
(utnut.
Prototypic end-stage respiratory failure (C2 state) seen in a patient (real name not used) with severe ARDS shown with the time
sequence of previous transitions from normal (R state)to D state, to A state, to B state,
and to early respiratory failure C1 state beforethe C
state transition shown in the circle diagram; note that each physiological state is characterized by a unique shape, pattern, and color
In the C2 statethere is a high cardiac index (Cl)andheartrate(HA) with increasedV02 and VCO2,but decreasedlung tidal volume (TV) and a markedly reduced
pulmonarycompliance (coMm.). The Al is increasedand the p,0, (correctedfrom Tc) falls despite a very high fraction of inspired oxygen (F102).VCO2is normal
or high, with a reduced TV and a decreased minute ventilation (VE), andthereis a markeddisparity in respiratory gas exchange, as evidenced by an increased
Al. The degree of respiratory insufficiency has reached the point at which CO2 elimination is impairedand
(corrected from Tcco2) increases. ThisC2state
demonstrates failure of gas exchange for both oxygenandCO3at high V02 and VCO2
Fig. 10.
ular adaptive state (A, B, C, or D) can be quantitatively
assessed. By utilizing
both pattern and color changes to
communicate this information,
the attending physician,
resident, or nurse can immediately recognize a significant
transition from one state to another, so that corrective
therapy can be undertaken and its results immediately
evaluated. By using this computer-based pattern-recognition methodology, therapeutic modalities to increase flow
and tissue perfusion and to enhance cardiac contractile
adequacy can be assessed with regard to their success or
failure in improving oxygen consumption, and the overall
result can be quantified by the patient’s transition from a
poorer to a better set of physiological values at steady state.
The Lung as a Window for Noninvasive Viewing of
Physiological Adaptation in Critical Illness
In summary, in hypovolemic shock and in low-flow cardiogenic failure states, the measurement of V02 can provide a quantitative
guide to the adequacy of tissue perfusion, and the likelihood of an oxygen deficit can be evaluated by comparing the noninvasively obtained V02 with
the normally expected stress-increased
values of V02, and
with the determinations of plasma lactate and base excess.
In critically ill hyperdynamic
patients managed in the
post-injury
or septic phases of their disease process, the
adequacy of and the need for disease-specific
types of
1592 CLINICAL CHEMISTRY, Vol. 36, No. 8(B), 1990
nutritional
support can be evaluated by analysis of V02
and VCO2 and the relationship between them, as manifested in the RQ. The measurement of oxygen consumption
gives an indication of the total caloric need, and the CO2
production and RQ indicate the required nutritional
fuel
mixture.
The relationship
between oxygen consumption
and CO2 production provides an estimate of the caloric
utilization
of carbohydrates and lipids, and the magnitude
of protein oxidation can be estimated from a modification of
the Weir equation (22), if the daily urinary urea nitrogen
excretion is also obtained. Thus, from the dynamics of
oxygen consumption and carbon dioxide production, giving
the qualitative aspects of metabolic fuel utilization, and the
measures of the effectiveness of lung gas exchange (quantified by the respiratory index), the state of physiological
adaptation of critically ill patients can be quantitatively
assessed by a set of nomnvasive sensors whose output is
analyzed in real time. From these data, and with use of a
diagnostic pattern-recognition
technique implemented on a
bedside computer work-station,
appropriate therapeutic
measures can be devised and their response quantified on
an ongoing basis, so that appropriate tailoring of cardiodynamic, respiratory, and metabolic therapies can be done to
fit the patients’ needs.
In conclusion, to succeed in critical care, as to be successful in business, one must follow Marshall Field’s law: “Give
the customer [the patient] what she wants [physiologically
as well as psychologically]”.
To understand these physiological requirements
[wants], the physician must have a
rapid and dynamic means of observing the critical variables and a methodology to facilitate the recognition of the
patient’s specific pattern of pathophysiological
adaptation.
The newer developments in noninvasive sensor methodology and computer data analysis make it possible to utilize
the respiratory gas-exchange data as well as transcutaneous blood gas and cardiodynamic
measurements in a realtime information-generating
system. From these data,
processed into an information-rich
format, a physiologically compatible therapeutic program can be devised and
its effects monitored
to achieve the optimum possible response.
Supported in part by Contract NAS9-17186 from the National
Aeronautics and Space Administration.
References
1. Siegel JH, Linberg SE, Wiles CE. Therapy of low-flow shock
states. In: Siegel JH, ed. Trauma: emergency surgery and critical
care. New York: Churchill Livingstone, 1987:201-84.
2. Siegel JH. Abnormal glucose regulation, alternative fuel utilization and muscle catabolism in sepsis and the multiple organ
failure syndrome. In: Fry DE, ed. Multiple organ failure: pathogenesis and management. Chicago: Year Book Publ., 1990 (in
press).
3
Crowell JW, Smith EE. Oxygen deficit and irreversible hem-
orrhagic shock. Am J Physiol 1964;206:313-6.
4. Dunham CM, Siegel JH, Weireter L, et al. Oxygen debt and
metabolic acidemia as quantitative predictors ofmortality and the
severity of the ischemic insult in hemorrhagic shock. Crit Care
Med 1990 (accepted).
5. Siegel JH, Rivkind Al, Dalal SM, Goodarzi S. Early physiologic predictors of injury severity and death in blunt multiple
trauma. Arch Surg 1990;125:498-508.
6. Rivkind Al, Siegel JH, Guadalupi P, Littleton M. Sequential
patterns of eicosanoid platelet and neutrophil interactions in the
evolution of the fulminant post-traumatic adult respiratory dis-
tress syndrome. Ann Surg 1989210:355-73.
7. Pomposelli JJ, Flores BA, Bistrian BR. Role of biochemical
mediators in clinical nutrition and surgical metabolism. J Parenteral Enteral Nutr 1988;12:212-18.
8. Moore FD, Olesen KH, McMurrey JD, et al. The body cell mass
and its supporting environment. Philadelphia: WB Saunders Co.,
1963:535 pp.
9. Siegel JH, Vary TC. Sepsisabnormal metaboliccontrol, and
the multiple organ failure syndrome. In: Siegel JH, ed. Trauma:
emergency surgery and critical care. New York: Churchill Livingstone, 1987:411-501.
10. Siegel JH, Ceri-a FB, Coleman B, et al. Physiological and
metabolic correlations in human sepsis. Surgery 1979;86:163-93.
11. Nanni G, Siegel JH, Coleman B, et al. Increased lipid fuel
dependence in the critically-ill septic patient. J Trauma
1984;24:14-30.
12. Shaw JHF, Klein FS, Wolfe ER. Assessment of alanine, urea
and glucose interrelationships in normal subjects and in patients
with sepsis with stable isotopes tracers. Surgery 1985;97:557-68.
13. Wolfe RH. Nutrition and metabolismin burns. Critical care:
state of the art. Soc Crit Care Med 1987;7:19-59.
14. Vary TC, SiegelJH, Nakatani T, et al. Effect ofsepsison
activity of pyruvate
dehydrogenase complexinskeletal
muscleand
liver.
Am J Physiol1986;250(Endocrinol
Metab 13):E634-40.
15. Vary TC, Siegel JH, Zechnich A, et al. Pharmacological
reversal
of abnormal glucoseregulation,
BCAA utilization
and
muscle catabolismin sepsisby dichloroacetate.
J Trauma
1988;28:1301-1.
16. Vary TC, Siegel JH, Rivkind Al. Clinical and therapeutic
significance of metabolic patterns
of lactic acidosis. Pei-spect Crit
Care 1988;1:85-132.
17. Sans RM, Jolly WW, Harris BA. Studieson the regulation of
leucine catabolism.J Mol Cell Cardiol 1980;12:1-16.
18. Cerra FB, Siegel JH, Coleman B, et al. Septic autocannibalism. Ann Surg 1980;192:570-80.
19. Vary TC, SiegelJH, Nakatani T, et al. A biochemicalbasis
for depressedketogenesisin sepsis.J Trauma 1986;26:419-25.
20. Wolfe RB, Jahoor F, Peters E, et al. Substrate cycling in
severely burned patients [Abstract]. Circ Shock 1986;18:359.
21. Sganga G, Siegel JH, Brown G, et al. Reprioritization of
hepatic plasma protein release in trauma and sepsis. Arch Surg
1985;120:187-99.
22. Weir JB. New methodsfor calculating metabolic rates with
special reference to protein metabolism.J Physiol 1949;109:1-9.
23. Sganga G, Siegel JH, Coleman B, et al. The physiologic
meaningofthe respiratoryindex in various types ofcritical illness.
Circ Shock
1985;17:179-93.
24. Laghi F, SiegelJH, Rivkind Al, et al. Respiratory index/
pulmonary shunt relationship:quantification of severityand prognosis in the post-traumatic adult respiratory distress syndrome.
Crit Care Med 1989;17:1121-8.
25. Stokes C, Blevins 5, Stoklosa JC, et al. Prediction of arterial
blood gas by transcutaneous 02 and CO2 in critically ill hyperdynamic trauma patients. J Trauma 1987;27:1240-60.
26. SiegelJH, Stoklosa JC, Borg U. Cardiorespiratory management of the adult respiratorydistresssyndrome.In: SiegelJH, ed.
Trauma: emergencysurgery and critical care. New York: Churchill Livingstone, 1987:581-673.
27. Siegel JH, Farrell EJ, Goldwyn RM, Friedman HR. The
surgical implicationsof physiologicpatterns in myocardial infarction shock. Surgery 1972;72:126-41.
28. Friedman HP, Goldwyn RM, Siegel JH. The use and interpretation of multivariable methods in the classification of stages in
the serious infectious disease processes in the critically ill. In:
Elashoff R, ed. Perspectives in biometrics. New York: Academic
Press, 1975:81-122.
29. Coleman WP, SiegelJH, Giovannini I, et al. Probability and
the patient state space. Int J Clin Monit Comput 1990 (accepted).
CLINICAL CHEMISTRY, Vol. 36, No. 8(B), 1990
1593