Download 17 Respiratory and cardiovascular support

17
Respiratory and cardiovascular
support
John A. Myburgh
17.1
Introduction
Our survival depends on the production of energy in
the form of adenosine triphosphate by the process of
oxidative metabolism. This in turn depends on a
continual supply of substrates, particularly oxygen
and glucose, to vital tissue beds such as the brain and
myocardium. In a critically ill patient, the supply of
oxygen may be compromised while at the same time
metabolic demand is increased. Inadequate oxygen
delivery may become the factor mitigating against
recovery or even survival of the patient (Schumacker
and Cain, 1987). In a patient with brain injury, cerebral
metabolic reserve is already reduced by the primary
insult and subsequent oxygen lack is a potent cause of
avoidable secondary brain injury (Chesnut et al.,
1993).
This chapter will address fundamental physiological principles of oxygen delivery from the atmosphere to the cells, the pathophysiological effects of
head injury on oxygen delivery and principles of
resuscitation and the maintenance of cardiopulmonary homeostasis.
17.1.1
THE PATIENT AT RISK
Although severe head injury is most common in
healthy young adults, particularly males, a proportion
of patients may suffer from pre-existing medical
conditions, which may be exacerbated by the stress of
neurotrauma and complicate its management (Luce,
1993; Kaufman et al., 1993). Ischemic heart disease and
hypertension mandate meticulous attention to mean
perfusion pressures of brain and myocardium; these
patients may require higher pressures than the normal
population to ensure optimal oxygen delivery. Control
of excessive centrally mediated sympathetic activity is
important in reducing both cerebral and myocardial
oxygen demand. Respiratory problems such as
asthma or chronic obstructive pulmonary disease may
affect mechanical ventilation.
Pre-existing nervous system disorders such as epilepsy, cerebrovascular or neuromuscular disease may
make assessment of effects of trauma and prognosis
more difficult.
Renal and hepatic dysfunction, particularly if there
is pre-existing impairment, may alter pharmacokinetics and drug interactions, leading to either toxic or
subtherapeutic effects. As an example, phenytoin,
frequently administered for prophylaxis of post-traumatic seizures, is a potent inducer of hepatic enzymes
and will alter the pharmacokinetic profiles of lipid
soluble drugs such as -adrenergic blockers and
opiates (Runciman, Myburgh and Upton, 1990).
Operative intervention may potentiate the acute
stress response, even in the young previously fit
patient (section 15.8).
17.1.2
THE ROLE OF INTENSIVE CARE
Management of head-injured patients in the Intensive
Care Unit (ICU) aims to optimize oxygen and substrate delivery to vital tissues and to detect events
that may harm the injured brain before they do
damage.
This requires prompt detection of potentially harmful changes by accurate monitoring and early intervention (Borel et al., 1990; Andrews, 1993). All aspects
of the patient’s state must be taken into account, with
particular attention to the cardiorespiratory, renal,
metabolic and nutritional systems (Bloomfield, 1989).
Although ICUs were established to preserve life and
improve the quality of survival, they are frequently
the setting for complications that increase morbidity
and mortality. Complications may result from intravascular catheters or from mechanical ventilators.
Nosocomial infection (Rodriguez et al., 1991) is a major
source of increased morbidity and prolonged ICU stay
(Hsieh et al., 1992). The acute respiratory distress
syndrome and multiple organ system failure may
occur as a result of the initial insult (Piek et al.,
1992).
Head Injury. Edited by Peter Reilly and Ross Bullock. Published in 1997 by Chapman & Hall, London. ISBN 0 412 58540 5
334
RESPIRATORY AND CARDIOVASCULAR SUPPORT
It is therefore important that the risk of each
procedure or drug administered is balanced by a net
benefit to the patient. To minimize complications and
to maximize patient benefit, all clinicians involved in
management must understand cardiopulmonary and
metabolic pathophysiology and the particular effects
of head injury. This applies both to neurosurgeons and
intensivists.
The main determinants of alveolar oxygen tension
(PAO2) are inspired oxygen tension and alveolar
ventilation.
17.2
Inspired oxygen tension (PIO2) is determined by the
fractional concentration of oxygen (FiO2) and the
ambient barometric pressure (BP) corrected for saturated water vapor pressure:
Physiology
Optimal cardiorespiratory function exists when oxygen delivery from the atmosphere to the cell matches
oxygen demand (Figure 17.1; Nunn, 1987a).
After a severe head injury, the brain is very
vulnerable to secondary insults, and remains so for a
period of perhaps 2–3 weeks. The oxygen cascade
may be threatened at a single link in the pathway, for
example by airway obstruction, but more often several
links are threatened simultaneously. The threat may be
primarily to the respiratory system (e.g. by chest
trauma or pneumonia) or to the circulatory system
(e.g. by hypovolemia or myocardial failure). Continued impairment of oxygen delivery may lead to
multiple organ failure, especially after traumatic injuries and sepsis.
Understanding the physiological determinants of
the various stages of the oxygen cascade may provide
guidance for optimizing cellular oxygen tension.
(West, 1985; Nunn, 1987a).
17.2.1 DISTRIBUTION OF OXYGEN FROM THE
ATMOSPHERE TO THE ALVEOLI
(a)
Inspired oxygen tension
PIO2 = FiO2 (BP – Pwat)
where Pwat represents the saturated water vapor
pressure (47 mmHg).
PIO2 is proportional to FiO2 if barometric pressure and
saturated water vapor pressure remain constant.
Ambient air has an oxygen concentration of 21%, so
that by administering 100% oxygen with a closed
circuit PIO2 may be raised from 150 mmHg in air to
700 mmHg. This is achieved by alveolar denitrogenation since ambient air contains 80% nitrogen.
Hence, since PIO2 can be effectively and markedly
raised by the administration of supplemental oxygen,
all severely injured patients should receive supplemental oxygen at the highest percentage available as
early as possible, to maximize the supply of oxygen to
the cells. (The effect of altitude on PIO2 is discussed in
Chapter 15.)
Figure 17.1 Transfer of oxygen from ambient air to mitochondria and summary of factors that influence oxygenation at
different levels in the oxygen cascade. PEO2 = end-expired gas. See text for details. (Source: modified from Nunn, 1987a,
p. 243, with permission.)
PHYSIOLOGY
Table 17.1
systems.
Capacities of available oxygen delivery
Apparatus/device
Oxygen flow
(l/min)
Inspired oxygen
concentration (%)
2
4
6
4
6
8
10
12
4
8
4–8
28
35
45
35
50
55
60
65
35
40, 60
40–80
Variable
Variable
21–100
21–100
Nasal catheters
Semi-rigid masks (e.g.
Hudson, Edinburgh,
Harris, Mary Catterell)
Venturi type mask (e.g.
Ventimask, Accurox)
Reservoir plastic masks
(e.g. Pneumask,
Polymask)
Anesthetic circuits
Ventilators
Table 17.1 shows the oxygen delivery capacities of
available oxygen circuits.
(b)
Alveolar ventilation
Adequacy of ventilation is reflected by the PaCO2,
normally 40 mmHg at sea level. Hypercapnic respiratory failure is defined as a PaCO2 greater than
50 mmHg at sea level unless there is a compensatory
metabolic alkalosis. (Serum bicarbonate concentration
indicates whether hypercapnia is acute or chronic, as
renal compensation for chronic hypercapnia takes 2–3
days.)
PAO2 is related directly to alveolar ventilation and
inversely to arterial carbon dioxide tension (PaCO2).
This relationship is expressed in the alveolar gas
equation:
PAO2 = PIO2 – PaCO2/RQ
where RQ is the respiratory quotient and reflects the
ratio of oxygen consumption to carbon dioxide production; normally 0.8.
The PAO2 in normal adults breathing air at sea level is
150 mmHg.
Hypercapnia may reduce PAO2 by displacing oxygen from the alveoli. Hypercapnia-induced respiratory acidosis may also reduce the consciousness level.
However this acidosis rapidly resolves on restitution
of effective alveolar ventilation. The acute stress
response to injury also increases basal metabolic rate
and carbon dioxide production which may further
reduce PAO2. The respiratory quotient usually remains
within the normal range due to an associated increase
in oxygen consumption (V̇O2).
335
The other determinant of PaCO2 is carbon dioxide
production (V̇CO2); normally 200 ml/minute in a
healthy person. An increase in carbon dioxide production or decrease in effective alveolar ventilation (V̇A)
will increase PaCO2 as expressed in the relationship:
PaCO2 = V̇CO2/V̇a.
Adequacy of ventilation is assessed by arterial blood
gas analysis obtained from repeated arterial punctures
or preferably from indwelling arterial catheters. A
continuous approximation of PaCO2 may be gained
from the end-tidal carbon dioxide tension measured
by a capnograph. Continuous end-tidal CO2 tension
may not represent PaCO2 in ventilated patients, particularly if there is coexistent lung injury or disease, and
should be checked frequently against PaCO2 (Liu, Lee
and Bongard, 1992). End-tidal carbon dioxide is
usually slightly less than PaCO2, but wide limits of
agreement between these parameters have been demonstrated (Russell, Graybeal and Strout, 1990;
Myburgh, 1991, Pansard et al., 1992).
Ventilatory variables such as respiratory rate (f),
tidal volume (VT) and their product, minute volume
(V̇E), may be accurately measured with most modern
ventilators. Alveolar ventilation (V̇A) is difficult to
determine and cannot be measured directly, although
alveolar ventilation may be inferred if V̇E and dead
space ventilation (V̇D) are calculated:
V̇E = V̇A + V̇D.
Dead space refers to those airways that are ventilated
but take no part in gaseous exchange and consists of
both conducting airways (anatomical dead space) and
non-perfused alveoli (physiological dead space;
Nunn, 1987a). The normal ratio of dead space to tidal
volume (VD/VT ratio) is 0.3. This may be calculated in
patients whose PaCO2 and expired carbon dioxide
tensions (PECO2) are known, using the modified Bohr
equation:
VD
VT
=
PaCO2 – PECO2
PaCO2
This ratio is an index of effective ventilation and is
helpful in determining the time to wean from mechanical ventilation. The VD/VT ratio in mechanically
ventilated patients with normal lungs is approximately 0.5. A ratio of 0.6 or greater is associated with
inadequate alveolar ventilation and hypercapnia.
In acute head injury, impaired conscious state is
associated with a loss of protective airway reflexes
and hypoventilation. Abnormal breathing patterns
may be exacerbated by alcohol, drugs, respiratory
injury or disease, resulting in alveolar hypoventilation and a rise in PaCO2. (North and Jennett, 1974;
Demling and Riessen, 1990). Hypoxia then rapidly
336
RESPIRATORY AND CARDIOVASCULAR SUPPORT
ensues and is potentiated by an increase in oxygen
consumption (Pfenninger and Lindner, 1991). This
reinforces the need for an adequate airway, ventilation and supplementary oxygen in these patients.
(c)
Lung mechanics
In addition to respiratory rate, tidal and minute
volume, other aspects of lung mechanics, in particular
lung compliance and airways resistance, are important
for effective alveolar ventilation in both spontaneously breathing and mechanically ventilated
patients (Bone, 1985; Nunn, 1987b).
17.2.2 DETERMINATION OF THE ARTERIAL OXYGEN
TENSION
Arterial oxygen tension (PaO2) is determined by the
diffusion capacity of the alveolar–capillary membrane,
pulmonary shunting and the relationship between
ventilation and perfusion. PaO2 is frequently a major
limiting step in the transfer of oxygen to the tissues.
Diffusion of oxygen across the short diffusion
distances of the alveolar capillary membrane normally
results in a negligible reduction in PaO2.
Ventilation–perfusion scatter or mismatching of
ventilation and perfusion (normally by ± 4 mmHg)
occurs through shunt and increased dead space
ventilation and reduces both oxygen and carbon
dioxide transfer.
Shunt is defined as perfusion of unventilated areas
of the lung. Normally a small shunt of approximately
4 mmHg exists as a result of the bronchial and cardiac
venous (Thebesian) circulations. Important features of
true shunting are that hypoxia due to shunting cannot
be abolished by increasing the FiO2, and the degree of
resulting hypoxia is proportional to the shunt
fraction.
Mismatch is increased in patients with underlying
chronic airways disease or pulmonary hypertension.
Neurally induced ventilation–perfusion mismatch has
been described as an uncommon complication of acute
head injury (Schumacker et al., 1979). This may be
exacerbated by acute intercurrent pulmonary pathology such as contusion, aspiration or infection, or
simply by placing patients on mechanical ventilators
(Hubmayr, Abel and Rehder, 1990). Pulmonary thromboembolism may cause severe hypoxia through both
ventilation–perfusion mismatch and reduction in cardiac output (Demling and Riessen, 1990).
These three factors – diffusion gradient, shunting
and ventilation perfusion mismatch – are responsible
for a normal alveolar–arterial oxygen gradient of
approximately 10 mmHg, determined by subtracting a
measured PaO2 from PAO2 calculated by the alveolar
gas equation.
Potentiation of any of these physiological causes of
reduced arterial oxygen tension by pathological processes will result in hypoxia.
An increased alveolar–arterial oxygen gradient
after head injury is most often due to alveolar
obstruction or collapse caused by contusion, aspiration of gastric content or blood or consolidation from
infection. Neurally induced ventilation–perfusion
mismatch is rare.
17.2.3 DELIVERY OF OXYGEN FROM THE ALVEOLI
TO THE TISSUES
Oxygen delivery (ḊO2) to individual cells involves
convective and diffusive oxygen transport (Leach and
Treacher, 1992).
Convective oxygen transport refers to the bulk
movement of oxygen in the blood. It depends primarily on the distribution of cardiac output to individual
organs and the mechanisms regulating tissue
microcirculation.
Diffusive oxygen transport is the process of uptake
and extraction of oxygen from blood by the tissues
(Leach and Treacher, 1992).
ḊO2 may be calculated as the product of arterial
oxygen content (CaO2) and cardiac output (Q̇):
ḊO2 = Q̇ CaO2.
The normal value for ḊO2 is 900–1300 ml/min.
(a)
Oxygen content
Arterial oxygen content (CaO2) is the total oxygen per
milliliter of blood and includes both oxygen combined
with hemoglobin and dissolved oxygen. CaO2 (normal
value 16–20 ml O2/100 ml blood) may be measured
directly with oxygen content analyzers or calculated
by the following equation:
CaO2 = (Hb 1.34 SaO2) + (PaO2 0.003)
where Hb is hemoglobin (g/dl); 1.34 is the oxygen
carrying capacity of hemoglobin (ml/g); SaO2 is
arterial oxygen saturation of hemoglobin (%); 0.003
represents the solubility of oxygen in blood in plasma
at 37°C.
The chemical relationship between oxygen and hemoglobin is represented by the oxygen–hemoglobin
dissociation curve. The idealized curve assumes a
normal hemoglobin and P50 (oxygen tension with 50%
hemoglobin saturation). A number of logarithmic
equations have been designed to determine oxygen
saturation from a single PaO2 measurement (Breuer et
al., 1989). However, P50, which is influenced by pH
and 2,3-diphosphoglycerate, a marker of intracellular
metabolism and utilization of oxygen, is frequently
reduced in critically ill patients (Myburgh, Webb and
PHYSIOLOGY
Worthley, 1991). Wide limits of agreement between
calculated and measured saturations have been demonstrated. Therefore oxygen saturations, either arterial
or venous, should be measured directly by oximetric
analysis when calculating oxygen delivery or consumption (Myburgh, 1992). This also applies to the
measurement of jugular venous bulb oxygen saturation (see below).
The major determinant of arterial oxygen content is
hemoglobin. Dissolved oxygen is a minor component
of the total. Anemic or bleeding patients have a
reduced oxygen transport capacity; hence transfusion
of blood is an integral part of resuscitation in trauma
patients.
There is contention on the optimal hemoglobin level
in severely head-injured patients with coexisting
extracranial injuries (Dhainaut et al., 1990; Greenberg,
1990) and consideration must be given to the risk and
benefits of small blood transfusions intended to raise
the hemoglobin level by only 1–2 g/dl (Dietrich et al.,
1990). There is increasing evidence that a fall in
hemoglobin is part of the early response to trauma
(Crosby, 1992), and that single unit transfusion does
not significantly increase tissue oxygen utilization and
may cause splanchnic ischemia (Silverman and Tuma,
1992). Impairment of immune function and the risk of
transmitting infection are additional arguments
against single unit transfusion in critically ill
patients.
Optimal tissue oxygen extraction from hemoglobin
to the brain and myocardium occurs at a hematocrit
of around 0.32 (Dantzker, 1989). It is not possible to
recommend a specific hemoglobin or hematocrit for
all clinical situations. However a hemoglobin of 8 g/dl
is regarded as the transfusion limit for young and
otherwise healthy trauma victims.
(b)
Cardiac output
Cardiac output is the product of heart rate (HR) and
stroke volume (SV). Any reduction in cardiac output
will reduce oxygen delivery.
Heart rate, normally 70 beats per minute, is determined by autonomic influences on intrinsic cardiac
pacemakers and is the principal factor maintaining
cardiac output in the traumatized patient (Schumacker and Cain, 1987; Guyton, 1991). The heart rate
response may be blunted in patients on beta-blockers
or in patients with high spinal injuries causing cardiac
denervation (Luce, 1985).
Stroke volume, normally 70 ml/beat, is determined
by three factors:
preload, defined as end diastolic cardiac muscle
fiber length and representing the volume status of
the patient;
337
contractility,
the
inotropic
state
of
the
myocardium;
afterload, the impedance to ventricular ejection.
Preload is the most important determinant of stroke
volume after trauma (Dhainaut et al., 1990). Reduced
preload, i.e. hypovolemia, must be aggressively treated to ensure adequate tissue oxygen delivery.
Impaired contractility may result from direct myocardial injury or pre-existing ischemic cardiomyopathy. Drugs (inotropes) are being increasingly used to
maintain adequate contractility once hypovolemia
has been corrected, not only to promote oxygen
delivery but also to maintain cerebral and myocardial
perfusion (Schumacker and Samsel, 1989; Dhainaut et
al., 1990).
Mean arterial pressure (MAP) is the product of
cardiac output and total peripheral resistance, derived
as systemic vascular resistance (SVR) from pulmonary artery catheter measurements.
17.2.4 PERIPHERAL EXTRACTION AND UTILIZATION
OF OXYGEN
The uptake and extraction of oxygen from blood by
the tissues are dependent primarily on diffusive
oxygen transport (Leach and Treacher, 1992). This
refers to the transfer of oxygen from blood through the
extracellular matrix along a capillary–intracellular
oxygen tension gradient; the process depends on
arterial oxygen tension. Alterations in microvasculature tone by neurohumoral (myogenic) and metabolic means adjust local oxygen delivery to changing
metabolic demands. Microvascular beds in different
tissues respond to increased oxygen requirements
either by increasing flow or by recruiting additional
vessels. For example, skeletal muscle can substantially
increase the number of functioning capillaries in
response to the increased metabolic demand, whereas
the myocardium and brain are more dependent on
changes in pressure and flow.
In acute neurotrauma, it is essential to maintain
adequate cardiac output and cerebral perfusion pressure in order to sustain flow-dependent oxygen transport. Interstitial edema, by increasing the distance for
diffusion, also jeopardizes oxygen transfer (Schumacker and Samsel, 1989).
Assessment of oxygen utilization by tissues is
difficult and in general only global oxygen consumption and delivery can be measured easily. Global (i.e.
whole body) oxygen consumption (normally
175–300 ml/min) is best measured by indirect calorimetry of expired respiratory gases using metabolic
monitors. The accuracy of these monitors is reduced at
high inspired oxygen concentrations (FiO2 > 0.8) and
by air leaks (Takala et al., 1989). Oxygen consumption
may be determined by the Fick equation, which
338
RESPIRATORY AND CARDIOVASCULAR SUPPORT
expresses the relationship between cardiac output (Q̇),
oxygen consumption (V̇O2) and the arteriovenous
oxygen content difference:
V̇O2 = Q̇ (CaO2 – CvO2).
The measurement and manipulation of oxygen delivery and extraction have been extensively debated
(Vincent, 1990; Dantzker, Foresman and Gutierrez,
1991; Shoemaker, Appel and Kram, 1992; Reed, 1993).
When oxygen delivery is reduced under physiological
conditions, oxygen consumption is maintained until
very low levels of oxygen delivery through increasing
oxygen extraction. Below a critical oxygen delivery
level oxygen consumption falls even though oxygen
extraction continues to increase (Schumacker and
Cain, 1987).
Under pathological conditions, such as trauma and
sepsis, the relationship between oxygen delivery and
consumption is more linear and a critical oxygen
delivery level cannot be identified. Recent studies
indicate that outcome in patients is improved if
oxygen consumption matches delivery, if necessary
through cardiovascular support (Shoemaker et al.,
1988; Reinhart, Hannemann and Kuss, 1990).
The clinical measurement of both oxygen consumption and delivery remains difficult and is compounded by mathematical errors (Archie, 1987; Reed,
1993). Elevated serum lactate does not necessarily
imply anaerobic metabolism, and its value as an index
of regional oxygen demand is limited (Silverman,
1991; Mizock and Falk, 1992).
17.3 Effects of head injury upon oxygen
delivery
Severe head injury may impair oxygen delivery
through the development of abnormal breathing patterns, activation of a systemic inflammatory response
and centrally mediated sympathetic overactivity.
17.3.1 BREATHING PATTERNS IN HEAD-INJURED
PATIENTS
Variable breathing patterns may result from head
trauma and other insults to the central nervous system
(North and Jennett, 1974). Cheyne–Stokes or periodic
respiration is associated with bilateral hemispheric
supratentorial lesions; central neurogenic hyperventilation or apneustic respiration with pontine lesions,
and ataxic respiration or apnea with medullary or
midbrain lesions. However, in practice, the diffuse
nature of the primary and secondary effects of trauma
upon central brain structures means that these patterns are seldom of diagnostic value.
Abdominal paradox, an inward movement of the
abdominal wall during inspiration, indicates abnor-
mal diaphragmatic contraction and may be seen in
patients with high cervical spinal cord injury above or
involving the phrenic nerve nuclei (level C3–C5; Luce,
1985). Chest trauma causing flail chest and pulmonary
contusion results in tachypnea and intercostal muscle
retraction (Demling and Riessen, 1990; Luce, 1993).
17.3.2 SYSTEMIC PHYSIOLOGICAL RESPONSE TO
ACUTE INJURY
The systemic physiological response to acute injury
and infection causes a potent, mediator-induced, nonspecific syndrome termed the systemic inflammatory
response syndrome (SIRS). This syndrome has been
extensively studied (Bone, 1991a, b; Bone et al., 1992a;
Vincent and Bihari, 1992) and includes one or more of
the following clinical manifestations:
body temperature greater than 38°C or less than
36°C;
heart rate greater than 90 beats/min;
tachypnea, with a respiratory rate greater than
20 breaths/min or hyperventilation, as indicated by
a PaCO2 less than 32 mmHg;
white blood cell count greater than 12 000/mm3 or
less than 4000/mm3, or containing more than 10%
immature neutrophils (bands) representing an acute
alteration from baseline in the absence of other
known causes such as chemotherapy, induced
neutropenia and leukopenia.
Infection may produce a systemic inflammatory
response syndrome. Non-infectious pathological
events such as multiple trauma, hemorrhagic shock,
ischemia and pancreatitis result in an identical syndrome (Bone et al., 1992a).
SIRS is mediated through a number of cytokines
such as tumor necrosis factor (Strieter, Kunkel and
Bone, 1993; Tracey and Cerami, 1993), interleukins
(Dinarello, 1984; Platanias and Vogelgang, 1990) and
arachidonic acid metabolites (Lewis, Austen and
Soberman, 1990). The organ system responses, particularly the cardiovascular, respiratory and renal responses, may either regress or progress to well-defined
clinical conditions such as acute lung injury, septic
shock, acute renal failure and multiple organ dysfunction syndrome (Clarke, 1985; Cerra, 1992). Hemodynamic hallmarks of this response are a high cardiac
output state due to reduced systemic vascular resistance and global blood flow redistribution. Cardiac
output and stroke volume are usually well maintained during the initial acute stages of this response
and may reflect a mediator-induced increase in oxygen
delivery to vital tissue beds.
It is important to recognize the systemic inflammatory response syndrome and to actively seek infectious and/or non-infectious causes. If infection occurs,
EFFECTS OF HEAD INJURY UPON OXYGEN DELIVERY
339
SIRS may progress to septic shock, maldistribution of
blood flow, myocardial depression and end organ
failure. (Bersten and Sibbald, 1989).
cess. Beta-blocking drugs appear to be preferable,
particularly when ICP is elevated (Robertson et al.,
1983), but are seldom needed.
17.3.3
(b)
NEUROHUMORAL RESPONSE
In addition to the acute systemic inflammatory
response, sympathetically mediated responses following acute head injury result in three clinical
symptom complexes: neurogenic hypertension, myocardial ischemia and neurogenic pulmonary edema.
(a)
Neurogenic hypertension
Neurogenic hypertension is common after head injury.
It appears to be sympathetically mediated and is
directly proportional to the degree of catecholamine
release (Graf and Rossi, 1978). Brain areas implicated
include the hypothalamus, the nucleus of the tractus
solitarius, the area postrema and areas A1 and A5 of
the medulla (Kirland and Wilson, 1991). Increased ICP,
brain-stem compression or medullary ischemia may
cause severe systemic hypertension, accompanied or
preceded by vagally mediated bradycardia (Graf and
Rossi, 1978). This is the classical Cushing response but
in practice it is not always seen; the hypertensive
response can be masked by coexisting hypotension
due either to hypovolemia or to a hyperdynamic
systemic inflammatory response.
In as many as 25% of patients with neurogenic
arterial hypertension, there is no associated increase
in ICP (Simard and Bellefleur, 1989).
Unlike essential hypertension, neurogenic hypertension may be quite labile, with varying cardiac
output and systemic vascular resistance, probably due
to mediator-induced vasoresponsiveness. When or
whether to treat neurogenic hypertension is still
controversial, since an elevated systemic arterial
pressure may represent an intrinsic compensation for
impaired cerebrovascular autoregulation and low
cerebral perfusion pressure. Cerebral autoregulation
may be abolished or impaired in the injured brain,
either regionally or globally, so that the normal
relationship between CPP and blood flow is disrupted
(Obrist et al., 1979) and cerebral blood flow becomes
perfusion pressure-dependent (Simard and Bellefleur,
1989). Theoretically, elevating systemic arterial pressure may benefit ischemic areas, but may also increase
hydrostatic pressure and vasogenic cerebral edema.
Conversely, low arterial pressures may increase ischemia, cellular hypoxia and cytotoxic cerebral edema
(Kong et al., 1991).
Neurogenic hypertension should only be treated
when mean arterial pressure approaches 150 mmHg
and sedation or analgesia with opiates and/or benzodiazepines has already been employed without suc-
Myocardial injury
Myocardial injury, in the absence of coronary artery
disease, may occur in up to 50% of head-injured
patients. The lesions are similar to those seen after
acute myocardial infarction, subarachnoid hemorrhage, pheochromocytoma or prolonged catecholamine infusions, where subendocardial hemorrhages
are commonly found at autopsy. Inhibition of both
catecholamine release (Talman, 1985) with adequate
sedation and beta-adrenergic receptor blockade (NeilDwyer et al., 1978; Cruikshank et al., 1987) have been
shown to prevent or ameliorate myocardial injury
after head injury.
Although the brain effector site is unknown, there
may be an association between hypothalamic lesions
and myocardial damage. In patients with myocardial
injury, cardiac isoenzymes may be elevated and
electrocardiogram changes consistent with an infarction pattern can appear. These changes may be
potentiated by pre-existing ischemic heart disease or
hypertension. It has been suggested that electrocardiogram changes may indicate patients with more
serious brain injury and thus a higher mortality
(Kirland and Wilson, 1991).
Similarly, a wide range of cardiac dysrhythmias
may occur as a result of centrally mediated sympathetic and vagal discharge. The true incidence of
dysrhythmias is unknown and all patients with a
severe head injury should have electrocardiograph
monitoring. Tachyarrhythmias are predominantly
supraventricular and include atrial fibrillation or
flutter and paroxysmal or multifocal atrial tachycardias (Cruikshank et al., 1987). Ventricular tachyarrhythmias are uncommon. Reducing sympathetic
tone with adequate sedation or occasionally with
beta-blockade is usually sufficient treatment for
tachyarrhythmias. In patients with poor ventricular
function persistent hemodynamically significant
tachyarrhythmias are best treated with class 3 agents
such as amiodarone. Nodal and sinus bradycardias
are quite commonly associated with sustained intracranial hypertension and rarely require treatment.
Atropine, beta-agonists or, rarely, transvenous pacing
may be indicated if hemodynamic compromise
results.
These cardiovascular conditions are seen frequently
during the initial stages of neurotrauma. However,
hypertension may persist into the rehabilitation phase.
In a study of head-injured patients in a rehabilitation
center, 11% had systemic hypertension yet only one
had a history of essential hypertension and 68% were
340
RESPIRATORY AND CARDIOVASCULAR SUPPORT
less than 30 years of age. At discharge, 63% still
required antihypertensive medications. A total of 21%
of these patients had non-specific ST–T wave changes
on their electrocardiograms (Kalizky et al., 1985).
(c)
Neurogenic pulmonary edema
The most dramatic of these sympathetically mediated
processes is neurogenic pulmonary edema. It was
reported many years ago to be present in most
patients who died of intracranial hemorrhage (Weismann, 1939) and in patients with severe isolated head
injury (Simmons et al., 1969), but was much less
common in patients who survived (Singbartl, Cunitz
and Hamrouni, 1982). The common locus of damage is
the brain stem. The mechanism is a large sympathetic
discharge (Demling and Riessen, 1990), which leads to
a redistribution of blood from the systemic to the
pulmonary vasculature and a rapid increase in pulmonary vascular pressures. This is a transient phenomenon but it may damage the pulmonary endothelium,
causing a persistent permeability defect (Theodore
and Robin, 1976)
This mechanism alone does not explain all cases of
neurogenic pulmonary edema. Other mechanisms
that may be involved include: neurogenic depletion
of pulmonary surfactant (Beckman, Bean and Baslock, 1974); direct sympathetic control of the number
and size of pulmonary endothelial pores (Rosell,
1980); constriction of the pulmonary lymphatics by
the sympathetic discharge (McHale and Roddie,
1983); lung microembolization from activation of the
clotting cascade, possibly caused by release of brain
tissue thromboplastin into the systemic circulation
(Kaufman et al., 1980; Simpson, Speed and Blumbergs, 1991); and the direct effects of increased
sympathetic activity on lung vasculature resulting in
pulmonary hypoperfusion and hypoxia (Balk and
Bone, 1983).
Typically, neurogenic pulmonary edema has an
immediate onset and becomes clinically evident 2–12
hours after injury. Occasionally clinical onset may be
delayed from 12 hours to several days. It is usually
self-limiting and, if the patient survives, it will resolve
in hours to days. A prolonged episode of pulmonary
edema would suggest other causes of either cardiogenic or non-cardiogenic origin.
The clinical findings in patients with neurogenic
pulmonary edema are tachypnea, hypoxia, decreased
pulmonary compliance, diffuse alveolar and interstitial ‘fluffy’ infiltrates on chest X-ray and low
filling pressures measured by right heart catheters (cf.
acute respiratory distress syndrome).
Treatment is primarily supportive and aims at
ensuring adequate oxygenation and ventilation. This
usually requires endotracheal intubation and mechan-
ical ventilation with positive end-expired pressure
(PEEP).
Sympathetic overactivity is usually abolished with
adequate sedation and beta-blockade is usually
unnecessary. Diuretics are effective but must be
titrated against the volume status of the patient so that
cerebral perfusion pressure is not compromised.
17.4
Management: resuscitation
The resuscitation phase is defined as the period from
on-site resuscitation and evacuation to stabilization.
Priority must be given to the most life-threatening
problems and this is the principle of basic and
advanced trauma life support systems (Buss, Nel and
van den Heever, 1990; Trauma Committee, Royal
Australasian College of Surgeons, 1992).
17.4.1
AIRWAY MANAGEMENT
Loss of airway patency and hypoxia at any stage of
head injury management is a life-threatening event. It
is the main cause of secondary brain injury and
significantly increases morbidity and mortality (Gentleman, 1992; Chesnut et al., 1993).
All severely head-injured patients are at risk from
inadequate airway protection. The degree of airway
intervention necessary will depend on the level of
consciousness and adequacy of protective glottic
reflexes.
The mouth and upper airway must be inspected for
foreign bodies, bleeding or ill-fitting dentures and
cleared under direct vision using a rigid sucker.
Simple maneuvers such as chin lift and jaw thrust and
the use of oropharyngeal airways may be enough to
allow efficient oxygenation in self-ventilating patients.
Nasopharyngeal airways must be used with caution if
cribriform plate fracture is suspected since they have
been passed directly into the cranial cavity. They may
also cause trauma to the nasal mucosa with bleeding
into the nasopharynx and upper airway. The same
cautions apply to the use of nasotracheal and nasogastric tubes.
Emergency endotracheal intubation is indicated
whenever the airway is threatened, especially in the
comatose head-injured patient. It is also indicated in
faciomaxillary trauma or upper airway obstruction
from direct laryngeal trauma. This is best performed
where there is expertise in anesthesia and specific
airway techniques such as rapid sequence induction
and intubation, blind nasal intubation, fiberoptic
laryngoscopy, cricothyroidotomy and tracheostomy.
However, urgency may dictate intubation in less
optimal circumstances.
Indications for early endotracheal intubation after
head injury are:
MANAGEMENT: RESUSCITATION
coma;
inability to obey commands with evidence of
respiratory compromise:
– tachypnea > 30/min
– sternal recession
– upper respiratory bleeding;
blood
gas
analysis
–
PaO2 < 70 mmHg;
PaCO2 < 45 mmHg.
Table 17.2 Priorities of rapid sequence induction of
anesthesia for emergency intubation in neurotrauma. See
text for further details. ETCO2 = end-tidal carbon dioxide
tension
Optimize conditions
for intubation
Skilled assistant
Adequate light
Suction
Neutral position
Equipment
Oropharyngeal airways
Self inflating hand ventilating
assembly and mask
100% oxygen
Two laryngoscopes
Magill forceps
Introducer/bougie
Endotracheal tubes (mm):
– 7.0, 8.0
– 8.0, 9.0
Cricothyroidotomy equipment:
Scalpel
6.0 mm cuffed endotracheal tube
Drugs
Induction agent
Suxamethonium (1–2 mg/kg)
Atropine (0.6–1.2 mg)
Adrenaline (10 ml 1:10 000
solution)
Sedation (fentanyl, diazepam,
morphine, midazolam)
Monitoring
Pulse oximetry
Capnography
Arterial blood pressure
Electrocardiograph
Procedure
Preoxygenation 100% oxygen for
3–4 minutes
Preload with 250–500 ml colloid
intravenously
Manual in-line cervical spine
immobilization
Cricoid pressure applied
Induction agent + suxamethonium
Direct visualization of vocal cords
Oral endotracheal intubation
Inflation of cuff until sealed
Confirmation of end-tidal CO2 and
chest auscultation
Cricoid pressure released
Secure tube at correct length
Ensure adequate sedation ± muscle
relaxant
Consider naso- or orogastric
intubation
Chest X-ray
Confirm blood gas analysis
Indication for cricothyroidotomy or tracheostomy
are:
coma and inability to intubate;
trauma, bleeding or fractures involving the upper
airway.
Comatose patients (Glasgow Coma Score less than 8)
should be intubated electively as soon as possible to
prevent aspiration of gastric contents, to allow effective oxygenation, ventilation and control of PaCO2 and
definitive imaging with computed tomography (CT;
Pfenninger and Lindner, 1991). Similarly, head-injured
patients who are agitated or combative may best be
managed by early intubation and controlled ventilation until diagnostic and therapeutic interventions are
completed (Redan et al., 1991). These patients should
be intubated orally using a rapid sequence induction
of anesthesia after preoxygenation and with the
application of cricoid pressure as described in Table
17.2 (Walls, 1993). Patients with significant extracranial injuries causing ventilatory failure must also be
intubated early. Patients with moderately depressed
or fluctuating conscious states may be observed
provided adequate glottic reflexes are present and the
cardiorespiratory state is stable.
All unconscious head-injured patients must be
regarded as having cervical spine injuries and intubation should be performed with in-line cervical spine
immobilization with the head in the neutral position
so that neck flexion and extension is avoided. Patients
may be intubated with rigid collars in place and these
should be left in situ until definitive radiological
views of the cervical spine are obtained (Crosby and
Lui, 1990).
The best method of assuring correct placement of an
endotracheal tube is by directly visualizing the passage of the tube through the vocal cords (Szekely et al.,
1993). End-tidal carbon dioxide and a respiratory
waveform on capnography (Lillie and Roberts, 1988),
auscultation of the chest and misting on the endotracheal tube during exhalation give further confirmation of correct placement, although false positives
have been described following inadvertent esophageal
intubation (Holland, Webb and Runciman, 1993).
Advancing the tube down the right main bronchus is
a common error especially in children. It is best
prevented by repeated auscultation and chest X-ray
341
examination. Tubes must be secured at the appropriate
length (normally 21 cm at the teeth in women and
23 cm in men). Adhesive tape is the most effective
means of securing endotracheal tubes.
Tracheostomy is only indicated acutely in patients
who cannot be intubated; e.g. those with laryngeal or
342
RESPIRATORY AND CARDIOVASCULAR SUPPORT
tracheal trauma or some patients with high cervical
spinal cord injury.
17.4.2
VENTILATION
Once the airway is secured, all injured patients must
receive oxygen at the highest concentration available
(Oh and Duncan, 1988). The decision to continue
ventilation is primarily a clinical one and is based on
the severity and pattern of head injury, associated
injuries and pre-existing lung function. Arterial blood
gas analysis of PaO2, PaCO2 and pH are important but
should not delay this decision.
For self-ventilating patients, face mask oxygen
using the delivery systems outlined in Table 17.1 are
suitable. For intubated patients, manual or mechanical
ventilating assemblies can reliably deliver an FiO2 of
1.0. Hand-held, self-inflating assemblies can effectively ventilate both intubated and non-intubated
patients and also give an impression of lung compliance during inflation, thereby reducing the risk of
disconnection from either the endotracheal tube or the
face mask (Table 17.1).
Emergency assessment of oxygenation may be
difficult when an impaired conscious state may be due
to head injury, alcohol, drugs or sedatives. Pulse
oximetry has greatly aided this assessment and is now
mandatory in trauma patients. A saturation greater
than 95% generally indicates a PaO2 of at least
75 mmHg and is recommended.
Pulse oximeters are less reliable in hypoperfused,
hypothermic and agitated patients (Runciman et al.,
1993). Nevertheless, even under conditions of poor
perfusion, a recorded saturation of less than 95% with
a pulse waveform indicates hypoxia (Clayton et al.,
1991). This must be corrected by delivering the highest
possible oxygen concentration, reassessing the airway
and ventilation, and excluding potential causes of
unexplained hypoxia such as tension pneumothorax
or cardiac tamponade. Arterial blood gas analysis
should be performed as soon as possible, as this is the
only means of calibrating a pulse oximeter.
For 30–40 years hyperventilation has been advocated for the initial reduction of intracranial pressure.
Global and regional cerebral blood flow may be
reduced by as much as 50–60% during the first several
hours following injury. Acute hypocapnia may further
reduce cerebral blood flow in areas of the brain where
vasoresponsiveness to hypocapnia is relatively preserved. This may or may not reduce ICP, yet the
reduction in blood flow may worsen ischemia. Early
hyperventilation is now only recommended in
patients with clinical signs of herniation or rapid
neurological deterioration, prior to imaging or measurement of ICP and providing adequate circulatory
resuscitation is complete. The level of PaCO2 required
to reduce cerebral blood flow and achieve optimal ICP
is not well defined but, on balance, moderate hypocapnia of PaCO2 to not less than 30 mmHg is recommended. In the patient without clinical evidence of
intracranial hypertension or brain shift, ventilation
should be set to achieve normocapnia.
17.4.3
CIRCULATION
Hypotension after severe head injury is independently
associated with significant increases in morbidity
and mortality. Resuscitation protocols for braininjured patients should aim assiduously at avoiding
hypovolemic shock. (Buss, Nel and van der Heever,
1990; Chesnut et al., 1993; Hill, Abraham and West,
1993).
Restoring circulating blood volume is the basis of
hemodynamic resuscitation in trauma patients. The
initial assessment of circulatory status may be difficult
since blood pressure may be well maintained by
sympathetic activity. Tachycardia and reduced capillary return are cardinal signs of hypovolemia. However
tachycardia is common after trauma and peripheral
perfusion may be difficult to assess in hypothermic
patients. Hence, 1–2 liters of balanced salt solution
(lactated Ringer’s or physiological saline solution)
should be infused initially in all patients through largebore peripheral venous access until definitive monitoring of blood pressure and volume status is established.
Extracranial sources of hemorrhage must be identified
and rapidly treated. This requires a coordinated
‘trauma team’ approach where priorities are given to
the most life-threatening injuries in the shortest
possible time (Trunkey, 1991; Trauma Committee,
Royal Australasian College of Surgeons, 1992).
Hypotensive patients should be given colloid solutions to expand the intravascular volume and restore
preload. Normal (5%) serum albumin solution or
synthetic polygeline are suitable solutions. The use of
hypertonic solutions in neurotrauma remains controversial and is discussed in Chapter 16. Blood
transfusion must be administered to patients who
have lost more than 20–30% blood volume or where
significant hemorrhage is anticipated. Patients who
require anesthesia for intubation should have adequate preload and arterial blood pressure monitoring
in situ as sudden ablation of sympathetic tone by
anesthetic agents may unmask a compensated hypovolemic state and induce hypotension.
As a general principle vasopressors should not be
used in hypovolemic patients until volume resuscitation is completed.
Causes of refractory hypotension in these patients
include acute spinal injury, tension pneumothorax,
cardiac tamponade and severe myocardial contusion.
These must be excluded early.
MAINTENANCE OF CARDIOPULMONARY–CEREBRAL HOMEOSTASIS
17.4.4
SEDATION AND ANALGESIA
In the acute phase, sedation for intubated and ventilated patients must be titrated against the hemodynamic stability. High-dose narcotics (15–25 g/kg
fentanyl) and non-depolarizing muscle relaxation
(pancuronium 8–10 mg) provide sufficient sedation
and allow control of ventilation for 1–2 hours while
imaging and other investigations are performed. This
combination provides hemodynamic stability; moreover tachycardias due to vagal blockade by pancuronium are reduced with high doses of fentanyl
(Salmenperä et al., 1983). Frequent reassessment of
conscious state is essential as excessive sympathetic
activity may potentiate raised intracranial pressure or
myocardial ischemia in a paralyzed but awake patient
(a state that should be carefully avoided). Sedation
may need to be supplemented with intermittent doses
of opiates or benzodiazepines.
17.4.5
(a)
MONITORING
Arterial pressure
During the initial resuscitation phase, a clinical
impression of cardiac output and tissue perfusion can
be gained from the strength of peripheral pulses and
the time required for superficial capillaries to refill
after they have been blanched (normally 2–3 s).
Accurate measurement of mean arterial blood pressure and heart rate is essential.
Non-invasive blood pressure devices are inaccurate
outside the normal range, tending to over-read in
states of hypotension and under-read in hypertension
(Rutten et al., 1986; Cockings et al., 1993). Radial,
brachial or femoral arterial catheterization is recommended. The arterial impulse may be connected via a
fluid-filled catheter to a transducer. This is the
standard method for measuring systemic blood pressure in the ICU and may also be used to obtain
samples for blood gas analysis (Runciman, Ilsley and
Rutten, 1988). Local vascular responses at the radial
arterial level may result in false elevations of or
reductions in blood pressure (Pauca et al., 1992).
Mean arterial pressure correlates more closely with
organ perfusion than systolic pressure and is less
susceptible to the errors of measurement that occur
with suboptimally damped transduced measuring
systems. Optimal mean arterial pressure is usually
60–80 mmHg, although patients with pre-existing
hypertension or raised intracranial pressure may
require higher mean arterial pressures.
Complications of arterial catheterization such as
local hematoma formation, ischemia distal to the site
of insertion or local infection may be minimized by
slow continual flushing of these catheters with dilute
343
solutions containing heparin and by sterile catheter
insertion and maintenance techniques.
(b)
Electrocardiography
Electrocardiography is routine, but since changes in
heart rate in response to hypoxia and hypotension
occur late, the ECG is a poor marker of vital organ
oxygenation (Ludbrook et al., 1993).
(c)
Central venous pressure
Central venous pressure monitoring is a useful guide
to volume status, but the insertion of these catheters
should not delay the restoration of volume. Catheters
should be placed under sterile conditions by skilled
operators on a semi-elective basis and pressures
should be measured with electronic transducers (see
section 17.5.2(b)).
The response to fluid resuscitation is best assessed
by improvements in heart rate, capillary return and
mean arterial blood pressure, and by a maintained
urine output of 0.5–1 ml/kg/h.
17.4.6
NEUROSURGICAL DIAGNOSIS
Once patients have been adequately resuscitated the
definitive diagnosis of intracranial pathology, usually
by CT scanning, is essential. In a patient who shows
signs of brain-stem compression, evacuation of an
intracerebral hematoma is a matter of great urgency.
Resuscitation should therefore be conducted pari passu
with the CT scan, or the patient should be taken
directly to the operating room. Following CT diagnosis and/or evacuation of intracerebral hematomas,
ICP and other monitoring devices such as jugular bulb
oxygen catheters should be inserted
Accurate, lightweight and portable monitors allow
continuous display of ECG, heart rate, oxygen saturation and arterial pressure and are recommended
during transport. This phase of treatment is covered in
more detail in Chapters 15 and 18.
17.5 Management: maintenance of
cardiopulmonary–cerebral homeostasis
After acute resuscitation and any necessary operative
procedures comes the phase of management in which
the chief aim is to prevent secondary hypoxic–ischemic
brain injury. This phase continues until the patient no
longer requires homeostatic control and support.
During this period patients are usually managed in the
Intensive Care Unit, with mechanical and/or pharmacological cardiorespiratory and metabolic support. The
period of vulnerability to secondary brain injury is not
344
RESPIRATORY AND CARDIOVASCULAR SUPPORT
clearly defined and varies from patient to patient, but it
may last up to 10–14 days after injury and beyond the
time when intracranial pressure is elevated. Management of the patient requires close liaison between
intensivists, neurosurgeons and other involved
specialists.
All observations should be accurately recorded on
24-hour charts that correlate all vital signs and
cumulative hourly fluid balance. This facilitates early
identification of trends, a more important indicator of
progress than single isolated recordings.
17.5.1
(a)
RESPIRATORY MANAGEMENT
Airway management
Patients should be intubated until ICP is controlled
(i.e. consistently below 20 mmHg) and respiratory
function indices are satisfactory (see below). Endotracheal intubation may be performed either translaryngeally, through the nose or mouth, or transtracheally, via a tracheostomy. The development of
polyvinylchloride tubes with low-pressure, high-volume cuffs has reduced the incidence of tracheal
mucosal damage and ulceration from nasal or oral
endotracheal tubes. As a result, translaryngeal intubation may continue as long as the head injury and lung
function require it (Stone and Bogdonoff, 1992).
Nasal intubation gives good support to the endotracheal tube and patients often tolerate nasal tubes
with little sedation. These tubes are less prone to move
in the trachea than oral tubes and therefore cause less
tracheal mucosal damage. Patients are able to swallow
their secretions and oral hygiene is more easily
maintained than with an oral tube. However suction
through nasal tubes is more difficult because of their
narrower diameter and greater length. Nasal intubation is contraindicated in patients with basal skull
fracture especially when the cribriform plate is
damaged.
Oral intubation is preferred in an emergency. It
allows passage of a tube with a larger diameter (8 mm
or more) and easier bronchial toilet. However these
tubes are more prone to movement and more difficult
to secure. The incidence of nosocomial sinusitis is
significantly lower with oral intubation (Rouby et al.,
1994).
Tracheostomy is generally indicated in patients
where prolonged ventilation due to respiratory failure
is needed or where a slow return of consciousness is
anticipated, e.g. of more than 7–10 days.
These patients usually have poor protective glottic
reflexes and inadequate clearance of secretion, and are
prone to aspiration and nosocomial pneumonia. Tracheostomy tubes are more comfortable than translaryngeal tubes and permit easier tracheobronchial toilet.
Enteric feeding via a soft, small-bore nasoenteric tube
is safer after tracheostomy. Tubes with a device that
directs a stream of air retrogradely through the larynx
above the cuff site allow some patients to talk.
Tracheostomy should only be performed by experienced operators. The operation should not be done
until pulmonary reserve is adequate to tolerate a
procedure on the airway. In ventilated patients this
means that an FiO2 of 0.5 or less achieves a PaO2
greater than 80 mmHg and peak inflation pressures are
less than 30 cmH2O. Coagulation status should be
normal.
Surgical tracheostomy may be performed in the
operating room or in the ICU. Percutaneous dilatational tracheostomy techniques (Ciaglia, Firsching and
Syniec, 1985; Griggs et al., 1990) can be performed in
the ICU. They are quick and allow the passage of a
standard 8.0 or 9.0 mm tube. Tracheal stenosis is a
recognized complication of tracheostomy, although
the relative incidence of stenosis following the percutaneous procedure compared to surgical tracheostomy
is not yet known (McFarlane et al., 1994).
Irrespective of the route of intubation, cuffed tubes
must be inflated with sufficient volume to seal the
airway and prevent aspiration. Cuffs may develop
leaks as a result of changing cuff compliance, tracheal
dilatation with prolonged intubation or tube movement (Kearl and Hooper, 1993). Excessive cuff pressure may cause tracheal ulceration and later stenosis.
Frequent volumetric cuff checks and cuff pressure
measurements should be performed to ensure an
adequate and safe seal (Rischbieth, Blight and
Myburgh, 1994).
(b)
Principles of ventilation
Ventilation must provide adequate oxygenation and
control of arterial carbon dioxide tension.
Morbidity and mortality of patients supported by
mechanical ventilation, especially those receiving prolonged support, remain very high. There has been
increasing concern that iatrogenic problems relating to
mechanical ventilatory support may be important
contributing factors (MacIntyre, 1993).
Accidental disconnection from mechanical ventilators is the most common and potentially lethal
complication of mechanical ventilation. All ventilators must have a dedicated disconnection alarm.
Other complications of ventilation are patient–
ventilator dyssynchrony, alveolar overdistention and
adverse effects of intrinsic positive end-expiratory
pressure (auto PEEP) .
Patient–ventilator dyssynchrony occurs either
when the breath delivered by the ventilator is out of
cycle with the spontaneous effort by the patient or
when the response of the ventilator is not adequate to
MAINTENANCE OF CARDIOPULMONARY–CEREBRAL HOMEOSTASIS
345
meet the flow demands of the patient’s effort. This
results in an increase in the pressure load placed on
the respiratory muscles and a substantial increase in
oxygen consumption and carbon dioxide production.
This may augment respiratory muscle fatigue and
failure (Mador, 1991; Oh et al., 1991). This is frequently
and incorrectly described as the patient ‘fighting the
ventilator’. The reverse is more appropriate: in reality
the ventilator is fighting the patient. The resulting
sympathetic overactivity may elevate ICP (Ersson et
al., 1990). Similar effects occur as a result of coughing
induced by endotracheal suctioning or of Valsalva
maneuvers in restless, agitated patients (Walsh et al.,
1989).
Optimizing the patient–ventilator interface requires
adequate sedation and the use of a mode of ventilation
that reduces the work of breathing.
Sedation is best achieved initially with a rapid onset
sedative/narcotic combined with a muscle relaxant,
such as fentanyl with pancuronium. Thereafter, combined infusions of a narcotic with a benzodiazepine
(e.g. morphine and midazolam) should be titrated
against the conscious state of the patient (Fisher and
Raper, 1991). The degree of sedation required may
vary widely according to the severity of the underlying head injury and individual pharmacokinetic
differences. The total dose of drug infused is not
important, but rather the titrated effect.
Propofol, a short-acting anesthetic induction agent
given by infusion, is becoming popular for the
treatment of head-injured patients. It has the theoretical advantage of a rapid and potent titrated sedation
without prolonging ventilation. Consciousness
returns rapidly on cessation of the drug because of its
rapid distribution and metabolism. (Farling, Johnston
and Coppel, 1989a, b). Propofol is a potent myocardial
depressant and cerebral perfusion should be monitored carefully. Definitive studies analyzing the longterm use of this agent as a sedative have yet to be
carried out.
Alveolar overdistention
There is now a clearer understanding of the pathophysiological consequences of alveolar overdistention in
mechanically ventilated, critically ill patients. Previously, patients were ventilated with tidal volumes
much larger (10–15 ml/kg) than those occurring in
spontaneously breathing individuals. This is the
standard method of ventilating patients during general anesthesia and was advised in order to prevent
atelectasis by promoting recruitment of collapsed
alveoli, to negate the adverse effects of increased dead
space and to increase functional residual capacity.
There is increasing evidence that this technique is
dangerous in critically ill patients with injured lungs
or in those who require prolonged (> 24 h) ventilation.
The consequences include initiating or exacerbating
an acute lung injury (Tsumo, Prato and Kolobow, 1990;
MacIntyre, 1993), barotrauma to the lung and impairment of cardiac output (Dreyfuss and Saumon, 1992).
Pulmonary edema due to increased permeability has
also been demonstrated in human and animal studies
after several hours of ventilation with high tidal
volumes (Dreyfuss and Saumon, 1993). This phenomenon, termed ‘volutrauma’, is related to the degree of
alveolar volume and to distention rather than to peak
airway pressure.
In patients and animals with an acute lung injury,
large tidal volumes increase edema accumulation (and
extravascular lung water), leading to increased shunt,
impairment of oxygenation and a significant reduction
in pulmonary compliance (Carlton et al., 1990; Sohma
et al., 1992; Tsumo, Prato and Kolobow, 1990).
The inflating airway pressure is transmitted along
interstitial bronchovascular planes and may result in
pneumothorax, pneumomediastinum, pneumopericardium or pneumoperitoneum and increase in intrathoracic pressure. This will reduce venous return to
the heart, increase the impedance to ventricular
ejection and reduce cardiac output (Marini, 1990).
Neuromuscular blockade
(c)
The use of non-depolarizing muscle relaxants in
critically ill, ventilated patients should be questioned
for two reasons. Firstly a paralyzed but awake patient
is more susceptible to autonomic sympathetic swings
which will exacerbating raised ICP and predispose to
increased oxygen consumption. Secondly, a polyneuropathy has been demonstrated in ventilated
patients who have received regular non-depolarizing
muscle relaxants with or without steroids. This axonal
degeneration significantly prolongs ventilation and is
associated with increased extracranial complications
and morbidity (Gooch et al., 1991; Hansen-Flaschen,
Cowen and Raps, 1993; Hsiang et al., 1994).
Positive pressure ventilation is the routine form of
mechanical ventilation in current practice. Gas is
delivered under positive pressure to either a preset
pressure or preset volume. The mode selected for any
patient should be based upon the lung mechanics of
that patient (Bone and Eubanks, 1991; Sassoon, 1991;
Burchardi, Sydow and Criee, 1994). For most patients
with normal lungs a volume control mode will suffice.
Volume-controlled ventilation delivers a preset
tidal volume at a preset rate. This mode of ventilation
is suitable in patients with normal lung mechanics but
may cause barotrauma and alveolar overdistention in
patients with poorly compliant lungs. The mean
Modes of mechanical ventilation
346
RESPIRATORY AND CARDIOVASCULAR SUPPORT
inspiratory pressure may be reduced (pressure-limited
volume control ventilation) to prevent barotrauma,
but at the expense of the tidal volume. This mode
ignores spontaneous respiratory efforts by the patient
and should be used only in paralyzed, deeply unconscious or anesthetized and sedated patients. Most
portable transport ventilators are of this type and
patients being transported usually need muscle paralysis and increased sedation to optimize ventilation
(Phillips and Skowronski, 1986).
Assisted mechanical ventilation of various types is
designed to optimize the patient–ventilator interface
by using a volume control mode and allowing the
patient to trigger the ventilator to deliver a preset tidal
volume. This is accomplished by generating a small
negative airway pressure or flow that activates the
inspiratory cycle, the duration of which is determined
either by the ratio of inspiration to expiration (I:E
ratio) or by reaching a pressure limit. The sensitivity of
the triggering mechanism must be such that the
inspiratory efforts of the patient are detected without
imposing an increased work of breathing and oxygen
consumption, but not so sensitive that the ventilator
will cycle in response to fluctuations in airway
pressure.
Synchronized intermittent mandatory ventilation
(SIMV) is the standard form of assisted volume
controlled ventilation. This mode sets the mandatory
minute volume delivered to the patient while reducing dyssynchrony by only delivering ventilated
breaths at the end of a spontaneous expiration. The
degree of mandatory ventilation is adjusted according to the spontaneous minute volume of the patient.
Initially the ventilator respiratory rate may be set
high enough to provide most, if not all, of the
patient’s minute volume and then reduced as sedation is reduced or the underlying pathology
improves. The potential benefits of SIMV include less
patient ventilator dyssynchrony, lower sedation
requirements, reduced mean airway pressure by
combining spontaneous and ventilated breaths, and
improved respiratory muscle function by allowing
patients to breathe spontaneously (Bone and
Eubanks, 1991). Disadvantages may include insufficient mandatory minute volume in unstable patients
and the possibility of respiratory muscle fatigue in
patients with inappropriately low ventilator rates. As
with any volume control mode of ventilation, alveolar overdistention is a potential hazard; this can be
minimized by using an upper peak inspiratory pressure limit above which the ventilator will not deliver
a breath (usually 30–40 cmH2O).
It has become routine to combine inspiratory
pressure support ventilation (PSV) with SIMV in
order to reduce the work of breathing through
ventilator and endotracheal tubing during the sponta-
neous breaths of the patient (Bersten et al., 1989). In
this mode, spontaneous inspirations are augmented
with a level of positive airway pressure that is preset
(usually to 20 cmH2O) to achieve a desired tidal
volume. As patients are weaned from SIMV by
reducing the ventilator rate and/or tidal volume, the
inspiratory pressure support will continue so that
patients set their own minute volume without incurring increased respiratory effort. Pressure support
ventilation alone is used chiefly to ensure an adequate
minute volume during the weaning phase and in
patients with adequate respiratory drive and minimal
sedation. It is not suitable for the acute phase of injury,
where control of minute ventilation and particularly
of PaCO2 is desired (Brochard et al., 1989; Santak et al.,
1991; Sassoon et al., 1991).
Pressure controlled ventilation delivers gas until a
preset peak inspiratory pressure is reached at a given
ventilator rate (Blanch et al., 1993a, b). The tidal
volume is determined by the patient’s compliance and
the I:E ratio. Increasing the inspiratory time in this
mode has been advocated so that the I:E ratio is
reduced (pressure control inverse ratio ventilation –
PC-IRV) to progressively recruit collapsed alveoli.
Inspiratory pressure therefore remains constant in
order to maximize alveolar ventilation at the lowest
inflation pressure, not only through the whole period
of inspiration, but also when lung compliance
improves. The shortened expiratory time is intended
to prevent alveoli collapsing again and usually induces a higher auto PEEP. This mode of ventilation
remains controversial and manipulations of the I:E
ratio are best combined with frequent reassessment of
both lung mechanics and auto PEEP and the effects on
cardiac output and oxygen transport (Marcy and
Marini, 1991; Mercat et al., 1993).
(d)
Positive end-expiratory pressure (PEEP)
Auto PEEP
In addition to the factors outlined above, a subtle yet
potentially hazardous effect is the presence of an
elevated positive pressure on exhalation known as
intrinsic or auto PEEP (Schmidt and Wood, 1993).
Auto PEEP is seen typically in patients with airflow
obstruction, but it can also occur without obstruction
when minute ventilation is very high. Patients susceptible to the development of auto PEEP are those with
asthma, acute on chronic respiratory failure or acute
lung injury, where a prolonged expiratory time is
necessary.
Auto PEEP was first described in ventilated patients
(Pepe and Marini, 1982), but it can also occur in
spontaneously breathing patients as a result of exhalation against the glottis, varying in magnitude from
MAINTENANCE OF CARDIOPULMONARY–CEREBRAL HOMEOSTASIS
1–10 cmH2O. These patients usually have degrees of
airflow obstruction and the physiological effects of
auto PEEP are those of extrinsically applied PEEP,
namely alveolar recruitment to reduce shunt and
increase functional residual capacity. The work of
breathing to overcome auto PEEP may be significant.
When such patients are ventilated, auto PEEP together
with extrinsically applied PEEP may result in air
trapping at end expiration and the risk of barotrauma,
reduced cardiac output and increased physiological
dead space (Fessler et al., 1991; Pinsky, Desmet and
Vincent, 1992). The level of extrinsic PEEP applied
must be adjusted in accordance with the level of auto
PEEP so that alveolar overdistention and increased
work of breathing do not occur.
Auto PEEP can be dangerous if undetected. It is
therefore essential that auto PEEP be identified and
measured in ventilated patients by stopping air flow
at end expiration just before the next breath, allowing
the pressure in the airways and the ventilator tubing
to equilibrate and reading the pressure from a ventilator pressure gauge. Many new ventilators have the
facility to monitor and display the level of auto PEEP
(Gottfried, Reissman and Ranieri, 1992).
Extrinsic PEEP can be used to maintain alveolar
recruitment and functional residual capacity throughout the ventilatory cycle in order to improve arterial
oxygenation (Schmidt and Wood, 1993). Low levels of
PEEP have also been shown to significantly reduce the
work of breathing imposed by ventilators and endotracheal tubes, particularly when used with inspiratory pressure support (Bersten et al., 1989; Banner,
Blanch and Kirby, 1993). Consequently, PEEP of
5–10 cmH2O is usually added to most ventilatory
modes. Higher levels (more than 10 cmH2O) may
reduce cardiac output and cause barotrauma; they are
only used in patients with severe hypoxia who need
high levels of inspired oxygen.
The application of PEEP above 10 cmH2O should be
titrated against oxygen delivery to provide ‘optimal
PEEP’ (Lin and Oh, 1990). Although the efficacy of
PEEP in preventing atelectasis is not conclusively
proven, it may be of some benefit in neurosurgical
patients who cannot be positioned with head elevation. PEEP can also be used in spontaneously breathing patients, either through a tightly fitting face mask
or an endotracheal tube, in which case it is called
continuous positive airway pressure (CPAP). CPAP is
useful in weaning patients with critical levels of auto
PEEP or with reduced ventricular function. The
withdrawal or reduction of positive pressure ventilation or PEEP from these patients may worsen oxygenation and require hours or even days of therapy to
reverse (Zwissler et al., 1991).
The use of PEEP in the head-injured patient has
been controversial. The rationale for avoiding PEEP
347
was that the increased intrathoracic pressure, transmitted via the venous system to the sagittal sinus and
cerebral veins, would raise intracranial pressure.
However there is evidence that PEEP only increases
intracranial pressure if it reduces systemic arterial
pressure (Frost, 1977; Shapiro and Marshall, 1978).
The mechanism suggested is arteriolar vasodilatation
induced by reduced cerebral perfusion pressure. On
balance, PEEP probably does not adversely effect
cerebrovascular dynamics unless there is an associated reduction in cerebral perfusion pressure, and
PEEP should be used primarily to improve oxygenation in the head-injured patient.
(e)
Control of carbon dioxide
Hyperventilation is of great value when ICP rises
acutely and requires urgent reduction such as during
resuscitation or prior to evacuation of intracranial
hematomas.
Urgent hyperventilation must be employed carefully because vigorous manual inflation (‘bagging’) or
marked elevation of the rate of mechanical ventilation
may decrease mean arterial pressure or increase mean
airway pressures (Ersson et al., 1990). The temporary
benefits derived from hypocapnia-induced vasoconstriction may be lost if systemic arterial pressure and
hence cerebral perfusion pressure are reduced. Furthermore excessive vasoconstriction due to hyperventilation may result in a reduction of oxygen
delivery to ischemic levels. Hyperventilation during
early diagnosis and management should be used for as
short a time as possible. If diagnostic studies show no
mass lesion, and intracranial pressure is low, PaCO2
should gradually be returned to normal.
Prolonged reduction of PaCO2 to 25 mmHg or lower
has been advocated for the last two decades without
specific evidence that it reduces mortality or morbidity. Indeed, clinical studies have suggested that prolonged hyperventilation may be associated with a
worse outcome (Jennett et al., 1980; Proctor et al., 1984;
Muizelaar et al., 1991). Potential detrimental effects of
prolonged hyperventilation include increased cerebral
lactic acidosis and uncoupling of oxidative phosphorylation. Profound reduction in the cytochrome a1 –a3
system has been attributed to cerebral oligemia
induced by prolonged hypocapnia (Proctor et al., 1984;
Rosner, 1987). Furthermore the vasoconstriction
induced by hypocapnia is temporary. Animal and
human studies have shown that despite continuous
hyperventilation to maintain a PaCO2 of 25 mmHg
there is a steady loss of pial arteriolar constriction and
a return of vessel diameters to baseline by 20 hours.
During this period, arterial and cerebrospinal fluid pH
return to normal. Return to normocapnia will now
result in vasodilatation, an increase in cerebral blood
348
RESPIRATORY AND CARDIOVASCULAR SUPPORT
volume and rise in intracranial pressure (Muizelaar et
al., 1988; 1991; Bouma and Muizelaar, 1992).
Hence continuous hyperventilation is probably only
effective for 20–24 hours and subsequent attempts to
return to normal PaCO2 may increase intracranial
pressure. Therefore if hyperventilation is used initially, it should only be continued until the level of
intracranial pressure can be determined. Thereafter it
should be used to treat acute elevations in intracranial pressure while other measures such as osmotherapy, ventricular drainage and the evacuation of
mass lesions are being initiated (Pitts and Andrews,
1992). Routine prolonged hyperventilation, still practiced in many units, is difficult to justify.
(f)
Management of ventilated patients
Once mechanical ventilation is established, meticulous
attention to oxygenation and ventilation is essential. A
nurse–patient ratio of 1:1 is desirable. In an unstable
patient with multiple injuries, or multiple organ
failure, a ratio of up to 2:1 may be justified. The
methods and principles of respiratory function monitoring described in section 17.2.1 should continue for
the duration of ventilation.
Humidification
In all ventilated patients, inspired gases must be
artificially humidified, since the humidifying functions of the nasopharynx are bypassed by endotracheal intubation. Normal respiratory humidification is important for conserving heat and water.
Humidification and warming continue along the
airways so that alveolar gas is 100% saturated at 37°C.
Effective mucociliary function depends on a relative
humidity of 75–100% irrespective of temperature. In
ventilated patients, the relative humidity of inspired
gases falls to 50% or less and rapid heat loss can result
in hypothermia. Cold, dry gases in the trachea and
bronchi increase mucous viscosity and inspissation,
leading to mucosal ulceration, depressed ciliary function and microatelectasis due to obstruction of small
airways. Humidification should provide 75–100%
saturated gas in the trachea at a constant temperature
(32–36°C) without increasing the work of breathing,
dead space or resistance to either spontaneous or
controlled ventilation. Water-bath humidifiers and
aerosol nebulizers or atomizers are commonly used
but they may lead to infection from bacterial contamination of water reservoirs, to overhydration,
overheating and electrical hazards. Most of these
complications can be avoided by using heat and
moisture exchangers attached to the endotracheal
tube. These give safe and efficient humidification and
may be combined with a bacterial filter to maintain
sterile gases (Shelly, Lloyd and Park, 1988).
Patients with reactive airways due to pre-existing
asthma, chronic airway disease or acute bronchospasm from infection or pulmonary edema may
benefit from nebulization of 2-agonists (e.g. salbutamol) to improved gaseous exchange and reduced
inspired airway pressure. The addition of nebulized
anticholinergics (e.g. ipratropium bromide) may benefit patients with severe airflow obstruction, particularly in the weaning phase. The routine use of these
agents in ventilated patients, however, has not been
shown to be of benefit (O’Driscoll et al., 1989).
(g)
Weaning from positive-pressure ventilation
In the head-injured patient, mechanical ventilatory
support can generally be withdrawn when intracranial pressure is consistently less than 25 mmHg and
there is no significant mass effect on CT scan.
Extracranial injuries should also be stable and should
not need inotropic support or preload augmentation.
Lung function, gaseous exchange and lung compliance should approach that of the patient’s presumed premorbid status. The patient should be placed
on assisted ventilation (either synchronized intermittent mechanical ventilation or pressure support) so
that spontaneous efforts can eventually take over
respiration. Sedation and analgesia should be minimal
so that cough and glottic reflexes are present.
As a general rule, weaning is not commenced until
PaO2 is 70–80 mmHg with a FiO2 not greater than 0.4.
The PaO2/FiO2 ratio is a useful measure of the alveolar
arterial oxygen gradient (Table 17.3). Similarly, PaCO2
should be normal, as discussed in section 17.5.1(e).
Pulse oximetry has greatly facilitated assessment of
oxygenation, although the limitations outlined in
section 17.4.2 must be considered; frequent blood gas
analyses should be performed, particularly after a
change in ventilator settings or FiO2 (Linton, 1993).
Numerous indices, listed in Table 17.3, have been
used to predict success in weaning patients from
mechanical ventilation but these should be interpreted
within the specific clinical context. They are accurate
guides, providing the work of breathing through the
endotracheal tube and ventilator tubing is compensated by inspiratory pressure support of approximately 10–15 cmH2O (Banner et al., 1993b). Alternatively, lung volumes can be assessed by connecting the
endotracheal tube to a Wright’s respirometer. Of the
indices of lung mechanics, the ratio of respiratory rate
to expired tidal volume (f/VT) is easily determined
and values greater than 100 may accurately predict
successful weaning (Yang and Tobin, 1991). Lung
mechanics can also be assessed by the simple bedside
demonstration of a strong expulsive cough on tracheal
suction. This implies adequate forced vital capacity,
cough and glottic reflexes.
MAINTENANCE OF CARDIOPULMONARY–CEREBRAL HOMEOSTASIS
Table 17.3 Criteria for weaning head injured patients
from mechanical ventilation. SIMV = synchronized
intermittent mandatory ventilation, PaO2 = arterial oxygen
tension, FiO2 = fractional inspired oxygen concentration,
PaCO2 = arterial carbon dioxide tension, PEEP = positive
end expiratory pressure, VD/VT = dead space/tidal
volume ratio, f/VT = respiratory rate/tidal volume ratio,
PAO2 = alveolar oxygen tension
Clinical markers for
initiating weaning
Appropriate conscious state;
cooperative patient
Appropriate peripheral motor
function
Clinically and radiologically
resolving lung disease
Adequate analgesia
Hemodynamic stability
Metabolic, acid–base stability
Neurosurgical
factors
Stability of extracranial injuries
Control of cerebral perfusion
pressure
Ventilation
parameters
Patient trigger mode (SIMV,
pressure support)
PaO2 > 80 mmHg, FiO2 < 0.4
PaCO2 < 45 mmHg
PEEP < 10 cmH2O
Lung mechanics
and function
Respiratory rate < 35 breaths/
minute
Tidal volume > 4 ml/kg
Minute volume < 15 l/min
Maximum vital capacity > 10–15
ml/kg
Maximum inspiratory pressure <
–20 cmH2O
Static compliance > 35 ml/cmH2O
PaO2/FiO2 ratio > 150
Alveolar–arterial oxygen gradient
< 250 mmHg
VD/VT < 0.6
Shunt fraction (Q̇s/Q̇t) < 20%
Predictors of
weaning success
f/VT ratio < 100
Tidal volume > 4 ml/kg
Static compliance > 35 ml/cmH2O
PaO2/PAO2 ratio > 0.35
T-piece weaning
Patients with normal lungs who have been ventilated
for only a few days may be connected from full
ventilation to a humidified T-piece (‘blow-by system’)
or open ventilator circuit for 20–30 minutes. This is the
usual method of weaning following general anesthesia. In neurotrauma it is suitable for patients with
minor head injuries who have required intubation and
ventilation as part of resuscitation and assessment.
During the period of T-piece weaning, lung mechanics, in particular the f/VT ratio, and gas exchange
should be constantly monitored. If these parameters
are stable the patient can be left on this circuit until
extubation is appropriate.
349
Intermittent T-piece weaning is suitable for
patients with underlying airflow obstruction or
mechanical muscle weakness. The patient is connected
to a T-piece for trial periods, of perhaps 5–30 minutes
every 30–120 minutes. Each trial period is increased
by 5–10 minutes until the patient maintains adequate
mechanics and oxygenation without mechanical ventilation. A balance must be achieved between allowing
the patient to exercise and causing respiratory muscle
fatigue with increased oxygen consumption. T-piece
weaning has been largely superseded by pressure
support ventilation using newer ventilators such as
the Siemens Servo 900C, Engstrom Erica or Puritan
Bennett 7200a (Tomlinson et al., 1989; Bone and
Eubanks, 1991; Oh et al., 1991).
Weaning from synchronized intermittent mandatory ventilation may be accomplished by progressively reducing the ventilator rate and tidal
volume until the patient can maintain an adequate
minute volume with spontaneous respiration. Inspiratory pressure support should be added as soon as the
patient begins spontaneous efforts and then maintained throughout the weaning period. Once the
patient is breathing adequately on pressure support
ventilation alone this may be gradually reduced to
10 cmH2O and the patient extubated or, if tracheostomized, placed on a humidified T-piece.
Extubation may be considered if the level of
consciousness has improved and the pharyngeal and
tracheal reflexes are able to protect the airway from
aspiration of secretions or gastric contents.
With recovery of physical and mental status, a patient
with a tracheostomy may progress from a cuffed to a
non-cuffed or fenestrated tube and then be extubated.
(h)
Complications: nosocomial pneumonia
Ventilated patients are susceptible to respiratory infections. Early pneumonia, defined as respiratory infection already present or developing soon after intubation, may arise from aspiration of gastric contents or
infection of collapsed lobes. Late or true ventilatorassociated (nosocomial) pneumonia is defined as
infection occurring 48 hours after intubation and not
present or incubating at time of intubation (Meduri,
1993).
True ventilator-associated pneumonia is associated
with a significant morbidity and mortality, particularly in those patients who progress to develop an
acute lung injury (section 17.5.1) In patients with
severe head injury who require positive pressure
mechanical ventilation for more than 3 days, the
incidence of pneumonia has been reported to be
35–70%. Nosocomial pneumonia significantly prolongs ventilation and ICU stay and incurs a reported
mortality of up to 50% (Rodriguez, 1991).
350
RESPIRATORY AND CARDIOVASCULAR SUPPORT
Diagnosis of nosocomial pneumonia
The diagnosis and treatment of ventilator-associated
nosocomial pneumonia has been extensively debated
(Torres, Gonzalez and Ferrer, 1991; Meduri and Chastre, 1992; Meduri, 1993) and several criteria have been
proposed. Clinical criteria are fever, leukocytosis,
purulent tracheobronchial secretions and radiographic
evidence of a new or progressive pulmonary infiltrate.
Microbiological criteria are a sputum Gram stain
showing more than 25 polymorphonuclear leukocytes
and more than ten squamous epithelial cells per lowpower field with recovery of a significant pathogen
(by stain or culture). These findings in a previously
healthy person almost invariably indicate pneumonia
and appropriate antibiotics should begin as soon as
sensitivities have been established.
However, purulent secretions are invariably present
in patients during prolonged mechanical ventilation
and are usually not caused by pneumonia. Secretions
may originate from the sinuses, stomach or oropharynx and may accumulate above the endotracheal tube
cuff; these secretions can be aspirated by minor
manipulation of the tube. In addition, the proximal
airways of ventilated patients are colonized early by
potentially pathogenic organisms originating from
hospital cross-infection.
As a result, no combination of clinical variables is
completely accurate in diagnosing pneumonia in
ventilated patients. Postmortem studies have shown
that pneumonia is underdiagnosed in patients with
acute lung injury, but over diagnosed in acute respiratory failure from other causes. Although appropriately
selected antibiotics have been shown to improve
survival, empirical use of broad-spectrum antibiotics
in patients without infection is potentially harmful,
since it facilitates colonization and superinfection by
highly virulent organisms.
In one study, the mortality rate in patients who
received antimicrobial therapy before the onset of
pneumonia was 83% compared with a mortality of
48% in the group who did not receive prior antibiotics
(Meduri, 1993). A culture taken from lung tissue is the
‘gold standard’ in establishing the diagnosis, but is
often not feasible. In most patients quantitative cultures of tracheal aspirates, protected specimen brushing and bronchoalveolar lavage are recommended. It
is important to stress that systemic sources of infection
should be actively sought and treated.
Management
Management of ventilator-associated pneumonia is
twofold. The first essential step is aggressive pulmonary toilet, postural drainage and prevention of gastric
aspiration. Increased sedation may be necessary during
tracheal suctioning to prevent sympathetically mediated swings in blood pressure and intracranial pressure
(Ersson et al., 1990). There is increasing evidence that
ventilated patients nursed at 30° head elevation have a
lower incidence of pulmonary aspiration and improvement in PaO2 as a result of reduced shunting. The
positioning of head-injured patients has been the topic
of some controversy. This is discussed further in section
17.5.2(c). The second step is appropriate antibiotic
cover, using the least toxic bactericidal agent in
appropriate doses, guided by blood level monitoring if
necessary. Antibiotics should only be used when the
criteria for nosocomial pneumonia are fulfilled and the
bacteriological cultures are positive. Sensitivity testing
is the best guide to antibiotic choice.
(i) Complications: acute respiratory distress
syndrome/acute lung injury
The acute respiratory distress syndrome (ARDS) has
been defined as ‘a sudden clinical pathophysiological
state characterized by severe dyspnea, hypoxia, diffuse bilateral pulmonary infiltrations and stiff lungs
following massive acute lung injury, usually in persons with no previous lung injury’ (Ashbaugh et al.,
1967). The hallmark of this syndrome is increased
permeability of the alveolar–capillary endothelium.
The constellation of clinical, radiological and pathophysiological findings, resulting from diffuse injury to
the lung parenchyma, that defines this syndrome has
been the subject of intense debate (Petty, 1990; Oh,
1992; Repine, 1992; McNaughton and Evans, 1992;
Kolleff and Schuster, 1995). The terms acute respiratory distress syndrome or acute lung injury are
generally used interchangeably, but they have recently
been defined in terms of the severity of insult,
specifically the degree of hypoxia; ARDS is now
regarded as the more severe form of acute lung injury
(Bernard et al., 1994).
The insult is not uniform and spares some areas of
lung parenchyma (Figure 17.2). In damaged areas, the
lung is atelectatic, edematous and hemorrhagic.
Hypoxia occurs as a result of increased intrapulmonary
shunting of blood caused by the loss of functional
alveoli and hence reduction in functional residual
capacity. Lung compliance is reduced (< 30 ml/
cmH2O) and physiological dead space is increased,
leading to hypercapnia. Diffuse alveolar and interstitial
infiltrates appear progressively on chest radiographs
and may be quantified in accordance with the severity
of injury (Murray et al., 1988). Microscopic examination
reveals intra-alveolar collections of proteinaceous
fluid, red blood cells and inflammatory cells. Microthrombi or white cell aggregates may be seen in small
vessels. After 24–48 hours, hyaline membranes line the
alveoli. These are formed by fibrin that has escaped
MAINTENANCE OF CARDIOPULMONARY–CEREBRAL HOMEOSTASIS
Figure 17.2 Chest X-ray of a patient with acute respiratory
distress syndrome (ARDS), characterized by bilateral alveolar and interstitial infiltrates. Note the pulmonary artery
catheter inserted via the right internal jugular vein, which
has advanced too far into a segment of the right pulmonary
artery.
through the capillary walls. Subsequently, as repair of
the injury occurs, fibrosis may ensue.
Some but not all patients with ARDS develop
dysfunction or failure of one or more organ systems,
either sequentially or simultaneously. By contrast,
other patients develop multiple organ failure without
ARDS, although they may have less severe degrees of
parenchymal lung injury. This suggests that ARDS is
the respiratory manifestation of multiple organ failure syndrome, just as septic shock is the cardiovascular manifestation (Bone et al., 1992b).
The patient with head injury is at risk of developing
ARDS as a result of neurotrauma per se – neurogenic
pulmonary edema is a form of ARDS – or of
extracranial factors such as lung contusion, multiple
transfusion, burns, pneumonia, pancreatitis, aspiration of gastric contents and sepsis.
Treatment of ARDS
Treatment of patients with ARDS is primarily supportive and aims to maintain oxygen delivery to all organ
systems. Most patients with ARDS who die do so from
351
multiple organ failure or sepsis rather than respiratory
impairment (MacNaughton and Evans, 1992). Hence
the underlying cause must be treated and suspected
sites of sepsis need to be managed aggressively with
appropriate antibiotics and surgical drainage. Careful
fluid management is necessary to maintain euvolemia
without increasing lung water (Humphrey et al., 1990;
Hudson, 1992). Pulmonary artery catheters may be
used to facilitate the measurement of oxygen delivery
(Russell, Graybeal and Strout, 1990; Mitchell, 1992).
Pressure-limited ventilation and PEEP, as outlined in
section 17.5.1(d), form the mainstay of respiratory
support.
Permissive hypercapnia has been shown to reduce
mortality in ARDS (Hickling, 1990). This employs
pressure-limited ventilation at the expense of tidal
volume and a rise in PaCO2. In these studies PaCO2 was
allowed to rise to greater than 80 mmHg, without
adverse effects, despite incurring a moderate respiratory acidosis. However, this strategy may not be feasible
in head-injured patients, where control of PaCO2 is
essential for controlling ICP and CPP (section 17.5.1(e)).
The prognosis of ARDS depends mainly on the
cause and the presence of multiple organ failure. Fat
embolism and neurogenic pulmonary edema may
cause a severe but transient ARDS, which often has a
good outcome. Gram-negative sepsis and shock states
have a higher incidence of multiple organ failure and
a higher mortality.
Despite improvements in the understanding of
pathophysiological pathways and in mechanical support for patients with ARDS and the multiple organ
failure syndrome, no definitive ‘magic bullet’ treatment is available and mortality remains high at
40–70% (Kraus et al., 1993).
(j)
Complication: pulmonary thromboembolism
Fatal pulmonary embolus occurs in 2–3% of patients
with head and multiple injuries. It is usually caused
by thrombosis of the iliofemoral veins. Deep venous
thrombosis in the legs may occur in 30–40% of these
patients but has a lower incidence of embolism.
High-risk factors include a previous history of
thromboembolism, major orthopedic (especially pelvic) injuries, other major trauma, prolonged coma and
hemiparesis. Prophylactic measures includes early
mobilization, graduated compression stockings, pneumatic calf stimulators and low-dose anticoagulants
such as heparin, low-molecular-weight heparin or
warfarin. Although bleeding is an uncommon complication of low-dose anticoagulants, it is prudent to
monitor coagulation profiles in patients with neurotrauma, giving anticoagulants only during the convalescent phase because of the risk of intracranial
bleeding (Moser, 1990; Sudlow et al., 1992).
352
RESPIRATORY AND CARDIOVASCULAR SUPPORT
In high-risk patients or those with proven thromboembolism in whom anticoagulation is contraindicated, early placement of vena caval filters has been
advocated (Persson, Davis and Villavicencio, 1991).
(k) Effects of pre-existing physical limitations of
the chest on weaning
Patients with massive obesity and kyphoscoliosis may
require pressure-limited ventilation. These patients
and those with respiratory muscle weakness due to
myopathies, neuropathies or paralysis may need
prolonged weaning. Pleural and parenchymal pathology limits alveolar ventilation by restricting lung
expansion and by increasing dead space. These disorders also increase the work of breathing, thereby
increasing oxygen consumption and respiratory muscle fatigue.
Drainage of the pleural space, usually by means of
tube thoracostomy, is indicated in patients compromised by pneumothorax, hemothorax or pleural
effusion.
Nutrition, introduced early and preferably by the
enteral route, will improve respiratory muscle function to a limited extent. It is associated with a
reduction in nosocomial pneumonia and preservation
of gut integrity (Chapter 16).
17.5.2 HEMODYNAMIC MANAGEMENT OF THE
ACUTE HEAD-INJURED PATIENT IN THE ICU
Cardiovascular stability in patients with neurotrauma
is of particular importance because of the need to
maintain an adequate cerebral perfusion as well as
myocardial, renal and splanchnic perfusion (Fessler
and Diaz, 1993).
Vital organ perfusion depends on an adequate
vascular volume, hence accurate monitoring of vascular volume and systemic blood pressure is essential.
As discussed in section 17.4.5(a), mean arterial pressure should be monitored continuously, usually with
an intra-arterial catheter.
(a)
Volume status monitoring
Right atrial pressure (central venous pressure) will
suffice to monitor volume status in most patients but
more specific monitoring of cardiac output may
sometimes be indicated. As with arterial pressure
measurement, this is most reliable with electronic
transducers.
Right ventricular preload is measured by inserting
a catheter into the superior vena cava and measuring
mean right atrial pressure when the tricuspid valve
is open. The pressure measurement provides an
approximation of right ventricular preload if ventricular compliance is normal, as it usually is in
young patients.
The response of right atrial pressure to a fluid
challenge will yield useful information regarding the
volume status. This is particularly helpful when right
atrial pressure is less than normal (approximately
5 mmHg) as pressure in the right ventricle is seldom
much higher than in the left. In this context, a low
right atrial pressure suggests decreased intravascular
volume.
Factors that influence the accuracy of right atrial
pressure measurements include right heart dysfunction, superior vena caval obstruction, tricuspid valve
disease, cardiac tamponade and raised intrathoracic
pressure (right atrial pressure may be elevated when
left ventricular pressure is normal).
Left ventricular end-diastolic pressure and volume
may be assessed with right atrial pressures if right
and left ventricular pressures are similar in diastole.
This is usually the case in patients with normal left
ventricular compliance. However, in patients with
poor left ventricular compliance or mitral valve
disease, left ventricular pressure may be elevated
when right ventricular pressure is normal. In these
patients, in older patients and in those with multiple
organ dysfunction, direct left-sided measurements
may be appropriate for assessing left ventricular
pressure.
Pulmonary artery pressure monitoring is widely
used in the management of critically ill patients.
However there has been strong opinion expressed
against its routine use (Robin, 1985). Conversely,
protagonists claim improved outcome in selected
patients (Shoemaker et al., 1988; Mitchell, 1992).
Left ventricular preload may be estimated by
passing a balloon-tipped catheter into the right-sided
pulmonary circulation. With the balloon inflated, the
catheter traverses the right atrium, right ventricle
and main pulmonary artery until it occludes a
branch of the pulmonary artery (Figure 17.2). Flow
distal to the balloon ceases and the pressure measured at the catheter tip distal to the balloon reflects
the pulmonary artery occlusion or ‘wedge’ pressure
downstream. This approximates to left atrial pressure
and is equivalent to left ventricular end-diastolic
pressure when the mitral valve is open, assuming
that pulmonary venous pressure is not elevated and
that the mitral valve is normal. The left ventricular
end-diastolic pressure is assumed to approximate to
left ventricular end-diastolic volume and therefore to
left ventricular preload. With the balloon deflated, a
pulmonary arterial pressure waveform is continually
displayed. When the catheter is correctly placed right
atrial pressure is measured simultaneously from a
port in the right atrium. The pulmonary artery
MAINTENANCE OF CARDIOPULMONARY–CEREBRAL HOMEOSTASIS
353
Table 17.4 Hemodynamic indices obtained from pulmonary artery catheters. MAP = mean arterial pressure, SBP =
systolic blood pressure, DBP = diastolic blood pressure, CO = cardiac output, BSA = body surface area, RAP = right
atrial pressure, MPAP = mean pulmonary artery blood pressure, PAOP = pulmonary artery occlusion pressure.
Variable
Formula
Mean arterial pressure (MAP)
MAP = DBP +
Cardiac index (CI)
CI =
Stroke volume index (SVI)
SVI =
Systemic vascular resistance index (SVRI)
Pulmonary vascular resistance index (PVRI)
Left ventricular stroke work index (LVSWI)
Right ventricular stroke work index (RVSWI)
Indications for pulmonary artery catheterization in a
head-injured patient (Table 17.4)
There are no specific indications for pulmonary artery
catheters in managing patients with head injury. They
are most often used for acute intercurrent problems.
Clearly they should only be used when the information derived from them will significantly improve
management and the benefits outweigh their risks.
Situations in which they may be used include:
3
CO
CO
HR BSA
CI
79.98
MPAP – PAOP
measurement of right heart pressures (right atrial
pressure, pulmonary artery pressure)
– ARDS
– pulmonary hypertension
– pulmonary embolism
– cardiac tamponade;
estimation of left heart filling (pulmonary artery
occlusion pressure)
– left ventricular failure
– response to fluid loading;
hemodynamic measurement (cardiac output, stroke
volume, systemic vascular resistance)
60–90 mmHg
2.5–3.5 l/min/m2
BSA
79.98
CI
LVSWI = (MAP – PAOP) SVI 0.0136
RVSWI = (MPAP – RAP) SVI 0.0136
PVRI =
catheter may also be used to measure cardiac output
by the thermodilution method using a thermistor at
the distal end. New rapid-response catheters are able
to determine continuous right ventricular ejection
fraction and cardiac output (Dorman et al., 1992). A
wide range of cardiorespiratory variables (Table 17.4)
can be derived from these measurements, some of
which have been discussed in section 17.2.3, and
they play an essential part in providing optimal
oxygen delivery.
SBP – DBP
MAP – RAP
SVRI =
Normal values
33–47 ml/beat
1700–2400 dyn.s/cm5/m2
150–250 dyn.s/cm5/m2
45–60 g.m/m2/beat
7–12 g.m/m2/beat
– quantification of shock states (cardiogenic, septic,
hypovolemic)
– assessment of response to treatment in the
above;
derivation of oxygen variables (V̇O2, ḊO2);
measurement of intracardiac shunt (acute ventricular septal defect).
(b)
Clinical use of central venous catheters
In interpreting central venous and pulmonary artery
catheter measurements, account must take of the
factors which alter the gradient between the pulmonary artery occlusion pressure and left ventricular end
diastolic pressure (Nadeau and Noble, 1986). Intravascular pressures must be referenced to the extravascular pressures around them.
Pulmonary artery occlusion pressure is lower than
true ventricular filling pressure in a spontaneously
breathing patient during inspiration when pleural
pressure may be greatly negative. Conversely, pulmonary artery occlusion pressure is higher than the true
filling pressure in a mechanically ventilated patient
during an inspiration with positive pressure. To avoid
erroneous interpretation, the pulmonary artery occlusion pressure should be measured in end-expiration.
Ventilated patients should remain on the ventilator
during the measurement.
Other pulmonary factors that affect the pressure
measurements include catheter-tip location in regions
of the lung where alveolar pressure exceeds arterial
pressure, PEEP and raised intrapleural pressure. Cardiac factors include mitral or aortic valve disease and
poor left ventricular compliance.
354
RESPIRATORY AND CARDIOVASCULAR SUPPORT
(c) Contraindications to central venous
catheterization
These include:
lack of suitable vascular access;
untreatable bleeding disorders;
patient instability preventing adequate time for the
procedure.
(d)
Complications of central venous catheters
These include:
trauma to structures at insertion (pneumothorax,
arterial
laceration,
hematoma,
hemothorax,
chylothorax);
arrythmias on passage of the catheter through the
heart;
perforation of great vessels and atrium;
migration into a distal vessel, which may cause
vessel rupture or lung infarction.
These complications may be avoided by determining
the position of the catheter tip on the chest radiograph
(Figure 17.2), by continuous monitoring of the pulmonary artery pressure waveform and by inflating the
balloon with only 1–1.5 ml of air to obtain a pulmonary artery occlusion pressure. (Hagley et al., 1992).
The optimal placement for right atrial pressure catheters is within the superior vena cava 2 cm above the
pericardial reflection (McGee et al., 1993).
Central venous catheters are a major source of
nosocomial infection and carry a significant morbidity
and mortality. In general catheters inserted during
resuscitation should be changed as soon as possible
unless strict asepsis was used in their insertion. The
incidence of infection increases markedly after 5 days
and varies with the site of the catheter, increasing
from subclavian through to internal jugular to femoral venous catheters (Norwood et al., 1991; Cobb et
al., 1992; Mermel and Maki, 1994).
A persistent or new pyrexia, leukocytosis and an
inflamed insertion site strongly suggest catheter
sepsis and warrant a new catheter. Culture of the
intradermal portion and tip of the catheter may
confirm an infection, although removal of the catheter
may be all that is necessary for treatment (Goldmann
and Pier, 1993).
(e)
Effects of therapy on cerebrovascular function
Until recently, head injury management was focused
on reducing ICP and cerebral blood volume, particularly by osmotherapy, hyperventilation and barbiturate coma. Each of these methods may in part reduce
ICP by reducing mean arterial pressure and therefore
CPP. Recent studies have questioned their efficacy and
emphasis is now being placed more on maintaining
CPP (Stone, 1989; Paulson et al., 1990; Bouma et al.,
1992; Miller et al., 1992; Fessler and Diaz, 1993;
Myburgh and Lewis, 1996).
The hemodynamic effects of ventilation on ICP and
CPP must be considered. Increased intrathoracic
pressure may reduce venous return and hence preload. At the same time reduction of ventricular
transmural pressure may reduce afterload. Thus in
hypovolemic patients ventilation may increase hypotension and reduce cerebral perfusion. Conversely,
patients with cardiogenic or neurogenic pulmonary
edema often improve dramatically with ventilation
because of the mechanical reduction in preload and
afterload and increase in cardiac output and cerebral
perfusion.
Impedance to right ventricular ejection may be
increased by high intrathoracic pressures or by preexisting or acute pulmonary hypertension, resulting in
interventricular septal shift, increased left ventricular
end diastolic volume and reduced stroke volume
(Zwissler et al., 1991; Pinsky, Desmet and Vincent,
1992). As discussed in section 17.5.1(b), the effects of
different modes of ventilation on auto PEEP must also
be considered.
CPP-based algorithms are discussed in Chapter 19;
they have altered the role of osmotherapy and elective
dehydration. When the blood–brain barrier is intact,
mannitol, a six-carbon non-metabolic sugar similar to
glucose, remains within the intravascular space. It may
have several modes of action. These include reduction
in cerebrospinal fluid formation, reduction in brain
tissue volume (Donato et al., 1994), an immediate
increase in circulating blood volume and arterial blood
pressure (Roberts et al., 1987), reduction of blood
viscosity by hemodilution (Muizelaar, Lutz and Becker,
1984), increase in red blood cell deformability (Burke et
al., 1981) and scavenging of potentially toxic free
radicals (Fisher and Raper, 1988). Experimental studies
have shown that cerebral vasoconstriction occurs in
direct response to the decrease in blood viscosity after
mannitol infusion (Muizelaar et al., 1983). In other
words, a constant CBF is maintained by vasoconstriction if autoregulation is intact. Other studies have
supported the vasoconstrictor action of mannitol.
Intracranial pressure responded better to mannitol in
patients with a low initial cerebral perfusion pressure,
who are likely to be vasodilated, than those with a high
cerebral perfusion pressure where the cerebral vasculature is likely to be already fully vasoconstricted
(Rosner and Coley, 1987)
When the clinical goal is euvolemia and adequate
cerebral perfusion, osmotherapy may be seen as a
method of selectively reducing brain volume while
avoiding dehydration (Miller, Piper and Dearden,
1993).
MAINTENANCE OF CARDIOPULMONARY–CEREBRAL HOMEOSTASIS
Repeated administration of mannitol, especially in
combination with frusemide (furosamide), may result
in a systemic hyperosmolar state and dehydration
(Schettini, Stahurski and Young, 1982; Roberts et al.,
1987). Mannitol will eventually cross the blood–brain
barrier into the extracellular space, increasing free
water and cerebral edema. The hemoconcentration
caused by dehydration may falsely suggest that red
cell mass is adequate. Hyperviscosity and microvascular shunting may adversely affect tissue oxygen
delivery.
A substantial part of the action of mannitol depends
on its ability to draw extravascular water into the
vascular space. If extravascular water is markedly
reduced mannitol will have only a minimal effect and
its toxicity will be increased. When dehydration is
well established, the patient will become refractory to
mannitol therapy.
The renal toxicity of virtually all drugs commonly
used in the neurosurgical patient, such as aminoglycoside antibiotics, mannitol, frusemide and contrast
media, as well as transfusion reactions, are increased
by dehydration. Dehydration or hypovolemia makes
patients more liable to unstable and inadequate
systemic and cerebral perfusion and to orthostatic
changes in the blood pressure.
Serum sodium should be kept at 145–150 mmol/l
and serum osmolality should be no greater than
300 mosmol/l (although many authors will accept
320 mosmol/l). Osmolality should be measured
directly, as osmolalities calculated from sodium, urea
and glucose will underestimate true osmolality when
mannitol is used.
Mannitol doses should be limited to the minimum
amount needed to control ICP and CPP and should
not be given routinely with frusemide because of the
risk of excessive dehydration. Dehydrating agents –
not only mannitol – given appropriately can at times
be life-saving, but chronic dehydration is potentially
harmful.
(f)
355
ured at head level then CPP will not increase despite
the fall in ICP. When severe intracranial hypertension
exists, there is little cerebral venous volume available
for displacement and CPP may be higher with
patients nursed flat (Rosner, 1993). In one study, onethird of patients demonstrated an increase in ICP
with head elevation (Feldman et al., 1992). This may
be due to vasodilatation in response to decreased
CPP (Rosner, 1993).
Fluctuations in intracranial pressure and cerebral
perfusion pressure with head elevation may be minimized by maintaining euvolemia and the cardiorespiratory benefits of slight head elevation can then
be utilized.
Barbiturate therapy
There are few options available for treating patients
with raised ICP refractory to the above measures.
Sodium pentobarbital at doses causing ‘burst suppression’ on the EEG will lower intracranial pressure.
However, these doses of barbiturates cause profound
myocardial depression and hypotension and reduce
cerebral perfusion pressure. Theoretically, the reduction in cerebral perfusion pressure and blood flow
may be offset by the reduction in cerebral oxygen
consumption induced by barbiturates. However two
randomized clinical trials have found no improvement in morbidity, mortality, or intracranial pressure
control when barbiturates where given (Ward et al.,
1985; Eisenberg et al., 1988). Barbiturates are still
sometimes used when all other measures have failed.
A recent concept of aggressively maintaining cerebral
perfusion pressure with inotropes and vasopressors
and using barbiturates to lower intracranial pressure
has yet to validated (Miller et al., 1992).
Barbiturate therapy has not been shown definitively to alter outcome in neurotrauma. In the group
of patients refractory to other means of treatment, the
persistent intracranial hypertension probably reflects
the severity of the primary injury (Chapter 19).
Head elevation
Head elevation to about 30° is widely used in nursing
critically ill patients with pulmonary congestion from
cardiac failure or pulmonary edema. Oxygenation is
improved by increasing functional residual capacity
and chest wall compliance and by reducing cephalad
movement of the diaphragm. It also reduces the risk of
aspiration of gastric contents.
In patients with head injury, head elevation may
reduce ICP by reducing cerebral venous pressure.
Rosner and Coley (1986) found that ICP was
7–10 mmHg higher in 50–70% of patients nursed flat
compared with 30–50° head elevation. However, if
there is an equivalent fall in arterial pressure meas-
(g)
Hemodynamic manipulation
In managing systemic arterial pressure it is important
to consider that systemic hypertension may be a
response to intracranial hypertension and therefore
necessary to maintain an adequate cerebral perfusion
pressure (Bouma et al., 1992).
Patient manipulations such as endotracheal and
nasotracheal suction, neurological examination, painful stimulation and even family visits, may increase
ICP via sympathetic stimulation. This is unlikely to be
harmful unless ICP is already raised. In this situation
the increases can be prevented by increasing sedation
and analgesia.
356
RESPIRATORY AND CARDIOVASCULAR SUPPORT
In current clinical practice optimal cerebral perfusion pressure is 60–70 mmHg (Saul and Ducker, 1982;
Tsutsumi et al., 1985), although levels of
80–100 mmHg have recently been proposed as providing better control of raised ICP. These levels of CPP
may be achieved by manipulating volume status and
by the use of vasopressors (Rosner, 1993).
Fluid management of head-injured patients has
focused for the last 20 years upon reducing intracranial pressure via strategies of elective dehydration
and osmotherapy. This was based on the concept that
less water available in the body meant less water
available to potentiate cerebral edema. There is,
however, no evidence to substantiate this hypothesis
and indeed the practice may caused more harm than
good. Attention has now shifted to maintaining brain
perfusion.
The primary goal of cerebral perfusion pressure
fluid management is euvolemia, but avoiding systemic overhydration (Marmarou, 1992). A normally
hydrated, euvolemic patient is more hemodynamically stable and has fewer blood pressure changes
associated with ventilator manipulation and less need
for vasopressor support. Furthermore, adequate perfusion of the kidneys minimizes the potentially
nephrotoxic effects of mannitol, frusemide and
sepsis.
Establishing euvolemia, however, requires great
care. The assessment of volume status depends on
clinical markers such as blood pressure, heart rate,
fluid balance and a measurement of preload (as
discussed in section 17.5.2(a)). Biochemical markers
such as serum sodium, urea, creatinine and osmolality
have been advocated but usually change late with
evolving volume depletion and are frequently a poor
guide to volume status. Reaching and maintaining
euvolemia depends on a number of factors, which
should be considered in concert.
Maintenance fluids
Balanced salt solutions maintain normal hydration and
replace insensible losses from pyrexia and ventilation.
When osmotic diuretics and inotropes are used (see
below), urine output may increase to 100–200 ml/h,
resulting in dehydration. Fluids must be adjusted to
maintain a normal volume status as apparently
adequate fluid intake alone does not guarantee adequate vascular volume. It is possible for a patient to
have interstitial overhydration with an inadequate
intravascular volume.
Hypovolemic patients generally require colloid solutions. The best vascular volume expander in an anemic
patient is red blood cells. Serum albumin as either
normal (5%) or concentrated albumin is also useful.
Synthetic colloids such as polygeline are suitable,
although they have a short half-life and may interfere
with clotting. (The fluid management of head-injured
patients is discussed further in Chapter 16.)
Vasopressors
Augmenting mean arterial pressure and cardiac output by inotropes and vasopressors is standard practice
in managing critically ill patients with sepsis and
cardiac failure and is now more widely used to
maintain adequate cerebral perfusion in patients with
head injury (Chapter 19).
Euvolemia is an essential prerequisite for the use of
these agents. As discussed above, a hypovolemic
patient is susceptible to wide swings in blood pressure
and these may be accentuated by vasopressors. The
ideal inotrope and vasopressor is one that acts
rapidly and has a short half-life, thereby allowing
titration of dose against a desired end-point, which is
usually mean arterial pressure and/or oxygen
delivery.
Inotropes increase ventricular contractility by acting on myocardial alpha- and beta-receptors. They
include adrenalin (epinephrine), noradrenalin (norepinephrine), dopamine, dobutamine and isoprenaline
(Prichard et al., 1991; Runciman and Morris, 1993a).
The principal hemodynamic effect of inotropes is an
increase in cardiac output. Effects on peripheral
vasculature are variable and may be either vasoconstriction or vasodilatation. The catecholamines (adrenalin, noradrenalin and dopamine) are increasingly
regarded as first-line inotropes. They exert betaadrenergic effects predominantly at low doses( i.e.
increased heart rate, vasodilatation, inotropy and
bronchodilation) and alpha-adrenergic effects (vasoconstriction, inotropy) at higher doses (Moran et al.,
1993). Tachycardia and dysrhythmias are infrequent
(Runciman and Morris, 1993a).
Coronary perfusion and myocardial oxygen delivery
depend on aortic diastolic pressure and are at risk in
states of low systemic vascular resistance such as sepsis
(Schreuder et al., 1989). Patients with low cardiac
output and high systemic vascular resistance, as in
cardiogenic shock, are best treated initially with a
catecholamine to improve inotropy and maintain
coronary perfusion. The use of dobutamine and
isoprenaline to treat cardiogenic shock has been
increasingly questioned because of a high incidence of
tachycardia and reduction of aortic diastolic pressure.
The effects of ‘low-dose’ dopamine, frequently
advocated to improve renal blood flow, are unsubstantiated. Any such effects are probably not specific
but mediated via increased cardiac output (Duke and
Bersten, 1992).
Vasoactive drugs include vasoconstrictors (phenylephrine, metaraminol) and vasodilators (sodium
CONCLUSION
nitroprusside, glyceryl trinitrate, dobutamine, isoprenaline). Vasoconstrictors may increase afterload on
the left and right ventricle. If right atrial and pulmonary artery occlusion pressure appear to be normal in
the face of vasoconstrictor therapy, vascular volume
may need to be increased.
Although the role of vasopressors in neurotrauma is
not fully established, the rationale for their use is the
loss of the capacity to maintain cerebral perfusion
pressure by cerebral autoregulation.
In severe brain injury cerebral perfusion may
become ‘pressure-passive’, i.e. directly dependent on
systemic arterial pressure. There may be marked
regional differences so that the relationship is difficult
to quantify. Vasoactive drugs should be titrated
against a selected cerebral perfusion pressure. The
degree to which the cerebral vasculature is directly
affected by these vasoactive agents is unclear.
Catecholamine infusions must be given through a
dedicated central venous line and patients receiving
these infusions must be appropriately monitored.
Adrenalin or noradrenalin may be infused using a
solution of 6 mg drug/100 ml saline and dopamine as
400 mg/100 ml. At these concentrations, 1 ml/h approximates 1 g/min. Infusion usually begins at
3–5 g/min and is then titrated against mean arterial,
intracranial and cerebral perfusion pressures.
Hypotensive agents
When cerebral autoregulation is lost, severe systemic
hypertension (i.e. mean arterial pressures greater than
120–140 mmHg) may potentiate vasodilatation and
cerebral hyperemia, theoretically increasing cerebral
edema and intracranial hypertension (Paulson et al.,
1990). Thus control of severe systemic hypertension
may be necessary. This is best achieved initially by
sympathetic blockade with sedation or analgesia.
Hypotensive agents must be used with caution so as
to avoid hypotension and reduction in cerebral
perfusion.
Beta-adrenergic antagonists are effective in controlling sympathetically induced hypertension. They act
primarily at the beta-1 receptors in the heart, resulting
in negative chronotropy and inotropy. They inhibit
sympathetically-mediated release of renin from the
juxtaglomerular cells of the kidney, and this also
assists in lowering blood pressure. Thus beta-blockers
may be indicated when hypertension is due to
increased sympathetic tone with tachycardia and
increased cardiac output, as in severe head injury
(Robertson et al., 1983). Esmolol is ideally suited for
this. It has an extremely short half-life which allows
accurate titration by intravenous infusion at doses up
to 300 mg/kg/min (Hansson, 1991). All beta-blockers
may potentiate cardiac failure and bronchospasm and
357
must be used carefully in susceptible patients (Runciman and Morris, 1993b). Vasodilators are seldom
indicated in neurogenic hypertension.
Although glyceryl trinitrate and sodium nitroprusside are commonly used in other neurosurgical situations to control blood pressure, their propensity to
increase cerebral blood volume and intracranial pressure, and to cause preferential peripheral vasodilatation
over cerebral vasodilatation (cerebral steal), probably
contraindicates their use in head-injured patients, who
may have increased intracranial pressure (Anile et al.,
1981; Davis et al., 1981; Michenfelder and Milde, 1988).
17.5.3
FUTURE ADVANCES IN NEUROMONITORING
Neuromonitoring after head injury aims to provide
information about intracranial compliance and cerebral oxygen utilization (Lewis, Reilly and Myburgh,
1994). Measuring intracranial pressure alone does not
give any information regarding cerebral oxygen utilization (Miller et al., 1992), cerebral blood flow or
cerebral metabolic rate (Bouma et al., 1992).
Frequently, intracranial hypertension is present
without any mass lesion on CT scan, when the patient
has diffuse cerebral swelling. Hyperventilation and
other methods of reducing intracranial pressure have
been used empirically without there being any conclusive evidence that they improve outcome. Indeed,
they may cause harm.
Jugular venous oxygen saturation, measured by
fiberoptic catheters inserted into the jugular bulb,
allows calculation of arteriovenous differences and
oxygen uptake by the brain. Cerebral venous desaturation may occur as a result of intracranial hypertension, excessive hypocapnia, arterial hypoxia and
systemic hypotension. Preliminary studies suggest
that jugular bulb oximetry used in conjunction with
intracranial pressure and cerebral perfusion pressure
monitoring may allow more precise management of
cardiorespiratory variables so as to optimize cerebral
oxygen utilization (Sheinberg et al., 1992; Lewis, Reilly
and Myburgh, 1995).
17.6
Conclusion
At present little can be done to influence the outcome
from the primary brain injury. The duration of the
period of increased vulnerability to secondary ischemic and hypoxic insults remains unclear but may be
up to 10 days to 2 weeks in some patients. Secondary
insults must be prevented or rapidly detected and
treated. Adequate substrate delivery is the key to
preventing secondary insults. This requires assiduous
attention to cardiopulmonary homeostasis, cerebral
perfusion pressure and oxygen utilization. The interaction of these parameters must be recognized.
358
17.7
RESPIRATORY AND CARDIOVASCULAR SUPPORT
References
Andrews, B. T. (1993) The intensive care management of patients with head
injury, in Neurosurgical Intensive Care, (ed. B. T. Andrews), McGraw-Hill,
New York, pp. 227–242.
Anile, C., Zanghi, F., Bracali, A. et al. (1981) Sodium nitroprusside and
intracranial pressure. Acta Neurochirurgica, 58, 203.
Archie, J. P. (1987) Mathematic coupling of data: a common source of error.
Annals of Surgery, 193, 296–303.
Ashbaugh, D. G., Bigelow, D. B., Petty, T. L. et al. (1967) Acute respiratory
distress in adults. Lancet, ii, 319–323.
Balk, B. and Bone, R. C. (1983) The adult respiratory distress syndrome.
Medical Clinics of North America, 67, 685–700.
Banner, M. J., Blanch, P. B. and Kirby, R. R. (1993) Imposed work of breathing
and methods of triggering a demand-flow, continuous positive airway
pressure system. Critical Care Medicine, 21, 183–190.
Banner, M. J., Kirby, R. R., Blanch, P. B. et al. (1993) Decreasing imposed work
of the breathing apparatus to zero using pressure-support ventilation.
Critical Care Medicine, 21, 1333.
Beckman, D. L., Bean, J. W. and Baslock, D. R. (1974) Neurogenic influence on
pulmonary compliance. Journal of Trauma, 14, 111–115.
Bernard, G. R., Artigas, A., Brigham, K. L. et al. (1994) Report of the American–
European consensus conference on ARDS: definitions, mechanisms, relevant
outcomes and clinical trial coordination. Intensive Care Medicine, 20, 225–232.
Bersten, A. D., Rutten, A. J., Vedig, A. E. et al. (1989) Additional work of
breathing imposed by endotracheal tubes, breathing circuits, and intensive
care ventilators. Critical Care Medicine, 17, 671–677.
Bersten, A. and Sibbald, W. J. (1989) Circulatory disturbances in multiple
systems organ failure. Critical Care Clinics, 5, 233–254.
Blanch, P. B., Jones, M., Layon, A. J. et al. (1993a) Pressure-preset ventilation.
Part 1: Physiologic and mechanical considerations. Chest, 104, 590–599.
Blanch, P. B., Jones, M., Layon, A. J. et al. (1993b) Pressure-preset ventilation.
Part 2: Mechanics and Safety. Chest, 104, 904–912.
Bloomfield, E. L. (1989) Extracerebral complications of head injury. Critical
Care Clinics, 5, 881–892.
Bone, R. C. (1985) Monitoring respiratory and hemodynamic function in the
patient with respiratory failure, in Mechanical Ventilation (eds R. R. Kirby, R.
A. Smith and D. A. Desautels), Churchill Livingstone, New York, pp.
137–170.
Bone, R. C. (1991a) The pathogenesis of sepsis. Annals of Internal Medicine, 115,
457–469.
Bone, R. C. (1991b) Let’s agree on terminology: definition of sepsis. Critical
Care Medicine, 19(7), 973–976.
Bone, R. C. and Eubanks, D. H. (1991) The basis and basics of mechanical
ventilation. Disease-a-Month, 6, 324–406.
Bone, R. C., Balk, R. A., Cerra, F. B. et al. (1992a) Definitions for sepsis and
organ failure and guidelines for the use of innovative therapies in sepsis.
Journal, 55, 123–45.
Bone, R. C., Balk, R., Slotman, G. et al. (1992b) Adult respiratory distress
syndrome. Sequence and importance of development of multiple organ
failure. Chest, 101, 320–26.
Borel, C., Hanley, D., Diringer, M. N. et al. (1990) Intensive management of
severe head injury. Chest, 98, 180–189. (Comment in Chest, 99(6) (1991),
1551–1552).
Bouma, G. J. and Muizelaar, J. P. (1992) Cerebral blood flow, cerebral blood
volume, and cerebrovascular reactivity after severe head injury. Journal of
Neurotrauma, 9, S333–S348.
Bouma, G. J., Muizelaar, J. P., Bandoh, K. et al. (1992) Blood pressure and
intracranial pressure-volume dynamics in severe head injury: relationship
with cerebral blood flow. Journal of Neurosurgery, 77, 15–19.
Breuer, H-W. M., Groeben, H., Breuer, J. et al. (1989) Oxygen saturation
calculation procedures: a critical analysis of six equations for the determination of oxygen saturation. Intensive Care Medicine, 15, 385–389.
Brochard, L., Harf, A., Lorino, H. et al. (1989) Inspiratory pressure support
prevents diaphragmatic fatigue during weaning from mechanical ventilation. American Review of Respiratory Disease, 139, 513–521.
Burchardi, H., Sydow, M. and Criee, C. P. (1994) New ventilatory strategies in
severe respiratory failure, in Current Topics in Intensive Care (eds G. J. Dobb,
J. Bion, H. Burchardi and R. P. Dellinger), W. B. Saunders, London, pp.
81–100.
Burke, A. M., Quest, D. O., Chein, S. et al. (1981) The effect of mannitol on
blood viscosity. Journal of Neurosurgery, 55, 170.
Buss, C. S., Nel, C. J. and van den Heever, C. M. (1990) Incidence of
extracranial injuries in fatal head injury patients. South African Journal of
Surgery, 28, 141–144.
Carlton, D. P., Cummings, J. J., Scheerer, R. G. et al. (1990) Lung overexpansion
increases pulmonary microvascular protein permeability in young lambs.
Journal of Applied Physiology, 69, 577–583.
Cerra, F. B. (1992) Multiple Organ Failure Syndrome. Disease-a-Month, 12,
845–946.
Chesnut, R. M., Marshall, L. F., Klauber, M. R. et al. (1993) The role of the
secondary brain injury in determining outcome from severe head injury.
Journal of Trauma, 34, 216–222.
Ciaglia, P., Firsching, R. and Syniec, C. (1985) Elective percutaneous
dilatational tracheostomy: a new simple bedside procedure; preliminary
report. Chest, 87, 715–719.
Clarke, G. M. (1985) Multiple System Organ Failure. Clinics in Anaesthesiology,
4, 1027–1053.
Clayton, D. G., Webb, R. K., Ralston, A. C. et al. (1991) A comparison of the
performance of 20 pulse oximeters under conditions of poor perfusion.
Anaesthesia, 46, 3–11.
Cobb, D. K., High, K. P., Sawyer, R. G. et al. (1992) A controlled trial of
scheduled replacement of central venous and pulmonary-artery catheters.
New England Journal of Medicine, 327, 1062–1067.
Cockings, J. G. L., Webb, R. K., Klepper, I. D. et al. (1993) Blood pressure
monitoring – applications and limitations: an analysis of 2000 incident
reports. Anaesthesia and Intensive Care, 21, 565–569.
Crosby, E. T. and Lui, A. (1990) The adult cervical spine: implications for
airway management. Canadian Journal of Anaesthesia, 37, 77–93.
Crosby, E. T. (1992) Perioperative haemotherapy: I. Indications for blood
component transfusion. Canadian Journal of Anaesthesia, 39, 695–707.
Cruickshank, J. M., Degaute, J. P., Kuurne, T. et al. (1987) Reduction of stress/
catecholamine-induced cardiac necrosis by beta-selective blockade. Lancet,
ii, 585–589.
Dantzker, D. (1989) Oxygen delivery and utilization in sepsis, in Critical Care
Clinics, (eds R. A. Balk and R. C. Bone), W. B. Saunders, Philadelphia, PA,
pp. 81–98.
Dantzker, D. R., Foresman, B. and Gutierrez, G. (1991) Oxygen supply and
utilization relationships. American Review of Respiratory Disease, 143,
675–679.
Davis, R. F., Douglas, M. E., Heenan, T. J. et al. (1981) Brain tissue pressure
measurement during sodium nitroprusside infusion. Critical Care Medicine,
9, 17.
Demling, R. and Riessen, R. (1990) Pulmonary dysfunction after cerebral
injury. Critical Care Medicine, 18, 768–774.
Dhainaut, J. F., Edwards, J. D., Grootendorst, A. F. et al. (1990) Practical aspects
of oxygen transport: conclusions and recommendations of the Rountable
Conference. Intensive Care Medicine, 16, S179–S180.
Dietrich, K. A., Conrad, S. A., Herbert, C. A. et al. (1990) Cardiovascular and
metabolic response to red blood cell transfusion in critically ill volumeresuscitated nonsurgical patients. Critical Care Medicine, 18, 940–1044.
Dinarello, C. A. (1984) Interleukin-1 and the pathogenesis of the acute-phase
response. New England Journal of Medicine, 311, 1413–1418.
Donato T., Shapira Y., Artu A. et al. (1994) Effect of mannitol on cerebrospinal
fluid dynamics and brain tissue edema. Anesthesia and Analgesia, 78,
58–66.
Dorman, B. H., Spinale, F. G., Kratz, J. M. et al. (1992) Use of a combined right
ventricular ejection fraction-oximetry catheter system for coronary bypass
surgery. Critical Care Medicine, 20, 1650–1655.
Dreyfuss, D. and Saumon, G. (1992) Barotrauma is volutrauma, but which
volume is the one responsible? Intensive Care Medicine, 18, 139–141.
Dreyfuss, D. and Saumon, G. (1993) Role of tidal volume, FRC, and endinspiratory volume in the development of pulmonary edema following
mechanical ventilation. American Review of Respiratory Disease, 148,
1194–1203.
Duke, G. J. and Bersten, A. D. (1992) Dopamine and renal salvage in the
critically ill patient. Anaesthesia and Intensive Care, 20, 277–302.
Eisenberg, H. M., Frankowski, R. F., Contant, C. F. et al. (1988) High-dose
barbiturate control of elevated intracranial pressure in patients with severe
head injury. Journal of Neurosurgery, 69, 15.
Ersson, U., Carlson, H., Mellstrom, A. et al. (1990) Observations on intracranial
dynamics during respiratory physiotherapy in unconscious neurosurgical
patients. Acta Anaesthesiologica-Scandinavica, 34, 99–103.
Farling, P. A., Johnston, J. R. and Coppel, D. L. (1989a) Propofol infusion for
sedation of patients with head injury in intensive care. A preliminary
report. Anaesthesia, 44, 222–226.
Farling, P. A., Johnston, J. R. and Coppel, D. L. (1989b) Propofol infusion
compared with morphine and midazolam bolus doses for sedation of
patients with severe head injuries in the intensive care unit. Journal of Drug
Development, 2(Suppl. 2), 97–98.
Feldman, Z., Kanter, M. J., Robertson, C. S. et al. (1992) Effect of head elevation
on intracranial pressure, cerebral perfusion pressure, and cerebral blood
flow in head-injured patients. Journal of Neurosurgery, 76, 207.
Fessler, H. E., Brower, R. G., Wise, R. A. et al. (1991) Effects of positive end
-expiratory pressure on the gradient for venous return. American Review of
Respiratory Disease, 143, 19–24.
Fessler, R. D. and Diaz, F. G. (1993) The management of cerebral perfusion
pressure and intracranial pressure after severe head injury. Annals of
Emergency Medicine, 22, 998–1003.
Fisher M. M. and Raper, R. F. (1988) The optimization of renal function in
acute illness. Medical Journal of Australia, 149, 546–551.
Fisher, M. M. and Raper, R. F. (1991) Sedation in intensive care, in Update in
Intensive Care and Emergency Medicine 14, Update 1991, (ed. J. L. Vincent),
Springer-Verlag, Berlin, pp. 549–558.
Frost, E. A. M. (1977) Effects of positive end-expiratory pressure on
intracranial pressure and compliance in brain-injured patients. Journal of
Neurosurgery, 47, 195.
REFERENCES
Gentleman, D. (1992) Causes and effects of systemic complications among
severely head injured patients transferred to a neurosurgical unit. International Surgery, 77, 297–302.
Goldmann, D. A. and Pier, G. (1993) Pathogenesis of infections related to
intravascular catheterization. Clinical Microbiology Reviews, 6, 176–192.
Gooch, J. L., Suchyta, M. R., Balbierz, J. M. et al. (1991) Prolonged paralysis
after treatment with neuromuscular junction blocking agents. Critical Care
Medicine, 19, 1125–1130.
Gottfried, S. B., Reissman, J. and Ranieri, V. M. (1992) A simple method for the
measurement of intrinsic positive end-expiratory pressure during controlled and assisted modes of mechanical ventilation. Critical Care Medicine,
20, 621–629.
Graf, C. J. and Rossi, N. P. (1978) Catecholamine response to intracranial
hypertension. Journal of Neurosurgery, 49, 862–868.
Greenberg, A. G. (1990) To transfuse or not to transfuse – that is the question!
Critical Care Medicine, 18, 1045.
Griggs, W. M., Worthley, L. I., Gilligan, J. E. et al. (1990) A simple percutaneous
tracheostomy technique. Surgery, Gynecology, and Obstetrics, 170, 543–545.
Guyton, A. C. (1991) Cardiac output, venous return, and their regulation, in
Medical Physiology, W. B. Saunders, Philadelphia, PA, pp. 221–233.
Hagley, M. T., Martin, B., Gast, P. et al. (1992) Infectious and mechanical
complications of central venous catheters placed by percutaneous venipuncture and over guidewires. Critical Care Medicine, 20, 1426.
Hansen-Flaschen, J., Cowen, J. and Raps, E. C. (1993) Neuromuscular
blockade in the intensive care unit. American Review of Respiratory Disease,
147, 234–236.
Hansson, L. (1991) Review of state-of-the-art beta-blocker therapy. American
Journal of Cardiology, 67, 43B–46B.
Hickling, K. G. (1990) Ventilatory management of ARDS: can it affect
outcome? Intensive Care Medicine, 16, 219–226.
Hill, D. A., Abraham, K. J. and West, R. H. (1993) Factors affecting outcome in
the resuscitation of severely injured patients. Australian and New Zealand
Journal of Surgery, 63, 604–609.
Holland, R., Webb, R. K. and Runciman, W. B. (1993) Oesophageal intubation: an
analysis of 2000 incident reports. Anaesthesia and Intensive Care, 21, 608–610.
Hsiang, J. K., Chesnut, R. M., Crisp, C. B. et al. (1994) Early, routine paralysis
for intracranial pressure control in severe head injury: is it necessary?
Critical Care Medicine, 22, 1471–1476.
Hsieh, A. H., Bishop, M. J., Kubilis, P. S. et al. (1992) Pneumonia following
closed head injury. American Review of Respiratory Disease, 146, 290–294.
Hubmayr, R. D., Abel, M. D. and Rehder K. (1990) Physiologic approach to
mechanical ventilation. Critical Care Medicine, 18, 103–113.
Hudson, L. D. (1992) Fluid management strategy in acute lung injury.
American Review of Respiratory Disease, 145, 988–989.
Humphrey, H., Hall, J., Sznadjer, I. et al. (1990) Improved survival in ARDS
patients associated with a reduction in pulmonary capillary wedge
pressure. Chest, 97, 1176–1180.
Jennett, B., Teasdale, G., Fry, J. et al. (1980) Treatment for severe head injury.
Journal of Neurology, Neurosurgery and Psychiatry, 43, 289.
Kalisky, Z., Morrison, D. P., Meyers, C. A. et al. (1985) Medical problems
encountered during rehabilitation of patients with head injury. Archives of
Physiological Medicine Rehabilitation, 66, 25–29.
Kaufman, H. H., Moake, J. L., Olson, J. D. et al. (1980) Delayed and recurrent
intracranial hematomas related to disseminated intravascular clotting and
fibrinolysis in head injury. Neurosurgery, 7, 445–449.
Kaufman, H. H., Timberlake, G., Voelker, J. et al. (1993) Medical complications
of head injury. Medical Clinics of North America, 77, 43–60.
Kearl, R. A. and Hooper, R. G. (1993) Massive airway leaks: an analysis of the
role of endotracheal tubes. Critical Care Medicine, 21, 518–521.
Kirland, L. L. and Wilson, G. L. (1991) Extracranial effects of acute brain
injury. Problems in Critical Care, 5, 292–306.
Kolleff, M. H. and Schuster, D. (1995) The acute respiratory distress syndrome.
New England Journal of Medicine, 332, 27–37.
Kong, D. L., Prough, D. S., Whitley, J. M. et al. (1991) Hemorrhage and
intracranial hypertension in combination increase cerebral production of
thromboxane A2. Critical Care Medicine, 19, 532–538.
Kraus, P. A., Lipman, J., Lee, C. C. J. et al. (1993) Acute lung injury at
Baragwanath ICU. An eight-month audit and call for consensus for other
organ failure in the adult respiratory distress syndrome. Chest, 103,
1832–1836.
Leach, R. M. and Treacher, D. F. (1992) Oxygen transport: the relation between
oxygen delivery and consumption. Thorax, 47, 971–978.
Lewis, A. R., Austen, K. F. and Soberman R. J. (1990) Leukotrienes and other
products of the 5–lipoxygenase pathway. New England Journal of Medicine,
323, 645–655.
Lewis, S. B., Reilly, P. L. and Myburgh, J. A. (1994) Developments in
intracranial pressure monitoring and investigation of head injured patients,
in Current Topics in Intensive Care (eds G. J. Dobb, J. Bion, H. Burchardi and
R. P. Dellinger), W. B. Saunders, London, pp. 1–19.
Lewis, S. B., Reilly, P. L. and Myburgh, J. A. (1995) Detection of cerebral
venous desaturation by continuous jugular bulb oximetry following acute
neurotrauma. Anaesthesia and Intensive Care, 23, 307–314.
Lillie, P. E. and Roberts, J. G. (1988) Carbon dioxide monitoring. Anaesthesia
and Intensive Care, 16, 41–44.
359
Lin, E. S. and Oh, T. E. (1990) Ventilation – which mode?, in Clinical
Anaesthesiology, International Practice and Research, (ed. G. J. Dobb), Baillière
Tindall, London, pp. 441–473.
Linton, D. M. (1993) Difficult ventilatory weaning. Current Opinion in
Anaesthesiology, 6, 347–353.
Liu, S., Lee, T. and Bongard, F. (1992) Accuracy of capnography in
nonintubated surgical patients. Chest, 102, 1512–1515.
Luce, J. M. (1985) Medical management of spinal cord injury. Critical Care
Medicine, 13, 126.
Luce, J. M. (1993) Cardiopulmonary physiology and management in
neurosurgical intensive care, in Neurosurgical Intensive Care, (ed. B. T.
Andrews), McGraw-Hill, New York, pp. 1–30.
Ludbrook, G. L., Russell, W. J., Webb, R. K. et al. (1993) The electrocardiograph: applications and limitations – an analysis of 2000 incident reports.
Anaesthesia and Intensive Care, 21, 558–564.
McFarlane, C., Denholm, S. W., Sudlow, C. L. M. et al. (1994) Laryngotracheal
stenosis: a serious complication of percutaneous tracheostomy. Anaesthesia,
49, 38–40.
McGee, W. T., Ackerman, B. L, Rouben, L. R. et al. (1993) Accurate placement
of central venous catheters: A prospective, randomized, multicenter trial.
Critical Care Medicine, 21, 1118–23.
McHale, H. G. and Roddie, I. C. (1983) Peripheral lymph flow during
intravenous noradrenaline infusion in sheep (abstract) Journal of Physiology
(London), 334, 50P.
MacIntyre, N. R. (1993) Clinically available new strategies for mechanical
ventilatory support. Chest, 104, 560–565.
MacNaughton, P. D and Evans T. W. (1992) Management of Adult Respiratory
Distress Syndrome. Lancet, 339, 469–472.
Mador, M. J. (1991) Respiratory muscle fatigue and breathing pattern. Chest,
100, 1430–1435.
Marcy, T. W. and Marini, J. J. (1991) Inverse ratio ventilation in ARDS. Chest,
100, 494–504.
Marini, J. J. (1990) Paying the piper – the linkage of alveolar ventilation to
alveolar pressure. Intensive Care Medicine, 16, 73–74.
Marmarou, A. (1992) Increased intracranial pressure in head injury and
influence of blood volume. Journal of Neurotrauma, 9, S327–S332.
Meduri, G. U. (1993) Diagnosis of ventilator-associated pneumonia. Infectious
Disease Clinics of North America, 7, 295–329.
Meduri, G. U. and Chastre, J. (1992) The standardization of bronchoscopic
techniques for ventilator-associated pneumonia. Chest, 102, S557–S564.
Mercat, A., Graini, L., Teboul, J. et al. (1993) Cardiorespiratory effects of
pressure-controlled ventilation with and without inverse ratio in the adult
respiratory distress syndrome. Chest, 104, 871–875.
Mermel, L. A. and Maki, D. G. (1994) Infectious complications of Swan–Ganz
pulmonary artery catheters. Pathogenesis, epidemiology, prevention and
management. American Journal of Respiratory and Critical Care Medicine, 149,
1020–1036.
Michenfelder, J. D. and Milde, J. H. (1988) The interaction of sodium
nitroprusside hypotension and isoflurane in determining cerebral vasculature effects. Anesthesiology, 69, 870.
Miller, J. D., Piper, I. R. and Dearden, N. M. (1993) Management of intracranial
hypertension in head injury: matching treatment with cause. Acta Neurochirurgica (Supplement) (Vienna), 57, 152–159.
Miller, J. D., Dearden, N. M., Piper, I. R. et al. (1992) Control of intracranial
pressure in patients with severe head injury. Journal of Neurotrauma, 9,
S317–S326.
Mitchell, J. P. (1992) Improved outcome based on fluid management in
critically -ill patients requiring pulmonary artery catheterization. American
Review of Respiratory Distress, 145, 990–998.
Mizock, B. A. and Falk, J. L. (1992) Lactic acidosis in critical illness. Critical
Care Medicine, 20, 80–93.
Moran, J. L., O’Fathartaigh, M. S., Peisach, A. R. et al. (1993) Epinephrine as an
inotropic agent in septic shock: a dose-profile analysis. Critical Care
Medicine, 21, 70–77.
Moser, K. M. (1990) Venous thromboembolism. American Review of Respiratory
Disease, 141, 235–249.
Muizelaar, J. P., Lutz, H. A. and Becker, D. P. (1984) Effect of mannitol on ICP
and CBF and correlation with pressure autoregulation in severely headinjured patients. Journal of Neurosurgery, 61, 700.
Muizelaar, J. P., Wei, E. P., Kontos, H. A. et al. (1983) Mannitol causes
compensatory cerebral vasoconstriction and vasodilation in response to
blood viscosity changes. Journal of Neurosurgery, 59, 822–828.
Muizelaar, J. P., van der Poel, H. G., Li, Z. C. et al (1988) Pial arteriolar vessel
diameter and CO2 reactivity during prolonged hyperventilation in the
rabbit. Journal of Neurosurgery, 69, 923–927.
Muizelaar, J. P., Marmarou, A., Ward, J. D. et al. (1991) Adverse effects of
prolonged hyperventilation in patients with severe head injury: a randomized clinical trial. Journal of Neurosurgery, 75, 731–739.
Murray, J. F., Matthay, M. A., Luce, J. M. et al. (1988) An expanded definition
of the adult respiratory distress syndrome. American Review of Respiratory
Distress, 138, 720–723.
Myburgh, J. A. (1991) The clinical utility of oximetry, capnography and
laboratory bicarbonate compared with blood gas analysis. Anaesthesia and
Intensive Care, 19, 295.
360
RESPIRATORY AND CARDIOVASCULAR SUPPORT
Myburgh, J. A. (1992) Derived oxygen saturations are not clinically useful for
the calculation of oxygen consumption. Anaesthesia and Intensive Care, 20,
460–463.
Myburgh, J. A. and Lewis, S. B. (1996) Goal directed therapy in neurotrauma,
in Yearbook of Intensive Care and Emergency Medicine, (ed. J. L. Vincent),
Springer-Verlag, Berlin, pp. 716–730.
Myburgh, J. A., Webb, R. K. and Worthley, L. I. G. (1991) The P50 is reduced in
critically ill patients. Intensive Care Medicine, 17, 355–358.
Nadeau, S. and Noble, W. H. (1986) Misinterpretation of pressure measurements from the pulmonary artery catheter. Canadian Anaesthetists’ Society
Journal, 33, 352–363.
Neil-Dwyer, G., Walter, P., Cruickshank, J. M. et al. (1978) Effect of propranolol
and phentolamine on myocardial necrosis after subarachnoid haemorrhage.
British Medical Journal, ii, 990–992.
North, J. B. and Jennett, S. (1974) Abnormal breathing patterns associated with
acute brain damage. Archives of Neurology, 31, 338–344.
Norwood, S., Ruby, A., Civetta, J. et al. (1991) Catheter-related infections and
associated septicemia. Chest, 99, 968–975.
Nunn, J. F. (1987a) The oxygen cascade, in Applied Respiratory Physiology, 3rd
edn, Butterworths, London, pp. 100–283.
Nunn, J. F. (1987b) Elastic forces and lung volumes, in Applied Respiratory
Physiology, 3rd edn, Butterworths, London, pp. 23–41.
Obrist, W. D., Gennarelli, T. A., Segawa, H. et al. (1979) Relation of cerebral
blood flow to neurological status and outcome in head-injured patients.
Journal of Neurosurgery, 51, 292–300.
O’Driscoll, B. R., Taylor, R. J., Horsley, M. G. et al. (1989) Nebulised salbutamol
with and without ipratropium bromide in acute airflow obstruction. Lancet,
i, 1418–1420.
Oh, T. E. (1992) Defining adult respiratory distress syndrome. British Journal of
Hospital Medicine, 47, 350–353.
Oh, T. E. and Duncan, A. W. (1988) Oxygen therapy. Medical Journal of
Australia, 149, 141–146.
Oh, T. E., Bhatt, S., Lin, E. S. et al. (1991) Plasma catecholamines and oxygen
consumption during weaning from mechanical ventilation. Intensive Care
Medicine, 17, 199–203.
Pansard, J. L., Cholley, B., Devilliers C. et al. (1992) Variation in arterial to endtidal CO2 tension differences during anesthesia in the ‘Kidney Rest’ lateral
decubitus position. Anesthesia and Analgesia, 75, 506–510.
Pauca, A. L., Wallenhaupt, S. L., Kon, N. D. et al. (1992) Does radial artery
pressure accurately reflect aortic pressure? Chest, 102, 1193–1198.
Paulson, O. B., Strandgaard, S., Edvinsson, L. et al. (1990) Cerebral
autoregulation. Cerebrovascular Brain Metabolism Review, 2, 161–192.
Pepe, P. E. and Marini, J. J. (1982) Occult positive end-expiratory pressure in
mechanically ventilated patients with airflow obstruction. American Review
of Respiratory Disease, 126, 166–170.
Persson, A. V., Davis, R. J. and Villavicencio, J. L. (1991) Deep vein thrombosis
and pulmonary embolism. Surgical Clinics of North America, 71,
1195–1209.
Petty, T. L. (1990) Acute respiratory distress syndrome. Disease-a-month, 1,
9–58.
Pfenninger, E. G. and Lindner, K. H. (1991) Arterial blood gases in patients
with acute head injury at the accident site and upon hospital admission.
Acta Anaesthesiologica Scandinavica, 35, 148–152.
Phillips, G. D. and Skowronski, G. A. (1986) Manual resuscitators and
portable ventilators. Anaesthesia and Intensive Care, 14, 306–313.
Piek, J., Chesnut, R. M., Marshall, L. F. et al. (1992) Extracranial complications
of severe head injury. Journal of Neurosurgery, 77, 901–907.
Pinsky, M. R., Desmet, J. and Vincent, J. L. (1992) Effect of positive endexpiratory pressure on right ventricular function in humans. American
Review of Respiratory Disease, 146, 681–687.
Pitts, L. H. and Andrews, B. T. (1992) Intracranial pressure monitoring and
treatment of intracranial hypertension, in Principles of Critical Care Medicine,
(ed. L. Wood), McGraw-Hill, New York, pp. 2985–3011.
Platanias, L. C. and Vogelgang, N. J. (1990) Interleukin-1: biology, pathophysiology, and clinical prospects. American Journal of Medicine, 89, 621–629.
Prichard, B. N., Owens, C. W., Smith, C. C. et al. (1991) Heart and
catecholamines. Acta Cardiologica, 46, 309–322.
Proctor, H. J., Cairns, C., Fillipo, D. et al. (1984) Brain metabolism during
increased intracranial pressure as assessed by niroscopy. Surgery, 96, 273.
Redan, J. A., Livingston, D. H., Tortella, B. J. et al. (1991) The value of
intubating and paralysing patients with suspected head injury in the
emergency department. Journal of Trauma, 31, 371.
Reed, R. L. II (1993) Oxygen consumption and delivery. Current Opinion in
Anaesthesiology, 6, 329–334.
Reinhart, K., Hannemann, L., Kuss, B. (1990) Optimal oxygen delivery in
critically ill patients. Intensive Care Medicine, 16, S149–S155.
Repine, J. E. (1992) Scientific perspectives on adult respiratory distress
syndrome. Lancet, 339, 466–469.
Rischbieth, A. M., Blight, I. L. and Myburgh, J. A. (1994) Efficacy of
endotracheal tube cuff seal in three cuff management techniques. Anaesthesia and Intensive Care, 22, 219–220.
Roberts, P., Pollay, M., Engles, C. et al. (1987) Effect on intracranial pressure of
furosemide combined with varying doses and administration of mannitol.
Journal of Neurosurgery, 66, 440.
Robertson, C. S., Clifton, G. L., Taylor, A. A. et al. (1983) Treatment of
hypertension associated with head injury. Journal of Neurosurgery, 59,
455–460.
Robin, E. D. (1985) The cult of the Swan–Ganz catheter. Annals of Internal
Medicine, 103, 445–449.
Rodriguez, J. L., Gibbons, K. J., Bitzer, L. G. et al. (1991) Pneumonia: incidence,
risk factors, and outcome in injured patients. Journal of Trauma, 31, 907–912.
Rosell, S. (1980) Neuronal control of microvessels. Annual Review of Physiology,
42, 359–371.
Rosner, M. J. (1987) Cerebral perfusion pressure: the link between the cerebral
and systemic circulations, in Cerebral Blood Flow: Physiologic and Clinical
Aspects, (ed. J. H. Wood), McGraw-Hill, New York, pp. 425–448.
Rosner, M. J. (1993) Pathophysiology and management of increased intracranial pressure, in Neurosurgical Intensive Care, (ed. B. T. Andrews),
McGraw-Hill, New York, pp. 57–112.
Rosner, M. J. and Coley, I. (1986) Cerebral perfusion pressure, the ICP and
head elevation. Journal of Neurosurgery, 65, 636.
Rosner, M. J. and Coley, I. (1987) Cerebral perfusion pressure: a hemodynamic
mechanism of mannitol and the postmannitol hemogram. Neurosurgery, 21,
147–156.
Rouby, J. J., Laurent, P., Gosnach, M. et al. (1994) Risk factors and clinical
relevance of nosocomial maxillary sinusitis in the critically ill. American
Journal of Respiratory and Critical Care Medicine, 150, 776–783.
Runciman, W. B., Ilsley, A. H. and Rutten, A. J. (1988) Systemic arterial blood
pressure. Anaesthesia and Intensive Care, 16, 54–57.
Runciman,W. B., Myburgh, J. A. and Upton, R. N. (1990) Pharmacokinetics
and pharmacodynamics in the critically ill, in Clinical Anaesthesiology,
International Practice and Research, (ed. G. J. Dobb), Baillière Tindall, London,
vol. 4, pp. 271–303.
Runciman, W. B., Morris, J. L. (1993a) Adrenoceptor agonists, in Mechanisms of
Drugs in Anaesthesia, (eds S. A. Feldman, W. Paton and C. Scurr), Edward
Arnold, Sevenoaks, pp. 262–291.
Runciman, W. B., Morris, J. L. (1993b) Adrenoceptor antagonists, in
Mechanisms of Drugs in Anaesthesia, (eds S. A. Feldman, W. Paton and C.
Scurr), Edward Arnold, Sevenoaks, pp. 291–306.
Runciman, W. B., Webb, R. K., Barker, L. et al. (1993) The pulse oximeter:
applications and limitations – an analysis of 2000 incident reports.
Anaesthesia and Intensive Care, 21, 543–550.
Russell, G. B., Graybeal, J. M. and Strout, J. C. (1990) Stability of arterial to end
tidal carbon dioxide gradients during postoperative cardiorespiratory
support. Canadian Journal of Anaesthesia, 37, 560–566.
Russell, J. A., Ronco, J. J., Lockhat, D. et al. (1990) Oxygen delivery and
consumption and ventricular preload are greater in survivors than in
nonsurvivors of the adult respiratory distress syndrome. American Review of
Respiratory Distress, 141, 659–665.
Rutten, A. J., Ilsley, A. H., Skowronski, G. A. et al. (1986) A comparative study
of the measurement of mean arterial blood pressure by automatic
oscillometers, arterial cannulation and auscultation. Anaesthesia and Intensive Care, 14, 58–65.
Salmenperä, M., Peltola, K., Takkunen, O. et al. (1983) Cardiovascular effects
of pancuronium and vecuronium during high dose fentanyl anaesthesia.
Anesthesia and Analgesia, 62, 1059.
Santak, B., Radermacher, P., Sandmann, W. et al. (1991) Influence of SIMV plus
inspiratory pressure support on V̇A/Q̇ distributions during postoperative
weaning. Intensive Care Medicine, 17, 136–140.
Sassoon, C. S. H. (1991) Positive pressure ventilation. Chest, 100, 1421–1429.
Sassoon, C. S. H., Light, R. W., Lodia, R. et al. (1991) Pressure-time product
during continuous positive airway pressure, pressure support ventilation,
and T-piece during weaning from mechanical ventilation. American Review
of Respiratory Disease, 143, 469–475.
Saul, T. G. and Ducker, T. B. (1982) Effect of intracranial pressure monitoring
and aggressive treatment on mortality in severe head injury. Journal of
Neurosurgery, 56, 498.
Schettini, A., Stahurski, B. and Young, H. F. (1982) Osmotic and osmotic-loop
diuresis in brain surgery. Effects on plasma and CSF electrolytes and ion
secretion. Journal of Neurosurgery, 56, 679.
Schmidt, G. A. and Wood, L. D. H. (1993) Critical Care Medicine. Journal of the
American Medical Association, 270, 194–196.
Schreuder, W. O., Schneider, A. J., Groeneveld, A. B. J. et al. (1989) Effect of
dopamine vs norepinephrine on hemodynamics in septic shock. Chest, 95,
1282–1288.
Schumacker, P. T. and Cain, S. M. (1987) The concept of a critical oxygen
delivery. Intensive Care Medicine, 13, 223–229.
Schumacker, P. T. and Samsel, R. W. (1989) Oxygen delivery and uptake by
peripheral tissues: Physiology and pathophysiology. Critical Care Clinics, 5,
255–269.
Schumacker, P. T., Rhodes, G. R., Newell, J. C. et al. (1979) Ventilation–
perfusion imbalance after head trauma. American Review of Respiratory
Disease, 119, 33–43.
Shapiro, H. M. and Marshall, L. F. (1978) Intracranial pressure responses to
PEEP in head-injured patients. Journal of Trauma, 18, 254.
Sheinberg, M., Kanter, M. J., Robertson, C. S. et al. (1992) Continuous
monitoring of jugular venous oxygen saturation in head-injured patients.
Journal of Neurosurgery, 76, 212.
REFERENCES
Shelly, M. P., Lloyd, G. M. and Park, G. R. (1988) A review of the mechanisms
and methods of humidification of inspired gases. Intensive Care Medicine, 14,
1–9.
Shoemaker, W. C., Appel, P. L. and Kram, H. B. (1992) Role of oxygen debt in
the development of organ failure sepsis, and death in high-risk surgical
patients. Chest, 102, 208–15.
Shoemaker, W. C., Appel, P. L., Kram, H. B. et al. (1988) Prospective trial of
supranormal values of survivors as therapeutic goals in high-risk surgical
patients. Chest, 94, 1176–1186.
Silverman, H. J. (1991) Lack of a relationship between induced changes in
oxygen consumption and changes in lactate levels. Chest, 100, 1012–1015.
Silverman, H. J. and Tuma, P. (1992) Gastric tonometry in patient with sepsis.
Chest, 102, 184–188.
Simard, J. M. and Bellefleur, M. (1989) Systemic arterial hypertension in head
trauma. American Journal of Cardiology, 63, 32C–35C.
Simmons, R. L., Martin, A. M. Jr, Heisterkamp, C. A. III et al. (1969)
Respiratory insufficiency in combat casualties: II. Pulmonary edema
following head injury. Annals of Surgery, 170, 39–44.
Simpson, D. A., Speed, I. E. and Blumbergs, P. C. (1991) Embolism of cerebral
tissue: a cause of coagulopathy and infarction? Surgical Neurology, 35,
159–162.
Singbartl, G., Cunitz, G. and Hamrouni, H. (1982) Disturbed pulmonary gas
exchange in patients with cerebral trauma. Anaesthesist, 31, 228–233.
Sohma, A., Brampton, W. J., Dunnill, M. S. et al. (1992) Effect of ventilation
with positive end expiratory pressure on the development of lung damage
in experimental acid aspiration pneumonia in the rabbit. Intensive Care
Medicine, 18, 112–117.
Stone, J. L. (1989) Nonsurgical management of increased intracranial pressure.
Seminars in Neurology, 9, 218–224.
Stone, D. J. and Bogdonoff, D. L. (1992) Airway considerations in the
management of patients requiring long-term endotracheal intubation.
Anesthesia and Analgesia, 74, 276–287.
Strieter, R. M., Kunkel, S. L., Bone, R. C. (1993) Role of tumour necrosis factora in disease states and inflammation. Critical Care Medicine, 21,
S447–S463.
Sudlow, M. F. and the Research Committee of the British Thoracic Society
(1992) Optimum duration of anticoagulation for deep-vein thrombosis and
pulmonary embolism. Lancet, 340, 873–876.
Szekely, S. M., Webb, R. K., Williamson, J. A. et al. (1993) Problems related to
the endotracheal tube: an analysis of 2000 incident reports. Anaesthesia and
Intensive Care, 21, 611–616.
Takala, J., Keinanen, O., Vaisanen, P. et al. (1989) Measurement of gas exchange
in intensive care: laboratory and clinical validation of a new device. Critical
Care Medicine, 17, 1041–1047.
Talman, W. T. (1985) Cardiovascular regulation and lesions of the central
nervous system. Annals of Neurology, 18, 1–12.
361
Theodore, J. and Robin, E. D. (1976) Speculations on neurogenic pulmonary
edema (NPE). American Review of Respiratory Disease, 113, 405–411.
Tomlinson, J. R., Miller, K. S., Lorch, D. G. et al. (1989) A prospective
comparison of IMV and T-piece weaning from mechanical ventilation.
Chest, 96, 348–352.
Torres, A., Gonzalez, J. and Ferrer, M. (1991) Evaluation of the available
invasive and non-invasive techniques for diagnosing nosocomial pneumonias in mechanically ventilated patients. Intensive Care Medicine, 17,
439–448.
Tracey, K. J. and Cerami, A. (1993) Tumour necrosis factor: An updated review
of its biology. Critical Care Medicine, 21, S415–S422.
Trauma Committee, Royal Australasian College of Surgeons (1992) Early
Management of Severe Trauma (EMST), Royal Australasian College of
Surgeons, Melbourne, Victoria.
Trunkey, D. (1991) Initial treatment of patients with extensive trauma. New
England Journal of Medicine, 324, 1259–1263.
Tsumo, K., Prato, P. and Kolobow, T. (1990) Acute lung injury from mechanical
ventilation at moderately high airway pressures. Journal of Applied
Physiology, 69, 956–961.
Tsutsumi, H., Ide, K., Mizutani, T. et al. (1985) The relationship between
intracranial pressure, cerebral perfusion pressure and outcome in headinjured patients: the critical level of cerebral perfusion pressure, in
Intracranial Pressure VI, (eds J. D. Miller, G. M. Teasdale and J. O. Rowan),
Springer-Verlag, Berlin, pp. 661–666.
Vincent, J. L. (1990) The relationship between oxygen demand, oxygen
uptake, and oxygen supply. Intensive Care Medicine, 16, S145–S148.
Vincent, J. L. and Bihari, D. (1992) Sepsis, severe sepsis or sepsis syndrome:
need for clarification. Intensive Care Medicine, 18, 255–256.
Walls, R. M. (1993) Rapid-sequence intubation in head trauma. Annals of
Emergency Medicine, 22, 1008–1013.
Walsh, J. M., Vanderwarf, C., Hoscheit, D. et al. (1989) Unsuspected
hemodynamic alterations during endotracheal suctioning. Chest, 95,
162–165.
Ward, J. D., Becker, D. P., Miller, J. D. et al. (1985) Failure of prophylactic
barbiturate coma in the treatment of severe head injury. Journal of
Neurosurgery, 62, 383.
Weismann, S. J. (1939) Edema and congestion of the lungs resulting from
intracranial hemorrhage. Surgery, 6, 722–729.
West, J. B. (1985) Respiratory Physiology – The Essentials, Williams & Wilkins,
Baltimore, MD, pp. 11–67.
Yang, K. L., and Tobin, M. J. (1991) A prospective study of indexes predicting
the outcome of trials of weaning from mechanical ventilation. New England
Journal of Medicine, 324, 1445–1450.
Zwissler, B., Schosser, R., Schwickert, C. et al. (1991) Perfusion of the
interventricular septum during ventilation with positive end-expiratory
pressure. Critical Care Medicine, 19, 1414–1424.