|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
PEDIATRIC HEAD TRAUMA: CEREBRAL PERFUSION PRESSURE AS AN
INDICATOR OF OUTCOME
LESLEY CYNTHIA MORGAN
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Lesley Cynthia Morgan
This dissertation is dedicated to my parents who always believed.
I would like to thank my committee chairman, Dr. James Jessup, for his guidance,
encouragement, and patience throughout this research study. I have appreciated his
knowledge of physiology, nursing and his commitment to provide leadership for this
I gratefully acknowledge and extend my appreciation to the members of my
committee, Charles Wood, PhD., Claydell Home, PhD., and Hossein Yarandi, PhD.
Each of them provided unique talents, time, and moral support during this study. I
would like to thank Dr. Wood for his support during my physiology classes, Dr. Yarandi
for his patience in dealing with the data and more importantly for improving my
understanding of statistical analysis and Dr. Home for continual calm and encouragement
and editorial expertise.
I want to thank Joseph Tepas, M.D., and Pamela Pieper, R.N., M.S.N., for their
assistance with the Pediatric Trauma Registry, the DELTA outcome score and for sharing
their tremendous knowledge of the pediatric trauma patient. I also want to thank and
offer my undying appreciation to Shannon Bourne, L.P.N., for helping me weave through
the obstacle course that is the IRB.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ................................................................................................. iv
L IST O F FIG U R E S .............. ............................ ............. ........... ... ........ viii
ABSTRACT .............. .......................................... ix
1 IN TR OD U CTION ............................................... .. ......................... ..
Background of the Problem ........................................................... .................
History of Pediatric Traum a Care....................................................................... 2
E pidem biology ....................................................... 2
Pathophysiology of Pediatric Head Injury .........................................................4
M echanisms of Injury ........... ..... ..... .... ............ .... ..........7.
Mechanical Causes of Brain Injury .............. ....................... ...............8
P pattern of B rain Injury ............. .... .......................................... ........ .... ........
M ortality and M orbidity ......................................................... ...............
C clinical P presentation ........ ............................................................ .... .... ..... 10
Purpose of the Study ................................................. ........ .............. .. 11
C onceptual F ram ew ork ............................................. ........................................ 12
Significance of the Study ............................................................................ .... .......12
S tu dy Q u e stio n s ................................................................................ 14
D definition of Term s ..... ...................... ....................... .... .... .. ............ 14
G lasgow C om a Scale............ .................................. ........ .......... .... ......... 14
DELTA Disability and Injury Score......................................... ............... 14
G lasgow O utcom e Scale.............................................. ............................ 15
C erebral B lood F low ............................................ .. .. .. ...... ........... 15
Cerebral Blood Volume ......... ...................... ............... 15
Cerebral Metabolic Rate of Oxygen....................... ......................................16
Arteriovenous Difference of Oxygen ................................ ...... ............... 16
V olum e-Pressure Index ............................ .............................................16
2 REVIEW OF LITERA TURE ...................... .... ................................ ............... 17
Physiology of Cerebral Circulation and Metabolism ...........................................17
Cerebral Edema .................... .... ..................... .........19
Intracranial Pressure ................................_.. ............ .. ........ .. ........ .. 19
C erebral Perfusion Pressure ........................................ .......................... 25
C erebral A utoregulation ............................................. ............................. 35
P oiseu ille's L aw .............................................................4 0
C erebral C om pliance ........................................... .......... .. ........ .. ........ .... 41
M annitol and Fluid B balance ........................................ .......................... 42
H y p erc ap n ia ............... ... ... ..... .. .......... ........................................ 4 4
Physiology of Pediatric Cerebral Circulation and Metabolism.............................46
Pediatric Traum atic Brain Injury ...................................................... ..... .......... 46
D iffuse B rain Sw selling .................................................................................. 50
H y p erem ia ....................................................................... 5 1
H yperventilation ...... ... ....... .......... ................... .... ..... ..............54
Sum m ary ............... ..................................... ........................... 54
3 M E T H O D S ........................................................................................................... 5 7
R research D design ................................... .. .. ........ .. ............57
R e se arch S ettin g ................................................................................................... 5 7
S am p le S iz e ................................................................5 7
S am p le C criteria ............. ................ ............. ................................ .... 57
M measures .......... ......... .. ................ ............... 58
P ediatric B lood Pressure........................................................... ............... 58
M ean Arterial Pressure (M AP)................................ ......................... ........ 59
Intracranial Pressure .......................... ......... ........ .. .... .......... 59
C erebral Perfusion Pressure ........................................ .......................... 60
H heart R ate (Pulse)....... ............................ .. ................ .......... .... 60
T em perature ................... ........... .............. .. ........................ 60
D ata C collection Procedure ...................... .... .................. ................... ............... 60
D ata C collection .......................................... .. ..... ............... ............ 61
Procedure for the Protection of Human Subjects ........................................ 62
D ata A nalysis................................................... 62
4 ANALYSIS AND RESULTS.......................................... ................... ............... 63
Subject Dem graphics ................. .. ...... .. .. ................. .............. 63
Subject Age, Sex and Ethnicity ......................................63
M echanism of Injury .................................................... ........ ....... ............65
R research F in ding s........... .................................................................... ........ .. ...... .. 66
R research Q uestions............ ................................................................ ........ .. ... 66
5 CONCLUSIONS AND RECOMMENDATIONS............................................... 70
D iscu ssion of F in din g s ......................................... .......................... ....................7 0
C erebral Perfusion Pressure............................................... ............................. 70
C o n c lu sio n s........................................................................................................... 7 5
L im station s of the Stu dy ..................................................................... ..................76
Study Design Limitations ................. .................................. 76
Statistical A analysis Lim itations................................ ................................... 77
Strengths of the Study ............................................... .............. .............. 77
Recom m endations for Further R research ........................................ .....................77
Im plications for C clinical Practice ................................................................... ... ..78
A DATA COLLECTION TOOL ............................................................................80
B GLA SGOW COM A SCALE (G CS) ............................................... .....................83
C MODIFIED GLASGOW COMA SCORE FOR INFANTS .....................................85
D DELTA DISABILITY AND INJURY SCORE ................................................. 86
E GLASGOW OUTCOME SCALE ....................................................... ............. 87
L IST O F R E F E R E N C E S ........................................................................ .....................88
BIOGRAPHICAL SKETCH ............................................................. ...............100
LIST OF FIGURES
4-1. A ge D distribution ............ ..................... ............ ............. .... ...... 64
4-2. Distribution of DELTA severity and outcome score ................... .............. 64
4-3. Distribution of Admission GCS Score.............. ............................ ............... 65
4-4 M mechanism of Injury ........................................................................ .................. 66
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
PEDIATRIC HEAD TRAUMA: CEREBRAL PERFUSION PRESSURE AS AN
INDICATOR OF OUTCOME
Lesley Cynthia Morgan
Chair: James Jessup
Major Department: Nursing
Brain injury is a common and devastating event in the United States impacting both
the adult and the pediatric populations. In the pediatric population there are over 100,000
children under the age of 15 treated annually for neurological trauma with many suffering
significant long-term disability and death. There are extensive data and research
evaluating the impact of cerebral trauma in the adult population. Optimization of
cerebral perfusion pressure (CPP) has gained recognition as a therapeutic endpoint in the
management of brain-injured adults. This study evaluated the relationship between CPP
and intracranial pressure (ICP) in children with severe traumatic brain injury. This study
focused on the effectiveness of perfusion of injured neurons and the relationship to
Fifty-five children, aged 1.5 to 15 years, admitted for severe blunt head trauma
(Glasglow Coma Score < 8) were retrospectively studied looking at systolic blood
pressure, heart rate, temperature, mean arterial pressure, intracranial pressure, and
cerebral perfusion pressure as an indicator of outcome as measured by the DELTA injury
and disability score. Demographic data collected included age, sex, ethnicity, and
admission Glasgow Coma Scale score. Data were analyzed using the SAS software.
Results indicate that the first 48 hours after an injury is the critical period to influence
outcome with mean arterial pressure, intracranial pressure, and cerebral perfusion
pressure significant indicators of outcome (p <0.0005).
When subjects were divided into two groups, children who died and children who
survived, temperature was significant (p < 0.0005) for the first four days post injury.
Temperature continued to maintain significance when subjects were divided into
survivors with poor outcome and good outcome. The relationship between outcome and
therapeutic interventions is unclear and requires further evaluation in a larger prospective
Background of the Problem
The management of traumatic brain injury (TBI) is a fairly new discipline. Until
40 years ago TBI was regarded as an untreatable insult. Since that time, treatment for
TBI has improved significantly. The improvement in therapy is credited to emerging
techniques and concepts in intracranial pressure control and cerebral perfusion
management. However, there is little consensus regarding the management of these
patients (Chestnut, 1997b). This problem is exacerbated when discussing the
neurotrauma care of children. From research now available, there are strong supporting
data from prospectively collected, observational studies that there is a biological
advantage toward favorable outcome associated with younger age. There is also
evidence that the path of physiological events underlying the brain's response to injury in
children is somewhat different than seen in adults (Aldrich et al., 1992; Bruce et al.,
1979; Bruce, Schut, Brumo, Wood, & Sutton, 1978; Luerssen, Klauber, & Marshall,
1988; Muizelaar et al., 1989; Obrist, Langfitt, Jaggi, Cruz, & Gennarelli, 1984).
Injury is the leading cause of death among children and adolescents in the United
States. Death from unintentional injury for age groups 1 to 4 years, 5 to 9 years, 10 to 14
years, and 15 to 19 years was 36%, 43%, 39%, and 46% respectively for the year 1998.
This represents a total drop in the death rate of 38% between the years 1979 and 1998
(Guyer et al., 1999). The impact of childhood injury is immense in terms of direct costs
to society and the tremendous emotional toll of death and disability. Brain injury is a
common and devastating event. In the pediatric population, there are over 100,000
children under the age of 15 treated annually for neurologic trauma with many suffering
significant long-term disability and death (Guyer & Ellers, 1990; Lehr & Baethmann,
1997; Tepas, DiScala, Ramenofsky, & Barlow, 1990).
History of Pediatric Trauma Care
In 1917 a French munitions ship and a Norwegian freighter collided causing an
enormous explosion at a narrow point in the harbor of Halifax, Nova Scotia, Canada.
Two thousand were killed, 9,000 injured and over 31,000 left homeless. Pleas for
medical help were issued throughout Canada and the United States. A health care team
from Boston led by William E. Ladd responded. Dr. Ladd was moved by the special
medical needs of the children and upon his return to Boston dedicated himself to the
surgical care of infants and children. This is considered the birth of pediatric surgery as
an independent surgical specialty. At this time infectious diseases were the most lethal
childhood illnesses. But by 1947, with the advent of sulfa drugs, penicillin, and smallpox
and polio vaccines, the death rate from these illnesses began to fall. During the 1940s,
trauma became the leading cause of death in children (Dietrich, 1954; Godfrey, 1937;
Goldbloom, 1986; Press, 1947).
In 1937 Edward Godfrey, M.D., commissioner of the New York State Department
of Health, wrote of the failure of health care workers to give important consideration to
childhood accidents. "What is the net gain if a child, through breast feeding and
pasteurized milk, is prevented from dying of gastroenteritis if he pulls a stewpan of
boiling water off the stove and is fatally scalded? What use to protect him against
diphtheria to be killed by an automobile" (p. 153).
Today, trauma remains the major cause of death in children between the ages of 1
and 14, with head injury accounting for 40 % of fatal childhood injuries. Head injury
accounts for 100,000 pediatric hospitalizations per year in the United States. The
incidence of head injury is approximately 200 per 100,000, with an ensuing mortality rate
of 10 %. For comparison, the next leading cause of death in the pediatric age group is
leukemia, with a rate of death of approximately 2 deaths per 100,000 (Francel, Park,
Shaffrey, & Jane, 1996). When the pediatric population is evaluated as a whole, the most
common cause of head injury are falls at approximately 35% of patients and motor
vehicle accidents at 25% of patients. When severe trauma is isolated from the above
numbers, motor vehicle accidents account for about 75% to 80% of the injuries. Falls
drop to 15% (Waxweiler, Thurman, Sniezek, Sosin, & O'Neil, 1995).
Bicycle accidents are a common cause of traumatic injury in children and young
adolescents. Bicycle accidents resulted in more than 400,000 emergency department
visits and 500 to 600 deaths in the United States in 1986. In 1982, 70% of all bicycle-
related injuries occurred in children younger than 15 years of age. Most of these severe
accidents involved a motor vehicle. Similar to bicycle accidents are pedestrian injuries to
children between the ages of 1 to 14 years. Fatal pedestrian injuries are more common
than fatal passenger injuries in preschool and school-aged children (Campbell, 1992;
Pautler, Henning, & Buntain, 1995).
Young people are at risk of head injury as pedestrians, as cyclists, and as occupants
of motor vehicles. The risks and consequences vary with age and maturity (Simpson,
Blumbergs, McLean, & Scott, 1992). Infant and child safety is an ongoing concern. The
advent of seat belts with the subsequent addition of shoulder straps or the three point
restraints decreases the fatality rate of accidents by 40% to 50% and the severity of injury
by 55% to 60% (Rivera, 1999). Car seats, when used properly with a five-point restraint
system, make fatality 11% less likely. The use of bike helmets by children has
significantly decreased severe head injury (DiGuiseppi, Rivara and Koepsell, 1990;
Goldsmith, 1992). A study by Spaite, Murphy, Criss, Valenzuela, and Meisen (1991)
showed that not wearing a helmet in a serious crash was strongly associated with major
head injury in 22% of all patients evaluated, whereas, only 1 in 116 patients wearing
helmets during a crash had major head trauma. A mandatory bicycle helmet law
instituted in Victoria, Australia, reduced the number of cyclists admitted to the hospital
with a marked reduction in the proportion with head injuries when compared with the
same period in the year before (Ryan, 1992; Vulcan, Cameron, & Watson, 1992).
Pathophysiology of Pediatric Head Injury
One of the most striking differences between the pediatric and the adult brain is the
size, both in terms of the absolute size of the brain and its size relative to the rest of the
body. At birth the brain comprises 15% of total body weight decreasing to 3% of total
body weight in an adult. The brain grows rapidly relative to the rest of the body during
childhood, reaching 75% of adult size by the age of two and 90% of adult size by the age
of six (Ward, 1995).
In the young child, the skull is also different from that of the older child and adult.
Up until the age of three the skull has unfused sutures. The skull is thinner and more
pliable. This pliable, thin skull can expand markedly to avoid compression or expansion
of the brain. However this pliability of the immature skull makes the brain more
vulnerable to injury. Direct blows to the skull tend to cause more focal deformation, and
to a limited extent absorb more of the force of the impact and thus convey less force of
the blow to areas of the brain remote from the area of impact (Francel et al., 1996;
McLaurin & Towbin, 1990).
The young brain has significantly less myelination than that of the adult.
Recticular system myelination continues on into adult life. While there is limited
research on the elastic properties of brain tissue in the young, it does appear that
unmyelinated brain tissue of the infant and young child is more susceptible to shearing
injuries than that of the older child or adult. An additional physiological difference that
makes the young brain susceptible to head injury is the concept of plasticity of the central
nervous system. This idea is based on the observation that young children are able to
sustain injuries and recover function, whereas, their older counterparts are not (Ward,
At birth, cerebral perfusion appears to be fairly homogenous, but regional patterns
later emerge that reflect the ongoing sequence of functional maturation that occurs in the
brain. Cerebral blood blow changes are linked to maturation that occurs in the brain.
Regional blood flow changes are linked to changes in metabolism. Regional variation in
blood flow may explain altered susceptibility of certain areas of the developing brain to
head trauma. However, the immature brain appears to tolerate anoxia and hypoxia better
than the adult brain (Francel et al., 1996).
The immature brain has unique characteristics with regard to brain edema. The
immature brain shows less edema than the mature brain after significant head trauma.
Speculation of the reason for this difference ranges from differences in the blood brain
barrier to the lower mean arterial pressure seen in younger animals. Brain edema fluid
clears more rapidly in the immature brain (Ghajar & Hariri, 1992).
Cerebral perfusion pressure, which is the difference between the mean arterial
pressure and the intracranial pressure, appears to be lower in the child than in the adult.
Multiple studies on adults have indicated a minimal cerebral perfusion pressure of 50 mm
Hg is required to maintain adequate perfusion with 70 mm Hg felt to be optimum to
maintain adequate perfusion of the brain so that deleterious decompensation does not
occur (Shapiro & Smith, 1993). The mean systemic blood pressure in children is lower;
thus, the cerebral perfusion pressure in children may be normal below 50 mm Hg
(Rosner, Prineas, Loggie, & Daniels, 1993). The actual minimum cerebral perfusion
pressure for the immature brain has not been determined and may vary with the child's
When discussing the skull and brain of children, it is important to note that
whenever a child suffers a head injury, there is damage done not only to the current
structures but also to the process by which the immature nervous system progresses to a
mature state. The central nervous system has a very difficult time regenerating itself.
While there is evidence that some regeneration can take place, it is difficult for
significant damage to be repaired. It is therefore more imperative to prevent damage than
recoup lost function (Ward, 1995).
In general terms, the brain must have enough cerebral blood flow to meet its
metabolic needs. Cerebral blood flow is felt to be between 50-100 mm Hg in the young
child and adult. It is probably lower in the newborn (birth to 1 month) and infant (1
month to 12 months). Currently there is no definitive answer to what constitutes
adequate cerebral blood flow. The answer appears to vary with age, metabolic demands,
and other factors. Most researchers believe that ischemic damage can occur when
cerebral blood flow falls below 18 to 20 ml per 100 gm of brain tissue per minute. There
are four variables that have a significant effect on cerebral blood flow: (1) systemic blood
pressure, (2) arterial blood gases, (3) metabolic demands of the brain, and (4) intracranial
pressure (ICP). Blood pressure and cerebral blood flow can be evaluated through the
relationship of cerebral perfusion pressure (Mean arterial pressure (MAP) Intracranial
pressure (ICP) = Cerebral perfusion pressure (CPP) (Ward, 1995).
Mechanisms of Injury
As previously stated, motor vehicle accidents comprise the majority of severe head
trauma seen in the pediatric population. A motor vehicle accident can be viewed as a
series of four accidents, each with the capacity of causing serious injury. First, there is
the initial collision event, which occurs when the vehicle strikes or is struck by an object.
There is an abrupt change in velocity and direction in the movement of the vehicle.
Deformation of the vehicle may lead to direct injury of the occupant. The second
potential collision may occur as the occupant strikes the interior of the vehicle. This is
the most common cause of injury in a motor vehicle accident and the target of belt
restraint systems (Pautler et al., 1995).
The human body is triphasic, with solid, liquid, and gaseous systems combining to
provide function and protection. The brain is continually bathed in cerebral spinal fluid.
During a collision the head undergoes a sudden change in velocity. While the head
strikes or is stuck by an object, the change in motion is not immediately transferred to the
brain which continues forward until it strikes the anterior aspect of the cranial vault. This
is the third potential collision, as solid organs strike the limits of their confining space.
The most commonly seen phenomenon is the coup-contre-coup mechanism of injury.
The brain strikes the occipital aspect of the skull as the head accelerates; when the head
then reflects in the opposite direction, the brain strikes the frontal aspects of the cranium
(Pautler et al., 1995).
The fourth accident is when objects such as children's toys, automobile parts, or
groceries are unrestrained within the vehicle. These items may become projectiles during
a collision, striking an occupant and causing further injury (Pautler et al., 1995).
Mechanical Causes of Brain Injury
Traumatic brain injury may be classified into three mechanical causes: (1) an
impact defined as the collision of the head with a solid object at an appreciable velocity,
(2) an impulsive load which produces sudden motion of the head without significant
physical contact, and (3) a static or quasistatic loading situation in which the
consequences of speed of occurrence may be neglected. The first mechanism, impact
loading, causes brain injury through a combination of contact forces and inertial forces.
If the head is prevented from moving after it is struck, then the impact injury is imparted
to the head as contact force. Inertial force occurs when the head is set in motion with or
without a contact force and results in acceleration of the head. The second mechanism
occurs through what is commonly referred to as the whiplash effect. The head is
propelled forward rapidly and then just as quickly propelled backwards. While the
cranium has not directly been struck, the brain within the cranial vault has suffered
trauma. The third mechanism is rare but can occur if a slowly moving object traps the
head against a rigid structure, slowly squeezing the skull to produce numerous
comminuted fractures of the cranial vault (Halliday, 1999).
Injury to the brain results if contact of inertial forces strains the tissue beyond its
structural tolerance. Strain is the amount of tissue deformation caused by an applied
mechanical force and is either compressive, tensile, or shear in nature. Compressive
strain results from tissue compression, and tensile strain is produced by stretch of the
tissue. Shear strain is the type of distortion that occurs when one tissue slides against
another. Brain and vascular tissue, being virtually incompressible, typically sustains
damage via tensile and shear strain (Halliday, 1999).
Pattern of Brain Injury
Traumatic brain injury may be divided into two primary patterns of injury, those
that produce focal injury and those that produce diffuse injury. Focal brain injuries occur
as the result of contact forces. Examples of contact force injuries are skull fractures,
epidural hematoma, coup contusion, and subdural hematoma. Contracoup contusion,
intracerebral hematoma, and subdural hematoma are examples of inertial forces
translationall acceleration) that produce focal and diffuse injury. Intertial force injuries
(rotational acceleration) are seen as concussion, diffuse axonal injury, subarachnoid
hemorrhage, tissue tear hemorrhage, and gliding contusion (Halliday, 1999).
Mortality and Morbidity
The mortality rate after traumatic brain injury of all severities is lower for children
as a group than for adults. However for children with severe traumatic brain injury the
mortality rate is as high as 59% in one series and does not include those children who
died at the scene of their injury. Other studies show a mortality rate of 29% for victims
of severe traumatic brain injury who are less than twenty years old and a case fatality rate
of 33% for those 15 years of age or younger. The factors that contribute most to the
outcomes of traumatic brain injury in children are the mechanisms of injury and the
severity of the primary brain damage (Jennet, Teasdale, & Braakman, 1979; Kraus, Fife,
& Conroy, 1987).
If one uses the Glasgow Coma Scale (GCS), children who suffer severe injuries
(GCS scores < 8) often have better outcomes than those predicted for adults with the
same score. The reported mortality from severe injury in children ranges from 9% to
53%, with the average at approximately 20%. This mortality is about half of that
reported for adults admitted with the same degree of injury by GCS (Ghajar & Hariri,
1992; Luerssen et al., 1988; Nakayama, Copes, & Sacco, 1991; Tepas et al., 1990).
Children in different age groups clearly have different mortality rates and
functional outcomes. Infants and preschool aged children appear to have the worst
functional outcomes. Although the immaturity of the brain may protect against loss of
more focal functions, it may result in a more significant decrease in overall cognitive
function. For equivalent injuries, often the residual neuropsychological defects can be
more profound in infants and young children than in adolescents and adults (Francel et
An accurate neurological assessment, performed close to the time of trauma, is vital
to formulating the plan of care. The most reliable marker for determining outcome in the
brain damaged individual, whether adult or child, is the level of consciousness after
resuscitation. The Glasgow Coma Scale, which will be discussed in detail later, was
developed to provide clinicians with a rapid, standardized, easy-to-use system to assess
patients with altered levels of consciousness. The use of the Glasgow Coma Scale in the
assessment of patients after traumatic brain injury has become universally accepted and
forms the foundation for communication among clinicians responsible for the care of
these patients (Ghajar & Hariri, 1992).
Observation of a patient's eyes, speech, and motor responses to stimuli provide the
care provider with a reliable estimate of level of neurologic function. The Glasgow
Coma Scale is based on a score between 3 and 15 and represents the numerical sum of
these three determinants. The cerebral metabolic rate of oxygen (CMRO2), which
reflects mainly supratentorial brain metabolism, is considered the most accurate
measurement of brain function and correlates closely to the Glowgow Coma Scale.
Based on this relationship, care providers can predict mortality in this population of
patients as a function of neurologic function and indirectly as a function of brain
metabolic activity. Since infants and toddlers are unable to speak or follow commands
appropriately, a children's coma score (CCS) was developed with a maximum score of 11
compared with 15 on the adult scale (Ghajar & Hariri, 1992).
Purpose of the Study
The purpose of this study is to investigate the relationship between cerebral
perfusion pressure (CPP) and outcome in severely head injured children. This study
seeks to determine if there is a definite threshold of CPP in the severely head injured
child, and if there is, what that threshold is. Is there a predictable relationship between
increased intracranial pressure (ICP) and CPP? Can this defined threshold be predictive
of outcome? The answers to these questions will help determine the therapeutic range of
CPP and ICP that allows for the best outcome in pediatric brain injured patients. This
will permit development of standard protocols of care for severely head injured children
and predict outcomes. The assumption of this study is that there does exist an ideal or
threshold value for CPP for traumatic brain injured children regardless of the value of
ICP and MAP.
Disturbances of cerebral circulation and metabolism play an important role in the
pathophysiology of severe head injury. Once inflicted, primary cerebral injury resulting
from TBI is immutable and irreversible. A major concept is that secondary insults add to
the primary damage inflicted by the trauma and impair neuronal healing. Secondary
systemic insults exert their deleterious effects by impairing the transport of oxygen and
nutrients. These secondary effects include raised intracranial pressure, arterial
hypotension or hypoxia, acute anemia, and high blood viscosity. These factors have an
adverse effect on cerebral blood flow and metabolism. These potentially viable cells
should be the targets of clinical strategies (Chestnut et al., 1993; Ghajar, 2000; Ghajar &
Hariri, 1992; Miller, Sweet, Narayan, & Becker, 1978; Robertson et al., 1999; Sullivan,
Significance of the Study
It is currently apparent that vigilant neurologic and cardiopulmonary monitoring
combined with aggressive management of postinjury intracranial hypertension, cerebral
perfusion pressure, uncontrolled seizure activity, hypoxia, and hypotension improves the
functional survival after head injury. The major goal in management of traumatic brain
injuries is the prevention of secondary brain damage by the avoidance of cerebral
ischemia and maintenance of cerebral perfusion (Hilton, 2000).
A four year effort under the aegis of the Joint Section on Neurotrauma and Critical
Care of the American Association of Neurological Surgeons and the Congress of
Neurological Surgeons, supported by the Brain Trauma Foundation, resulted in an
empirical, evidence-based analysis of the existing state of the literature for 14 topics
integral to traumatic brain injury treatment. The overall summary demonstrated that the
present state of traumatic brain injury management is poorly founded in terms of well-
performed, empirical, clinical outcome studies. With respect to the management of
pediatric traumatic brain injury, the situation is even less secure. The overall conclusion
of pediatric neurotraumatologists is that there are insufficient clinical data to make
definite statements on any aspects of managing pediatric traumatic brain injury (Chestnut,
Ghajar et al. (1995) show this inconsistency in a review of the management of
patients with severe head injury in the United States. This point is further brought home
by a similar study by Bulger et al. (2002). Thirty-four academic trauma centers show
pronounced variation in all aspects of care of adult head trauma patients. This
inconsistency of care is even more pronounced when management of children is
evaluated. Tilford et al. (2001) examined therapies and outcome for pediatric head
trauma patients in three pediatric intensive care units (PICU) showing marked variations
in both variables. Research on pediatric head trauma patients has lagged far behind that
of adults. As a result, large variations in practice styles and outcomes persist across
centers that care for children with head injuries. The numerous studies in adults have led
to evidence-based guidelines to support their management; whereas, therapies for
children have been extrapolated from adult studies.
Data acquired and analyzed from this study will help identify trends that facilitate
management of TBI in children that may be of prognostic value, both in assessing the
effect of therapy as well as predicting outcomes in terms of mortality and morbidity. The
potential financial benefits of this study are the development of a standard protocol
leading to improved outcomes which will decrease the cost of hospitalization through a
reduction in length of stay, decrease the length and cost of rehabilitative care and reduce
the emotional toll and expense to families.
The following hypotheses will be tested:
Children with non-penetrating traumatic brain injury with a cerebral perfusion pressure
averaging 50 mmHg or higher will demonstrate better outcome, as measured by the
DELTA disability and injury score (DS), than subjects with a cerebral perfusion
pressure averaging less than 50 mmHg.
Children with non-penetrating traumatic brain injury will tolerate a lower average
cerebral perfusion pressure than adults with non-penetrating traumatic brain injury as
measured by DS.
There is an inverse relationship between intracranial pressure and cerebral perfusion
pressure in children with traumatic brain injury.
Definition of Terms
Glasgow Coma Scale
One of the important concepts in evaluating neurologic injury is the Glasgow Coma
Scale (GCS). In 1974 Teasdale and Jennet proposed a clinical scale for assessing the
depth and duration of consciousness and coma. Three aspects of behavior are
independently measured: motor responsiveness, verbal performance, and eye opening.
The patient's best responses are recorded and a cumulative score is obtained. The lower
the total score, the more severe the neurologic compromise (Appendix B). The Glasgow
Coma Scale is modified for use with infants and small children with open fontanels
(Appendix C) (James & Trauner, 1985).
DELTA Disability and Injury Score
The Delta disability and injury score (DS) is a means of evaluating the qualitative
degree of disability and tracking rate and extent of injuries. It is based on a categorized
assessment (Appendix D) of four components that describe autonomous function. The
basic premise of this assessment scheme is that injury-related disability produces a
negative score that can be periodically reevaluated to document rate and extent of
recovery (Tepas, 1989). Disability, as assessed at discharge, is considered mild if the
DELTA score is -2 or less, moderate if -3 to -5 and severe for greater than -5 (Kisson,
Tepas, Peterson, Pieper & Gayle, 1996).
Glasgow Outcome Scale
The Glasgow Outcome Scale (GOS) is designed to evaluate patient outcome
following, among other things, traumatic brain injury. Similar to the Glasgow Coma
Scale, the Glasgow Outcome Scale assigns numerical values based on clinical condition
(Appendix E). The Glasgow Outcome Scale is appropriate for use from the time of
hospital discharge until several years post injury. As with the Glasgow Coma Scale, a
lower numerical total on the Glasgow Outcome Scale indicates a more severe outcome.
Correlation between initial GCS and GOS is low; however, GCS scores six hours after
presentation correlated better with eventual outcome (Waxman, Sundine, & Young,
Cerebral Blood Flow
Normal mean cerebral blood flow (CBF) in adults is 50ml per 100g brain tissue per
minute at an arterial PCO2 of 40mg. In children and adolescents the CBF value is higher
and probably inversely related to age, although normal values in the younger age groups
have yet to be established (Obrist et al., 1984).
Cerebral Blood Volume
Cerebral blood volume (CBF) is the flow of blood through the brain, important for
the delivery of oxygen and removal of waste products. In normal circumstances when
CBF falls the physiological electrical function of the cell begins to fail. Further, an
increase or decrease in cerebral arterial blood volume will cause an increase or decrease
in cerebral arterial blood volume because of arterial dilatation or constriction.
Cerebral Metabolic Rate of Oxygen
The cerebral metabolic rate of oxygen (CMRO2) is calculated by multiplying CBF
and the arteriovenous difference of oxygen content (AVDO2). Normal CMRO2 is 3.2
ml/100 g min (Kennedy & Sokoloff, 1957). When CBF falls, CMRO2 is initially
maintained because the brain extracts more oxygen from the blood. A lowering of
CMRO2 of greater than 50% of normal causes disintegration of functional elements of
the cell, eventually leading to neuronal death or infarction (Bouma & Muizelaar, 1992).
Arteriovenous Difference of Oxygen
The arteriovenous difference of oxygen content (AVDO2) is the amount of oxygen
gas (in ml gas/100 g min) extracted from the brain. Normal AVDO2 is 6.3 ml/100 ml
(Kennedy & Sokoloff, 1957).
Pressure-volume index (PVI) is a measure of brain compliance and is the calculated
volume (in milliliters) required to raise ICP by a factor of 10. It is thought to be a
reflection of the vascular component of the intracranial compartment.
REVIEW OF LITERATURE
Physiology of Cerebral Circulation and Metabolism
The brain, unlike other tissues, is enclosed in a rigid shell. This fact makes volume
regulation of the brain more sophisticated. When the ability of the cerebrospinal fluid
and intracerebral blood to adjust for volume is exhausted, the intracranial pressure will
rise. A rise in intracranial pressure reduces cerebral perfusion pressure with a
corresponding decrease in cerebral blood flow (CBF). Under normal circumstances, the
volume-regulating mechanisms of the brain work to keep the brain volume within normal
limits. One of the mechanisms controlling normal brain volume is the fluid exchange
across the brain capillaries. The brain differs from most organs of the body in its highly
sophisticated semipermeable capillary membrane function. The capillary is permeable
for water but less permeable for other molecules and solutes. The interplay among water,
solutes, and crystalloids effectively controls brain volume within allowable limits
(Grande, Asgeirsson, & Nordstrom, 1997).
Under normal circumstances, energy requirements of the brain are high. Most of
the energy is needed for restoration of ionic gradients across the cell membrane, for
maintaining membrane integrity, and for molecular transport between central and
peripheral sites. The oxidation of glucose is the single most important source of energy
for the brain. The oxidation of glucose is much more efficient than anaerobic metabolism
in the production of adenosine triphosphate (ATP). The end product of aerobic
metabolism of glucose is carbon dioxide, which easily crosses the blood brain barrier.
The end product of anaerobic metabolism is lactic acid, which accumulates and is toxic to
the brain. The main purpose of cerebral circulation is to provide sufficient glucose and
oxygen to the brain. Since the brain itself has no storage capacity for either of these
substrates, a tight coupling between CBF and cerebral metabolism is necessary for
maintaining normal neuronal activity. The supply of glucose is usually sufficient, thus
oxygen transport is the limiting factor in determining adequate cerebral circulation
(Bouma & Muizelaar, 1992).
Cerebral metabolism can be assessed by the Fick principle as the product of
cerebral blood flow and the arterio-venous differences (AVD) of metabolites consumed
and produced by the brain. Under normal circumstances, CBF is closely coupled to
cerebral metabolism, and changes in CBF are accompanied by reciprocal alterations in
oxygen and glucose extractions, as calculated by AVD of oxygen (AVDO2) and glucose
(AVD-Glu). In ischemic conditions, extractions of oxygen and glucose increase in order
to compensate for the low CBF. Following severe head injury, cerebral metabolic rates
of oxygen (CMRO2) and glucose (CMR-Glu) are reduced by as much as 50%. In these
patients, as opposed to ischemic patients, metabolic dysfunction presumably occurs as a
result of cellular injury, rather than by ischemia. If so, cellular demand for oxygen and
glucose by the injured brain is low, and therefore, lower uptake is expected (Cruz, Jaggi,
Shalmon, Caron, Martin, Hoyda, and Becker (1994) assessed the association
between oxygen and glucose utilization in conjunction with CBF in severely head-injured
patients. Also studied was the hypothesis that the cerebral supply of these metabolites is
sufficient for metabolic demand. Fifteen severely head injured patients were
prospectively studied. Glasgow Coma Score ranged from 3-7. Conclusions reached were
that traumatic brain injury alters the metabolic demands of neuronal cells and hence the
uptake of oxygen and glucose. Low uptake of oxygen and glucose results in small
AVDO2 and AVD-Glu regardless of their supply, and CMRO2 and CMR-Glu are often
coupled to CBF in severely head-injured patients.
The pathophysiology of ischemic brain edema depends primarily on the duration
and severity of ischemia. Cerebrovascular permeability as well as hydraulic conductivity
of capillaries, hydrostatic and osmotic pressure gradients, and tissue compliance and
resistance are also associated with the ischemic edema process. A hydrostatic pressure
gradient across the capillary develops soon after the ischemic onset and is the driving
force for early accumulation of edema fluid. Hatashita, Hoff, and Salamat (1989) sought
to clarify whether blood brain barrier and an osmotic gradient across the capillary are
associated with the development of ischemic brain edema. Using adult male Sprague-
Dawley rats that were subjected to occlusion of the middle cerebral artery (MCA), the
researchers demonstrated that the accumulation of edema fluid was related to a
hydrostatic pressure gradient. This developed soon after the onset of ischemia, and was
followed by an osmotic pressure gradient as ischemic injury progressed.
Intracranial pressure is an estimate of the force required to displace blood and CSF
from the intracranial space in order to accommodate the new volume. A thick layer of
bone penetrated by several foramina bound the intracranial space. The tentorium divides
it into two large compartments, and the compartments communicate through the tentorial
incisura. Since the thick bone of the calvarium is essentially nondistensible, the volume
of the intracranial space is virtually constant, irrespective of the pressure generated within
it. The tentorium itself can be displaced upward or downward but the total intracranial
volume remains unchanged. The intracrainial space is occupied by fluid and solid
material and these contents are nearly noncompressible. This noncompressability is the
basis for the Monro-Kellie doctrine (Rosner, 1993).
The Monro-Kellie doctrine put forth in 1783 the concept that the intracranial space
contains only two compartments that can change in volume, brain matter, and
intravascular blood. In 1846 Burrows added the concept of cerebrospinal fluid. He
maintained that the blood volume of the brain does change in volume under a variety of
circumstances, and is accompanied by a reciprocal change in volume in one of the other
intracranial components, either the brain, intravascular blood or brain tissue water, or the
cerebrospinal fluid (Lang & Chestnut, 1994; Langfitt, 1969; Shackford, 1997).
The largest component of the brain is the parenchyma. It represents approximately
1100 to 1200 grams and should be considered constant under most conditions. The
vascular component represents blood distributed between arteries, arterioles, capillaries,
venules, and the larger venous system. This total volume is approximately 150 ml but
varies widely. The cerebralspinal fluid compartment represents approximately 150 ml of
volume and has tremendous therapeutic potential since a portion is usually available for
removal. When combined, the total volume is approximately 1500 ml of which the
majority parenchymaa) is fixed and 20 percent (CSF, blood and water) are variable. This
concept continues to form the basis of cerebral intracranial pressure. One exception to
this concept is infants since the skull is not yet rigid (Rosner, 1993).
Many pathological conditions affect the cerebral parenchyma, though rarely in
absence of effect upon the vasculature or CSF components. Cerebral edema is the
prototypical process capable of increasing the parenchymal component of the intracranial
volume. Cerebral edema is an increase in the water content of the brain. Cerebral edema
is further separated into two types of brain water accumulation. These two pathological
types of cerebral edema are cytotoxic edema and vasogenic edema (Rosner, 1993).
Cytotoxic edema refers to the accumulation of primarily intracellular water.
Vasogenic edema represents ultrafiltrate of plasma leaking at a greater than normal rate
into the cerebral parenchyma. This is usually due to insults that affect primarily
vasculature at the capillary endothelial level and the integrity of the blood-brain barrier
Cytotoxic edema is primarily intracellular and will respond to therapies aimed at
cellular mechanisms for maintaining salt and water balance. This can be viewed as
metabolic processes requiring therapies aimed at cellular metabolism. Vasogenic edema
is more mechanical in nature and therapies aimed at vasogenic edema will be more
mechanically based and directed at the blood-brain barrier (Rosner, 1993).
Some of the first research generated describing the behavior of ventricular fluid
pressure in patients with intracranial hypertension was presented by Lundberg in 1960.
Lundberg first described "A"-wave or "plateau wave". This intracranial phenomenon is
characterized by spontaneous and acute elevations in ICP rapidly rising above baseline
levels (usually 15 to 25 mm Hg). These elevations may reach levels of 50 to 100 mm Hg
and last anywhere from 2 or 3 minutes to as long as 20 or 30 minutes. These ICP spikes
usually abort as quickly as they begin. Lundberg concluded these spontaneous A-waves
reflected vasodilation with subsequent increases in CBV (Rosner & Becker, 1984).
While CBV has been shown to increase during the plateau, CBF is slightly
decreased. Matsuda, Yoneda, Handa, and Gotoh (1979) studied five patients with
increased intracranial pressure and noted a marked decrease in cerebralvascular resistance
during the plateau waves. The researchers proposed that plateau waves indicate a period
of marked cerebral vasodilation followed by an increase in CBV. Rosner and Becker
(1984) in a study of cats with fluid percussion head injuries concluded that plateau waves
occur when there is intact autoregulation responding to changes in CPP.
Changes in ICP waveforms occur under various physiologic and pathophysiologic
conditions and may provide valuable information about intracranial adaptive capacity.
Intracanial waveform analysis provides information about intracranial dynamics that can
help identify individuals who have decreased adaptive capacity and are at risk for
increases in ICP and decreases in CPP, which may contribute to secondary brain injury
and have a negative impact on neurologic outcome (Kirkness, Mitchell, Burr, March &
Intracranial hypertension is defined as a persistent elevation of intracranial pressure
with an intracranial pressure over 20 mm Hg. During the expansion of intracranial space-
occupying lesions, such as subdural and epidural hematomas, the mechanisms for spatial
compensation fail. The brain becomes less compliant so that small increases in volume
cause increasing rises in intracranial pressure. As intracranial pressure increases, the
decreasing difference between arterial and intracranial pressure becomes equivalent to the
cerebral perfusion pressure. Leech and Miller (1974) looked at eight anaesthetized,
ventilated adult baboons. The intracranaial volume-pressure response was examined
during different levels of raised intracranial pressure during induced changes in systemic
arterial pressure and cerebral blood flow. At normal intracranial pressure, the volume-
pressure response was unchanged by alterations in systemic arterial pressure and cerebral
blood flow. At raised intracranial pressure the systemic arterial hypertension rendered
the intracranial contents more sensitive to the effects of an addition to the ventricular
volume as shown by an increased volume-pressure response. As intracranial pressure is
increased there is a linear correlation between the volume-pressure response and both
arterial pressure and cerebral blood flow. The clinical implication of this phenomenon is
that arterial hypertension in patients with increased intracranial pressure is likely to have
a deleterious effect by increasing brain tightness.
A retrospective study of 245 patients with TBI evaluated the contribution of CPP
and ICP to neurological deterioration. It was found for this study that the most powerful
predictors of neurological deterioration was the presence of intracranial hypertension
(ICP > 30 mm Hg). The CPP also had a prognostic power on neurological deterioration
when its level was less than 60 mm Hg (Feng, Huang, Gao, Tan, & Liao, 2000).
Robertson et al. (1999) compared two management protocols with long-term
outcome in patients with severe head injury. One protocol was targeted at intracranial
pressure management and the second was targeted at cerebral blood flow management.
One hundred eighty-nine adults admitted with severe head injury at a Level 1 trauma
center were studied prospectively in a randomized clinical trial. Mean arterial pressure,
cerebral perfusion pressure, and PaCO2 were seen as the primary differences between the
two protocols. The CBF-targeted management protocol was found more successful in
reducing secondary ischemic insults.
This CBF-targeted management protocol is known as the Lund therapy and aims to
reduce ICP. It is based purely on physiologic principles for cerebral tissue and blood
volume regulation in a tissue with disrupted blood-brain barrier and allows for both the
risks of increased ICP and the risks of compromised microcirculation. The Lund therapy
emphasizes reduction in microvascular pressures to minimize edema formation in the
brain. The goal of the therapy is to preserve a normal osmotic pressure, to reduce
capillary hydrostatic pressure by reducing systemic blood pressures, and to reduce CBV
by vasoconstricting precapillary resistance vessels (Eker, Asgeirsson, Grande, Schalen, &
Nordstrom, 1998; Grande, Asgeirsson, & Nordstom, 2002; Robertson, 2001).
A study to evaluate the Lund therapy by the developers of the protocol involved a
prospective, nonrandomized outcome study over a five year period on severely head
injured patients with increased ICP. Results were compared with a historical control
group with the same selection criteria for patients who were treated according to
conventional principles. The results showed the clinical outcome in the Lund therapy
group of patients considerably better compared with the outcomes of patients treated with
conventional, CPP, based therapies (Eker, Asgeirsson, Grande, Schalen, & Nordstrom,
Naredi et al. (2001) conducted a prospective non-random study of patients with
severe head trauma using the Lund therapy. One of the two purposes of the study was to
determine whether the good outcome obtained with the Lund therapy in two previous
outcome studies could be reproduced. The outcome results from the two previous studies
were successfully reproduced.
The effect of hypothermia on ICP, systemic and intracranial dynamics, and
metabolism in patients with severe traumatic head injury was studied to clarify optimal
temperature. Thirty-one patients were studied. Data results showed that ICP decreased
significantly at brain temperatures below 37 degrees Celsius and decreased more sharply
at temperatures 35 to 36 degrees Celsius. There was no difference observed at
temperatures below 35 degrees Celsius. The author concluded that decreasing body
temperature, after traumatic head injury, can reduce intracranial hypertension while
maintaining sufficient CPP without cardiac dysfunction or oxygen debt. Temperatures of
35 to 35.5 degrees Celsius seemed to be optimal (Tokutomi, Morimoto, Miyagi,
Yamaguchi, Ishikawa, & Shigemori, 2003).
Cerebral Perfusion Pressure
Cerebral perfusion pressure is the pressure gradient across the vasculature tree.
When systemic arterial pressure is normal and the ICP is normal, the difference between
CPP and systemic arterial pressure is minimal and probably unimportant under most
circumstances. These conditions fail to hold in the damaged brain and/or in the face of
hypertension. While ICP and mean systolic arterial blood pressure may vary
independently of one another, neither can be affected without altering CPP. Cerebral
perfusion pressure correlates with cerebral blood flow when either blood pressure or ICP
is altered (Rosner, 1995).
Cerebral ischemia dominates traumatic brain injury as the single most important
event-determining outcome. The primary role of CPP maintenance is the preservation of
cerebral blood flow through early and accurate monitoring. Cerebral perfusion pressure
is amenable to physician manipulation. Monitoring of cerebral perfusion pressure is
achieved by measuring intracranial pressure via subdural or intraventricular access
Bouma, Muizelaar, Choi, Newlon, & Young, 1991).
In 1986 Kontos and Wei studied the appearance of superoxide anion radicals in
cerebral extracellular space during and after experimental fluid-percussion brain injury.
Experiments were carried out on cats. The results supported the following conclusions:
1) Fluid-percussion injury causes the generation of superoxide which continues for at
least one hour after injury. 2) Superoxide generated by brain injury and/or radicals
derived from it are responsible for the sustained vasodilatation and reduced
responsiveness of cerebral arterioles to arterial hypocapnia.
Rosner and Coley (1986) examined CPP as a function of head elevation. Eighteen
patients with intracranial hypertension were strictly monitored with head elevations of 0
degrees to 50 degrees in 10 degree increments. The results indicated that for CPP to be
maximized, patients with increased intracranial pressure should remain flat in bed. There
was no case in which CPP improved with head elevation even though ICP was usually at
its highest point. Conversely, Feldman et al. (1992) and Gopinath, Robertson, Narayan,
and Grossman (1994) looked at the effect of head elevation on ICP and CPP. Both
studies concluded that head elevation of 30 degrees significantly reduced ICP without
significantly changing CPP.
Changaris et al. (1987) monitored cerebral perfusion pressure and Glasgow Coma
Scale scores to guide the management of 136 consecutive patients admitted to a trauma
service. Cerebral perfusion pressure was chosen as the single determinant to correlate
with GCS because it represents the net effects of mean arterial pressure (MAP) and ICP.
The authors found that cerebral perfusion pressure has the greatest impact on potential
A study of 50 patients with severe head injury (GCS of 8 or lower) was undertaken
by Shigemori et al. (1989) to ascertain thresholds of ICP and CPP for CBF and brain
function. Extra- and intracranial blood velocities, determined by transcranial
ultrasonography, were found decreased when ICP increased to 20-30 mm Hg and when
CPP decreased to 40-50 mm Hg.
McGraw (1989) looked at the issue of determining the most beneficial CPP. A
review of data collected over a nine-year period at the University of Louisville Trauma
Center showed that a CCP of 80 mm Hg is a critical point at which mortality and
morbidity change. Graphing GOS versus CPP derived this figure. A look at mortality
showed a semi-linear relationship between GOS and GCS. There appeared to be a
natural break in the data at a CPP of 80 mm Hg. Outcomes were observed to be more
favorable when the CPP was kept above 80 mmHg. There was a decrease in percent
mortality, and increase of good outcomes and a decrease in the number of GCS
deteriorations. When CPP remained below 60 mmHg for six hours, death could be
predicted with the highest accuracy. This retrospective study further determined that
patients could withstand acute drops in CPP to levels below 60 mm Hg, however, a
sustained drop in CPP for six hours is less tolerable. The critical time is five hours or
under for depressions in CPP below 60 mm Hg.
Rosner and Daughton (1990) used CPP as a method of management for ICP.
Thirty-four consecutive patients with GCS < 7 were clinically managed to an CPP of 84
+ 11 mm Hg, ICP of 23 + 9.8 mm Hg, and SABP of 106 + 11 mm Hg. All patients were
nursed with the head of the bed at 0 degrees. The results indicated that CPP can be
artificially elevated by clinical manipulation without deleterious ICP or systemic effects.
Marmarou et al. (1994) studied 386 severely brain injured patients (GCS < 8)
obtained from the combined data banks of the Traumatic Coma Data Bank (TCDB) and
the Medical College of Virginia Neurocore. The objective was to determine which
factor, raised ICP or hypotension, was most responsible for the reductions in CPP during
the course of therapy. The amount of time CPP was less than 60 mm Hg during the first
72 hours of monitoring as a function of GCS and GOS was determined. The data
obtained suggests that for a CPP threshold of 60 mm Hg, arterial pressure should not be
allowed to fall below a mean of 80 mm Hg that corresponds to a systolic level of 114 mm
Rosner and Rosner (1994) looked at 158 patients (GCS < 7) with traumatic brain
injury. These patients were managed using volume expansion, CSF drainage via
ventriculostomy, systemic vasopressors, and normocapnia to maintain a minimal CPP >
70 mm Hg. This data led to the conclusion that CPP management yields favorable
clinical results up to six-fold better than traditional techniques directed at ICP. When
causes of mortality were examined, CPP management alone reduced death rates by 33-
Conversely, a 1994 study by Shalmon et al. purposed to correlate simultaneous
measurements of CPP and CBF in severely head-injured patients and to assess the ability
of predicting CBF from CPP. Fifty-two consecutive head-injured patients were
prospectively evaluated for this study. The data indicated that CPP and CBF are poorly
correlated in head injured patients. The maintenance of MABP above 90 mm Hg and
preventing intracranial hypertension above 20 mm Hg does not insure adequate CBF in
severely head injured patients.
The use of CPP management as the primary goal of therapy was thought to yield a
lower mortality than with the more traditional ICP-based techniques. Rosner, Rosner,
and Johnson (1995) studied 158 patients with GCS scores of 7 or lower. The patients
were clinically managed to maintain CPP at 70 mm Hg or higher. The mortality results
obtained were significantly better than other reported series across GCS. The researchers
surmised that cerebral perfusion pressure management can serve as the primary goal in
the treatment of TBI with substantially improved mortality and morbidity.
Cruz, Jaggi, and Hoffstad (1995) emphasized exploring the relationship of cerebral
perfusion pressure and vascular resistance to cerebral blood flow, as well as to cerebral
oxygen metabolism. These researchers felt that cerebral vascular resistance, more than
cerebral perfusion pressure, determined important cerebral hemodynamics. To that end,
66 adults with severe acute brain trauma were prospectively evaluated. All patients had
GCS of between 4 and 8, coma for at least 12 hours and intracranial pressure monitoring.
Cerebral vascular resistance was calculated as equal to cerebral perfusion pressure
divided by cerebral blood flow. Data analysis did not demonstrate any correlation
between cerebral perfusion pressure and cerebral blood flow, between cerebral perfusion
pressure and arterio-jugular oxygen content difference, and between cerebral perfusion
pressure and cerebral metabolic rate of oxygen consumption, over a broad range of
perfusion pressures ranging from 60 130 mm Hg. In contrast, a significant correlation
was found between cerebral vascular resistance and cerebral blood flow, where higher
values of cerebral vascular resistance were associated with lower blood flow levels, and
vice versa. Thus cerebral vascular resistance (not perfusion pressure) was more closely
correlated with different patterns of cerebral blood flow and metabolism.
Giulioni and Urino (1996) used a mathematical model to study the possible effect
of CPP changes on ICP. The model mimicked intracranial hemodynamics and CSF
dynamics. Specifically, the study aimed to clarify how a sudden CPP decrease, caused
by arterial hypotension, can affect CBV, ICP, and CBF. The model showed the relevant
role of CPP changes elicited by acute arterial hypotension on intracranial dynamics. To
achieve intracranial stability, CPP should be maintained above 80 to 90 mm Hg.
The goal of fluid resuscitation in a patient with TBI is restoration of CPP.
Shackford (1997) used a porcine model to determine if a hypotonic saline solution or
Ringers Lactate provided the best protection against secondary brain ischemia. The ideal
fluid must be effective in restoring CBF and CPP with small volumes and have little
effect on either ICP or cortical water content in the uninjured areas of the brain.
Shackford concluded cerebral perfusion pressure was elevated and CBF increased after
brain injury following infusion of hypertonic saline thus decreasing or limiting secondary
Elevated CPP and decreased ICP is seen in a similar study by Simma, Burger, Falk,
Sacher, and Fanconi (1998). Hypertonic saline verus Ringer's Lactate was studied in a
randomized prospective study of 35 consecutive children with head injury (GCS < 8).
All children received continuous infusions of 3% saline to raise serum osmalarity to the
level required to reduce intracranial pressure to < 20 mm Hg and < 15 mm Hg in patients
with an open fontanel. Cerebral perfusion pressure was maintained at 60-70 mm Hg for
older children and 50 mm Hg in infants. The children receiving hypertonic saline
showed decreased ICP and significantly improved CPP. An evaluation of outcome
showed hypertonic saline to be efficacious and safe.
The Traumatic Coma Data Bank (TCDB) was evaluated by Chestnut (1997a) to
delineate the influence of hypotensive episodes occurring during the early (time of injury
through resuscitation) and the late (in intensive care units) posttraumatic periods. The
analysis consisted of 493 patients who survived nine or more hours into their ICU course.
It was determined that the occurrence of hypotension during either the early or the late
period was statistically significant, and an independent predictor of outcome. Chestnut
determined the estimated reduction in unfavorable outcome (death or vegetative state) for
each of these occurrences independently and found that the elimination of early
hypotension was predictive of a 15-fold reduction and elimination of late hypotension
was predictive of an 11-fold reduction in relative risk.
Juul et al. (2000) examined the relationship and relative importance of ICP and
CPP in severely head injured (GCS 4 to 8) patients. The study of 407 patients was
prospective and involved more than 50 treatment centers in Europe, Australia, Canada,
and Argentina. Evaluation of patient outcome using the GOS failed to demonstrate any
significant benefit of a CPP greater than 60 mm Hg. Conversely an ICP greater than 20
mm Hg is associated with increased mortality rates compared to ICPs of less than 20 mm
Hg. The aim of increasing CPP to levels greater than 70 80 mm Hg in patients with
severe head injury, while not attempting to reduce the ICP if it is greater than 20 mm Hg,
was found unacceptable.
The retrospective analysis of prospectively collected data on 114 head-injured
patients between January 1997 and August 2000 evaluated optimal CPP. Mean arterial
pressure, ICP, CPP were continuously recorded and pressure reactivity index (PRx) was
calculated. Pressure reactivity index is defined as the ability of vascular smooth muscle
to respond to changes in transmural pressure, one of the key mechanisms responsible for
autoregulation in CBF. The CPP-oriented treatment strategy was used with a target CPP
of at least 70 mm Hg. Patient outcome was assessed using the GOS six months after
discharge. The researchers determined that the correlation between PRx and outcome
was significant, a higher PRx the less favorable the outcome. Pressure reactivity index
reacted dynamically to changes in CPP with breakpoints for a decline in cerebrovascular
pressure reactivity at approximately 60 and 85 mm Hg. It was concluded that patients
with a mean CPP close to optimal were more likely to have a favorable outcome than
those whose mean CPP was more different than optimal CPP, optimal CPP individually
defined for each patient (Steiner et al., 2002).
In a prospective study, 124 adult head injured patients were studied during their
stay in intensive care unit. A continuous computerized data collection system recording
minute by minute values for physiological variables was initiated. Monitored variables
included heart rate, systolic blood pressure, mean arterial pressure, intracranial pressure,
oxygen saturation, and peripheral and core temperatures. Cerebral perfusion pressure
was calculated on-line. Patient outcome was evaluated using the Glasgow Outcome
Scale score. Data obtained confirmed hypotension and papillary response as major
predictors of poor outcome. Therefore management of BP was found to be important in
the treatment of patients after head injury. Cerebral perfusion insults were found to be
more important than ICP insults. Hypotension and a low CPP were indicated as the best
predictors of death. Hypotension was a significant predictor of poor outcome. Low CPP,
patient age, hypocarbia, and papillary response were good predictors of good/poor
outcome. This study validates the premise that secondary insults play a significant role in
determining patient outcome and occur commonly with current clinical management
(Andrews et al., 2002)
The use of a CPP driven protocol was challenged by Oertel, Kelly, Lee, Glenn,
Vespa, and Martin (2002). This study tested the hypothesis that increasing MAP
decreases ICP after TBI. A total of 23 patients were continuously monitored for MAP,
ICP and jugular-venous oxygen saturation. The data indicated that in the majority of
cases ICP increased due to an increase in MAP. The authors suggested that the
vasoconstrictory cascade is not functioning in these patients. The authors proposed that
CPP therapy has an indication in patients with a high GCS and low jugular-venous
oxygen saturation. They further concluded that in the majority of subjects ICP increased
with increasing MAP and did not recommend CPP therapy as a general concept for
treating increased ICP.
A retrospective analysis of critical thresholds for ICP, MAP, CPP, and fluid
balance was associated with outcome after TBI. A total of 392 adult patients were
studied. The control group consisted of 193 randomly assigned patients. The study group
received standard treatment plus hypothermia for 48 hours. Intracranial pressure
measurements of 20, 25, and 30 mm Hg; MAP of 70 and 80 mm Hg; CPP levels of 50,
60, and 70 mm Hg; and fluid balance levels in quartiles were examined for their effect on
outcome as measured by the GOS six months after injury. Glasgow Coma Scale score at
admission, age, MAP < 70 mm Hg, fluid balance lower than -594 mL, and ICP > 25 mm
Hg, in that order, were the most powerful variables in determining outcome. The authors
recommended maintaining CPP of> 60 mm Hg. They also maintained that driving CPP
levels > 80 mm Hg risks increasing the possibility of medical complications (Clifton,
Miller, Choi, & Levin, 2002).
The driving pressure gradient for CPP is the difference between MAP and critical
closing pressure (CPP zero flow pressure). Therefore, determination of the difference
between MAP and CCP should provide an appropriate monitoring of effective CPP.
Based on this concept, the authors compared conventional measurements of CPP by
MAP and ICP with effective CPP, measured by blood flow velocities of the middle
cerebral artery. Seventy consecutive head trauma patients who received invasive ICP
pressure monitoring were included in the study. It was determined that the indirect
measurement of CPP (MAP ICP) is an equally, but less invasive method, of measuring
CPP (Thees et al., 2002).
In another study Thees et al. (2003) investigated whether CPP is a reliable
parameter of sufficient cerebral perfusion and oxygenation. In an animal model of
controlled ICP, the effect of decreasing CPP, due to increasing ICP, on cerebral tissue
oxygenation was studied. Reduced CPP due to increased ICP led to a continuous
decrease in cerebral oxygen saturation of hemoglobin and a decreased CPP in all animals.
The experimental findings further suggested that CPP thresholds may be misleading. In
some animals a CPP of 45 mm Hg was sufficient to provide adequate cerebral tissue
oxygenation while in other animals a cerebral perfusion pressure of 42 mm Hg was
associated with a severe reduction in cerebral blood flow. Furthermore, after severe brain
injury CPP shows a poor correlation with CBF. The authors concluded that perhaps an
individual determination of CPP thresholds with respect to cerebral oxygenation and
cerebral function is required for therapeutic management of intracranial hypertension.
Cerebral autoregulation is the ability to maintain a constant cerebral blood flow
despite a changing cerebral perfusion pressure. Cerebral perfusion pressure is the
stimulus to cerebral autoregulation, not the absolute systemic arterial pressure. Cerebral
vasculature maintains cerebral blood flow relatively constant across a wide range of
pressure gradients through vasodilation and vasoconstriction. Within the vascular
compartment, CBF remains relatively constant despite changes in CPP over the 50- to
170-mm Hg range (Lang & Chestnut, 1994). Autoregulation of CPP is accomplished by
changes in the cerebrovascular resistance achieved by alterations in the caliber e.g.,
constriction when the pressure rises and dilation when the pressure decreases, of the
vessels. The consequence of constant flow and changing vessel size is a change in the
CBV (Gray & Rosner, 1987). It is well established that CBF, over a wide range of
arterial blood pressure, changes proportionally less than associated changes in pressures
(Harper, 1966; Lassen, 1959).
Vasconstriction and vasodilatation occur by changes primarily in the cerebral
arteries, primarily the arterioles. The varying diameter changes cerebral vascular
resistance. A reduction in systemic pressure leads to a reduction in CCP. Cerebral
vascular resistance is reduced in a manner that compensates for the reduction in pressure
gradient and CBF is maintained (Giulioni & Ursino, 1996; Rosner, 1995).
Under normal conditions, most vasoconstriction occurrs by the time the CPP is
110-120 mmHg. At this level of CPP, cerebral vascular resistance is nearly maximal,
with little vasoconstriction left to be attained. As CPP is reduced, the vasculature begins
to dilate but only slowly. A study by Kontos et al. (1978) on anesthetized cats studied the
response of cerebral precapillary vessels to changes in arterial blood pressure. Vessel
responses were found to be size dependent. Between mean arterial pressures of 110 and
160 mm Hg autoregulatory adjustments in caliber occurred only in vessels larger than
200 microns in diameter. Small arterioles, less than 100 microns in diameter, dilated
only at pressures equal to or less than 90 mm Hg; below 70 mm Hg their dilation
exceeded that of the larger vessels. When pressure rose to 170-200 mm Hg, small vessels
dilated while the larger ones remained constricted. At very high pressures (greater than
200 mm Hg), forced dilation was frequently irreversible. Measurement of the pressure
differences across various segments of the cerebral vascularture showed that the larger
surface cerebral vessels (over 200 microns) were primarily responsible for the
adjustments in flow over most of the pressure range. Kontos et al. further determined in
animal studies that the amount of vasodilatation that occurred between 80-100 mm Hg
CPP is on the order of 10% to 15% of the total range. The remaining 80% to 85% of
vasodilatation occurs from a CPP of 70 to 80 mm Hg to 50 to 60 mm Hg. The
researchers concluded that the rate of radius change is logarithmic and not linear within
the autoregulatory range.
Below the lower limits of autoregulation of cerebral autoregulation, the vasculature
is passive. Vessels, once maximally dilated, cannot dilate further and must collapse as
the pressure gradient within them is further reduced. Once the CPP drops below the
lower limits of autoregulation, vessels collapse and blood flow declines rapidly
(Weinstein & Langfitt, 1967).
The active range of cerebral autoregulation is 50 to 160 mm Hg CPP. Within this
range blood flow is relatively constant across a wide range of CPP, and the vasculature
constricts as CPP is increased. Vasculature constriction causes a slight increase in blood
flow but a decrease in blood volume. Therefore as CPP increases, vasculature constricts
and ICP decreases. Thus within the active range of cerebral autoregulation, ICP is
expected to vary indirectly with CPP (DeWitt et al., 1986; Kontos et al., 1978).
At the upper limits of cerebral autoregulation, the vasculature is maximally or
nearly maximally constricted. Any changes in CPP near the upper limits or slightly
above the upper limits of cerebral autoregulation results in a net increase in cerebral
blood flow but little change in ICP. If the pressure increases dramatically above this
level, the vasculature passively dilates, blood flow increases markedly, and ICP
increases. While this phenomena is often seen in cerebral pathology such as cerebral
encephalopathy it rarely, if ever, occurs after cerebral trauma (Kontos et al., 1978).
This series of events assumes that cerebral metabolic rate and metabolism are
constant. This is not always the case as a reduction in CPP may be accompanied by
reductions in level of consciousness and reductions in metabolism. These reductions will
shift the blood flow curve to lower levels, similar to the Starling curve of the heart being
shifted upward or downward by sympathetic influences (Rosner, 1995).
When there is a cerebral insult, such as occurs by mechanical trauma, the ability of
the vasculature to respond is altered. Lewelt, Jenkins, and Miller (1980) tested the
hypothesis that concussive brain injury impairs autoregulation of CBV by studying 24
cats subjected to hemorrhagic hypotension. The authors observed an increase in cerebral
vascular resistance. The cerebral autoregulatory curve is depressed and shifted to the
right. This phenomenon of cerebral injury occurred as a result of severe trauma although
it was also observed in chronic hypertensive disease. The authors concluded the primary
result is the need for supranormal levels of CPP to achieve and maintain normal levels of
CBV. Thus, because cerebral vascular resistance was increased, the pressure gradient
across the vasculature must be increased to achieve relatively normal CBV.
Cerebral injury alters the critical closing pressure of the cerebral vasculature. The
cerebral critical closing pressure is the CPP pressure at which blood flow begins. This
CPP is usually between 0 and 5 to 10 mm Hg. Cerebral injury increases the critical
closing pressure of the cerebral vasculature and may raise it to 30 to 35 mm Hg or more.
This increase further shifts the cerebral autoregulatory curve to the right (Nelson et al.,
Bauer et al., (1999) developed a model using piglets to evaluate the effect of
cerebral volume expansion on CPP. Mild CPP reduction (30% of baseline values) was
completely compensated for by blood flow and oxygen update of the cerebrum. This
finding correlated with the autoregulatory threshold in piglets. Additional CPP reduction,
which surpassed the autoregulatory threshold, resulted in a decreased blood flow to the
cerebrum with a reduction in cerebral oxygen delivery and cerebral oxygen uptake.
The use of catecholamine infusion is commonly used to increase MAP in order to
preserve blood flow when increased ICP compromises CPP. This CPP protocol is based
on the premise that a high CPP can help stabilize ICP and prevent the vasodilatory
cascade when the CPP is close to the lower limits of the autoregulation plateau. A
prospective study of 42 patients examined the influence of autoregulation on the
amplitude and direction of changes in intracranial pressure in patients with severe head
injuries during the management of CPP. Continuous recordings of CBF velocity, ICP,
and MAP during the start or change of continuous norepinephrine infusion concluded that
cerebral perfusion pressure-oriented therapy can be a safe way to reduce ICP, whatever
the status of autoregulation (Ter Minassian et al., 2002).
In a prospective study involving 122 patients with severe head trauma Hilatky et al.
(2002) evaluated the extent and timing of impairment of cerebral pressure autoregulation.
An autoregulatory index (ARI) was used as the independent variable. A normal ARI
value is 5 + 1.1. The changes in ARI value over time were examined against other
physiological values. The authors concluded that inability of cerebral vessels to
autoregulate CBF may play a role in the vulnerability of the injured brain to secondary
insults. Further, the results indicated that this vulnerability increased beyond the first 24
Lang, Czosnyka, and Mehdorn (2003) studied the relationship between arterial
blood pressure, ICP, directly measured brain tissue oxygenation, and middle cerebral
artery blood flow veloscity in severely head injured adults. This prospective study
involved a total of 14 patients to whom blood pressure was pharmacologically
(norepinehrine) manipulated to achieve CPP ranging from 50 to 100 mm Hg. The
authors demonstrated that cerebral autoregulation and tissue oxygen reactivity were
mutually correlated. The better the cerebral autoregulation was preserved, the smaller
were the brain tissue oxygenation changes when CPP changed. There was a close link
between cerebral blood flow and oxygenation.
Poiseuille's Law relates flow through a tube directly to the pressure gradient across
the tube, the radius of the tube to the fourth power, and inversely to the viscosity of the
For the above equation Q, is the rate of blood flow, AP is the pressure difference
between the ends of the vessel, r is the radius of the vessel, 1 is length of the vessel, and r
is viscosity of the blood. Since the rate of blood flow is directly proportional to the
fourth power of the radius of the vessel, this demonstrates that the diameter of a blood
vessel (which is equal to twice the radius) plays the greatest role of all factors in
determining the rate of blood flow through a vessel (Guyton & Hall, 2000).
Adjusting the equation for cerebral perfusion pressure, we derive the following
Flow = t(CPP)r4
It can be seen that only the vascular radius is altered during cerebral autoregulation.
Under conditions of high CPP, the radius of the vessels is maximally small with
relatively little additional vasoconstriction available. Under these conditions, the radius
remains constant. Therefore, when CPP is high, any additional increase in perfusion
pressure must increase blood flow. Any decrease in viscosity must also increase blood
When CPP is 70 to 80, the vasculature is believed to be in the midrange of its
diameter. The cerebral vascular is, therefore, capable of vasoconstriction and
vasodilation. Changes in CPP around this normal range result in very little change in
CBF. There is, however, significant change in the radius of the vasculature. The change
in vascular radius size results in change in CBV, which manifest in ICP changes. A
smaller radius translates into a reduction in CBV and a reduction in ICP. Below the
lower limits of autoregulation, any increase in CPP is associated with not only an
increase in blood flow, but also expansion of the vasculature, tending to increase ICP
Traumatic brain injury increases the resistance of the cerebral vasculature. This
increased resistance is through alterations in the vascular endothelium. When the injury
is severe, the level of CPP required to move into the active range of the cerebral
vasculature is higher and probably narrower. The CPP must be higher and the range in
which ICP varies directly with CPP increases (El-Adawy & Rosner, 1988).
Brain compliance is the ability of the cranial contents to accommodate an extra
amount of volume without an increase in ICP. This additional volume may be in the
form of a hematoma, cerebral spinal fluid, or increased cerebral blood volume. Changes
in blood pressure or CPP have a major influence upon the intracranial blood vessels,
particularly the arterioles when autoregulation is intact. Outside the range of
autoregulation, brain compliance or pressure-volume index (PVI) become inversely
related to changes in CPP (Bouma, Muizelaar, Bandoh, & Marmarou, 1992). Muizelaar,
Marmarou, and Ward (1989) obtained measurements of ICP and PVI in comatose
patients with severe head injuries ranging in age from 4 to 50 years. The authors
determined that within the range of cerebral autoregulation PVI is not dependent on CPP
changes. When cerebral autoregulation is not intact, the higher pressure in the arterial
system is transmitted to the arterioles, capillaries, and veins. This leads to increased
transmural pressure transfer into brain parenchyma. This effect is magnified by passive
dilation of these vessels. This makes the walls thinner, thereby, decreasing the barrier to
Bishop, Bishop, and Rosner (1994) tested the hypothesis that the volume tolerance
of the brain as measured by PVI would decrease as the head was elevated from zero to 45
degrees. This hypothesis was based on the concept of a linear relationship of PVI and
CPP above the lower limits of autoregulation. Pressure volume index was measured in
15 adult neurosurgical intensive care patients at head elevations of zero, 15, 30, and 45
degrees. Pressure volume index decreased in all patients with increasing head elevation.
Intracranial pressure and CPP both decreased with increasing head elevation. The
elevated head position decreased ICP by increasing venous drainage with a secondary
increase in CBV. This study suggested that CBV may actually be increased with head
elevation. A flat position was determined to be superior to a raised head position in the
management of head injured patients. Improved outcome in patients treated in the flat
position suggested that ICP was less important than CPP in the brain injured patient.
Mannitol and Fluid Balance
Poiseuille effects are extremely important when understanding the effect of
mannitol and other osmotic agents. Administration of mannitol increases cerebral blood
flow when CPP is relatively high, and it also lowers ICP and helps maintain CBF when
CPP is in the midrange. Mannitol does not withdraw fluid from the edematous brain. It
does not act through cerebral dehydration. Kaufman and Cordoso (1992) studied cerebral
edema in cats and observed the administration of mannitol (0.33 gm/kg q4h) did not
reduce brain water.
Mannitol potentates blood flow and oxygen delivery through Poiseuille effects
rather than through cerebral dehydration in the adequately hydrated patient. Mannitol
exhibits its effect by decreasing the viscosity of blood. The viscosity of blood increases
as the hematocrit increases. The viscosity of blood at a normal hematocrit of 40 is about
three. This means that three times as much pressure is required to force blood through
vessels as would be necessary for water (Guyton & Hall, 2000). Research on patients
undergoing surgery for intracranial aneurysms by Burke, Quest, Shien, and Cerri (1981)
studied the effect of mannitol on cerebral edema. Mannitol reduced erythrocyte rigidity
and decreased whole-blood viscosity, thereby, enhancing tissue perfusion in the cerebral
Muizelaar, Lutz, and Becker (1984) described changes in ICP and CBF in a group
of severely head-injured patients with intact and defective autoregulation after mannitol
administration. For patients with intact autoregulation the decrease in resistance to flow
from decreased blood viscosity and increased resistance from vasoconstriction balanced
so that CBF remained the same. The authors termed this blood viscosity autoregulation
of CBF, analogous to pressure autoregulation. When autoregulation was not intact, there
was no vasoconstriction, CBF increased with decreased viscosity.
The effect of fluid resuscitation in pediatric head injury was evaluated in a
prospective, clinical study using Ringer's solution (sodium 131 mmol/L, 277 mOsm/L)
compared with hypertonic saline (sodium 268 mmol/L, 598 mOsm/L). A total of 35
consecutive children were studied. The findings showed that treatment of severe head
injury with hypertonic saline is superior to treatment with lactated Ringer's solution. An
increase in serum sodium concentrations significantly correlated with lower ICP and
higher CPP. Children treated with hypertonic saline required fewer interventions, had
fewer complications, and stayed a shorter time in the ICU (Simma et al., 1998).
A similar study by Peterson, Khanna, Fisher, and Marshall (2000) showed that
administration of hypertonic saline solution to children with closed head injury appeared
to be a promising therapy for control of cerebral edema. A retrospective chart review of
68 children with closed head injury used 3% hypertonic saline to increase serum sodium
to levels necessary to reduce ICP < 20 mm Hg.
In a follow-up, Khanna et al. (2000) conducted a prospective study to evaluate the
effect of prolonged infusion of 3% hypertonic saline (514 mEq/L) and sustained
hypernatremia on refractory intracranial hypertension in pediatric traumatic brain injury
patients. Ten patients with increased intracranial pressure resistant to conventional
therapy (head elevation at 30 degrees, normothermia, sedation, paralysis and analgesia,
osmolar therapy with mannitol, loop diuretic, external ventricular drainage in five
patients) showed decreases in ICP and increases in CPP with increases in serum sodium.
It is generally well accepted that hyperventilation reduces ICP by reducing cerebral
blood volume through constriction of the pial and cerebral arterioles. As C02 readily
crosses the blood brain barrier, a decreased PaCO2 is immediately reflected in a reduced
PCO2 in the interstitial brain fluid. This then leads to a reduction in hydrogen ion (H+)
concentration in the vicinity of the cerebral blood vessels. Because the H+ ion is one of
the most potent relaxants of smooth muscles of cerebral arterioles, a reduction in its
concentration will lead to rapid vasoconstriction. Vasoconstriction is not dependent on
low C02 and can be maintained only if the increased perivascular pH can be maintained.
Studies have shown that pH in blood and cerebrospinal fluid returns to normal during
prolonged hyperventilation, despite sustained hypocapnia (Christensen, 1974).
Muizelaar and van der Poel (1989) studied New Zealand white rabbits to examine
whether the return to normal pH during prolonged hyperventilation would be
accompanied by vasorelaxation. Pial arteriolar vasoconstriction was maintained with
prolonged hyperventilation. They concluded that hyperventilation is effective in
reducing cerebral blood volume for less than 24 hours and that it should be used only
during actual ICP elevations.
Cruz demonstrated improved outcome associated with TBI management based on
optimized hyperventilation. A 1995 study using hyperventilation as a primary therapy
showed improved outcomes secondary to improved glucose uptake. A 1998 study design
compared hyperventilation therapy to CPP therapy, in a group of TBI patients. The
hyperventilation therapy group showed a 12% mortality. For the patients treated with the
CPP therapy the mortality was 32%.
In an editorial to the 1998 Cruz study, Chestnut (1998) presented the study as
another "salvo in the vociferous and occasionally vitriolic interchange between two
preeminent groups in head injury management: the CPP camp and the optimized
hyperventilation camp" (p.210). Chestnut calls for the elimination of the continuing
polarization of management strategies. He believes the future of head injury
management lies in targeting therapy to the underlying brain pathophysiology based on
injury type and evolution over time. Further Chestnut sees that only by considering
therapeutic modalities as complimentary will there be a successful transition to a targeted
Physiology of Pediatric Cerebral Circulation and Metabolism
As discussed, cerebral vasculature is regulated by metabolism (metabolic
autoregulation), CPP (pressure autoregulation), blood viscosity (viscosity autoregulation),
and PaC02 (C02 reactivity). Cerebrovascular reactivity and regulation of CBF are
disturbed after severe traumatic brain injury (Bourma & Muizelaar, 1992). These
principles also form the template for the management of pediatric TBI.
In 1957, Kennedy and Sokoloff studied cerebral circulation in children using the
nitrous oxide method first described by Kety and Schmidt in 1948. Modifications to the
procedure included a reduction in the amount of blood drawn and achieving active
cooperation with the child. The results were some of the first to systematically quantify
the physiological differences between children and adults. The comparison of the results
found notable differences in pulse rate, blood pressure, hemoglobin concentration, blood
gas concentrations, and blood pH. The mean cerebral blood flow in children, 106
milliliters per 100 grams per minute, was found to be considerably greater than the mean
value, 60 milliliters per 100 grams per minute, observed in the young adults. Also
significantly higher in children, was the cerebral oxygen consumption, the mean value
being 5.2 milliliters per 100 grams per minute as compared to 4.2 milliliters per 100
grams per minute in the adults. The cerebral vascular resistance in the children was 0.8
mm Hg per 100 gram per minute, which is lower than the 1.4 mm Hg per 100 gram per
minute observed in adults. No difference in the cerebral respiratory quotient between the
two groups was observed.
Pediatric Traumatic Brain Injury
Pediatric head trauma is common and is the leading cause of death and disability in
childhood. Children are more likely to suffer increased intracranial pressure and diffuse
cerebral injury than adults, who tend to develop focal intracranial lesions (Graneto &
Soglin, 1993). Multiple studies have put forth the proposal that there is a relationship of
patients' age to mortality from head injury. This relationship applies to all levels of
A series of 8,814 head-injured patients admitted to 41 hospitals was prospectively
studied. Of the total patients studied 1,906 (21.6%) were 14 years of age or less. Except
for patients experiencing profound hypotension or subdural hematoma, the pediatric
patients exhibited a significantly lower mortality rate compared to the adults (Luerssen et
The Glasgow Coma Score is the accepted method of evaluating patients suffering
from traumatic brain injury. Lieh-Lai et al. (1992) observed a significant portion of
children with severe TBI and relatively low GCS scores who seemed to have a better
functional outcome than expected. They conducted a study to determine the validity of
the GCS score alone in predicting the outcome of severe TBI in children. Seventy-nine
children were studied retrospectively. The researchers concluded that a low GCS score is
not a sole predictor of poor outcome in children with TBI.
Pigula, Wald, Shackford, and Vane (1993) hypothesized that hypertension and/or
hypoxia occurring in children suffering from traumatic head injury (GCS < 8) would
have significantly higher mortality than normotensive children with normal blood gases.
Over a five year period, 58 children were prospectively evaluated on the basis of systolic
blood pressure and arterial blood gases. Children in the normotensive and normal blood
gas group had a significantly improved survival. These results were validated by a
retrospective review of 509 children from the National Pediatric Trauma Registry.
A study by Mendelow et al. (1994) identified children hospitalized for traumatic
brain injury during a six-year period (1988-1993). All children had a GCS of < 8 and
required ICP monitoring. The GOS was used to determine outcome at six months.
Twenty-nine children were studied of which eight children died. In the survivors four
were severely disabled; five moderately disabled; and 12 had good recoveries. There
were no vegetative survivors. For this study, the minimum CPP associated with better
outcomes was found to lie between 41 and 57 mm Hg. Further analysis showed that
improved outcome for children was an average minimum CPP greater than 50 mm Hg.
The authors proposed that this lower minimum CPP value reflects fundamental
differences in cerebral physiology between adults and children.
As previously stated, adult brain injury studies recommend maintaining CPP above
70 mm Hg. Downard et al. (2000) retrospectively evaluated CPP and outcome in 118
brain-injured children. No patient with mean CPP less than 40 mm Hg survived. Among
patients with mean CPP in deciles of 40-49, 50-59, 60-69, or 70 mm Hg, no significant
differences in GOS distribution existed. Thus low mean CPP was lethal. In children
with survivable brain injury (mean > 40 mm Hg), CPP did not stratify patients for risk of
adverse outcomes. The authors interpreted these data to suggest that there may be a
threshold for cerebral hypoperfusion, above which a CPP must be maintained with no
marginal benefit from additional CPP elevation. It was further suggested that it may be
more efficacious to concentrate on minimizing the variation of the CPP around the
clinically targeted level rather than elevating the level.
Intracranial pressure and cerebral perfusion pressure were monitored to establish
which one is more predictive of outcome and to examine whether there are significant
threshold levels. Data were obtained from 291 severely head injured patient (207 adults
and 84 children). A CPP of 55 mm Hg and ICP of 35 mm Hg appeared to be the best
predictors in adults. For the children, the levels were a CPP of 43 to 45 mmHg and an
ICP of 35 mm Hg. The CCP thresholds of 45 mmHg for children and 55 mmHg for
adults show that children have lower thresholds. This correlates with blood pressure,
which is normally lower in children than adults (Chambers, Treadwell, & Mendelow,
A retrospective chart review of 320 consecutive pediatric patients with TBI from
1992 through 1996 was undertaken to evaluate the relationship of patient care variables
to survival and functional outcomes. The review concluded that the ability to maintain a
CPP of> 50 mm Hg was the single most important predictor of TBI survival. Thus,
these authors concluded monitoring and optimizing CPP is critical to the management of
pediatric patients (Hackbarth et al., 2002).
Forty-five acutely comatose children who sustained severe, non-missile brain
trauma were prospectively evaluated and treated to a protocol to maintain normalized
values for ICP, CPP, and arteriojugular oxyhemoglobin saturation difference (CEO2).
Six-month clinical outcomes were assessed in relation to physiological abnormalities
observed during the acute phase of injury. At six months, 37 children were in the
favorable outcome (GOS) category, whereas six children exhibited unfavorable
outcomes. Two children died, and six exhibited severe disability. No children were in a
prolonged vegetative state. The children with unfavorable clinical outcome were
significantly related to more pronounced intracranial hypertension and more profound
concomitant decreases in CEO2 (Cruz, Nakayama, Imamura, Rosenfeld, de Souza, &
Diffuse Brain Swelling
One area felt to be different in the pediatric patient is the higher frequency of
diffuse brain swelling after trauma. This is believed to be due primarily to increased
cerebral blood volume and has been attributed to partial or complete disruption of
cerebral autoregulatory mechanisms, allowing vasodilation (Aldrich et al., 1992; Bruce et
al., 1978; Bruce et al., 1981). Aldrich et al. collected data prospectively on 753 patients,
111 children, and 642 adults. Diffuse brain swelling occurred twice as often in children
as with adults. Children diagnosed with diffuse brain swelling had a mortality rate three
times as great as children without diffuse brain swelling. Adults with and without diffuse
brain swelling had similar mortality rates. This data support the concept that children
with TBI respond differently.
In 1994, Lang, Teasdale, MacPherson, and Lawrence reported a study involving a
series of 118 patients with traumatic brain swelling. This retrospective review compared
clinical findings in children with adults to determine the occurrence of neurological
deterioration and outcome. Adults were more likely to have evidence of severe initial
injury than children. The level of consciousness was similar in both groups at admission;
although, adults showed secondary deterioration more often, and thus, had a poor
outcome twice as often. The authors concluded that a child's brain may more readily
respond to injury by developing diffuse swelling. A child has a proportionately greater
intracranial CSF volume available for displacement. Therefore, the swelling appears to
be benign in children in more than 75% of the cases; whereas, in adults its appearance
signifies a poor outcome in two-thirds of the cases.
During a 15 year period (1968-1982), 434 non missile head injured patients were
retrospectively studied by necropsy (Teasdale, Graham, & Lawrence, 1989). The
researchers sought to discover the incidence of brain swelling and to relate its occurrence
to features of primary damage and to other secondary complications such as hypoxia and
raised intracranial pressure. In 40% of the cases dying within the first week of injury
death occurred without evidence of brain swelling. Raised intracranial pressure and
hypoxia were common findings. Teasdale et al. concluded that brain swelling per se was
rarely the primary mechanism of brain damage in the fatally head injured patients. They
did, however, observe a small number of fatal cases, one or two per year in a population
of 2.7 million in whom brain swelling appeared to be the major intracranial lesions. Most
of these cases were children.
Evaluation of the broad range of developmental, psychological, demographic, and
social variables, as well as medical variables, on the neurobehavioral status of children
who were more than a year post traumatic head injury was undertaken. Two variables,
CPP and premorbid learning problem, were statistically significant predictors of outcome
(Woodward et al., 1999).
Hyperemia and vasodilation are a ubiquitous response of the pediatric brain to
trauma (Piper, 1994). Cerebral hyperemia is commonly related to high ICP, caused
mainly by increased CBV (Muizelaar et al., 1989). Bruce et al. (1979), in a study of 85
children suffering from severe head trauma (GCS 5-8), found that when hyperemia
occurred in isolation, with minimal parenchymal damage, control of hyperemia with
maintenance of a normal ICP should prevent death with a permanent recovery. When
hyperemia congestion occurred, hyperventilation seemed to control hyperemia and ICP
for the first 24-48 hours. After this time frame multifocal brain edema seemed to occur
and secondary increases in ICP were found. If these ICP increases were controlled,
useful recovery occurred in 80% of the children. Bruce et al. suggested that
cerebrovascular dilatation was a primary response of the pediatric brain to trauma and the
rapid, early clinical deterioration seen is associated with acute brain swelling caused by
increased cerebral blood flow, increased cerebral blood volume, and decreased
Muizelaar et al. (1989) in a sample of 32 children studied the relationship between
hyperemia, CBF, ICP, and PVI in severe head injury. These researchers were unable to
establish a correlation between hyperemia and high ICP or low PVI in head injured
children. They concluded that real hyperemia is uncommon. Secondly, they found no
linear relationship between CBF and CBV.
A retrospective review of children admitted to a Level 1 pediatric trauma center
with severe traumatic brain injury sought to determine variables in the acute care period
associated with survival (White et al., 2001). Children (0 17 years) admitted from 1991
to 1995 with nonpenetrating traumatic brain injury and an admission GCS of< 8 were
included in the study. The first 72 hours of hospitalization were analyzed in detail for
136 patients. The primary end point was survival. The data suggested that patients with
a higher six hour GCS score were more likely to survive. Adjusting for severity of
injury, survival was associated with maximum systolic blood pressure > 135, suggesting
that supranormal blood pressures are associated with improved outcome. Thus, actual
elevation of blood pressures may be required in children suffering from traumatic brain
Sharpies, Matthews, and Eyre (1995) participated in a two- part study to understand
the pathophysiology of pediatric head trauma. They sought to test the hypothesis that
children with TBI differ from adults with TBI with increased cerebral blood flow
(cerebral hyperemia). Cerebral blood flow, arteriojugular venous oxygen difference, and
cerebral metabolic rate for oxygen were performed on 21 children with severe head
trauma (GCS < 8, mean age 8). The data from this study do not support the proposal that
children have raised cerebral blood flow after severe injury. The study determined there
is no fundamental difference between adults and children in the pathophysiological
response of cerebral blood flow to severe head injury.
The objective of the second phase of the study by Sharpies, Stuart, Matthews,
Aynsley-Green, and Eyre (1995) was to explore whether cerebrovascular resistance is
responsive to normal physiological mechanisms in children with severe head injuries, and
to determine if the overall level of cerebrovascular resistance is abnormally low in these
patients. Data indicate cerebrovascular resistance values were normal or raised in most
cases and there was significant correlation between cerebral perfusion pressure and
cerebrovascular resistance, suggesting preservation of autoregulation. They concluded
that normal cerebrovascular reactivity was often preserved in children with severe head
injury. Further, despite evidence that cerebral hyperemia is more common in the
youngest children, there was no correlation between cerbrovascular resistance and age.
No data were found to support the premise that the pathophysiology of traumatic
encephalopathy in children was essentially different from that in adults.
A study undertaken to determine the role of acute CBF alterations in the
pathophysiology of clinical head injury placed emphasis on the occurrence of hyperemia,
its time course and relationship to ICP, and its response to hyperventilation (Obrist et al.,
1984). Seventy-five adult patients with closed head injury (mean GCS score 6.2) were
studied within 96 hours of trauma. Fifty-five percent of patients developed acute
hyperemia while 45% had subnormal flows. There was a correlation between hyperemia
and the occurrence ofintracranial hypertension. Their findings suggested that cerebral
circulatory factors exerted an important influence on ICP, particularly in the presence or
absence of acute hyperemia.
The use of hyperventilation is widely debated for use in the care of patients with
TBI. Cerebral metabolic and vascular responses to head injury in children are less
predictable than previously claimed and have raised concerns about the safety of routine
hyperventilation. A prospective study assessed the effect of hyperventilation on regional
cerebral blood flow in children. Twenty-three children with TBI were treated by altering
minute ventilation to PaCO2 levels of >35, 25 35, and <25 toor. Measurements taken
showed a clear relationship between the frequency of cerebral ischemia and hypocarbia
suggesting that hyperventilation should be used with caution in children with TBI
(Skippen et al., 1997).
The review of literature for both the adult and pediatric population suffering TBI
clearly shows the lack of definite protocols for treating these patients. Past efforts to
develop guidelines for the management of patients with TBI relied on author's expert
opinion and practice experience, and thus, had the element of subjectivity. There is a
lack of class I evidence available for many current management practices. A task force
developed guidelines for the management of severe head injury using a meticulous
process relying on scientific evidence rather than expert opinion (Bullock et al., 1996;
Chestnut, 1997b; Prough & Lang, 1997). The Brain Trauma Foundation (1996) was only
able to issue three standards based on class I evidence (prospective randomized
controlled trials) and only eight guidelines based on class II evidence (data collected
prospectively with retrospective analyses based on clearly reliable data).
The recommendations issued by The Brain Trauma Foundation (1996) had a
positive influence in establishing consistent treatment therapies for severe traumatic brain
injury. Marion and Spiegel (2000) conducted a survey to determine management of
head-injured patients in 1997 and to identify differences compared with a survey done in
1991. A forty percent response rate, from a total of 3,156 neurosurgeons contacted,
showed a significant increase in the proportions of neurosurgeons who felt these patients
should have ICP monitoring (28% vs. 83%), and a decrease in the proportion who used
prophylactic hyperventilation therapy (83% vs. 36%) and steroid therapy (64% vs. 19%).
Ninety-seven percent of respondents felt that the CPP should be maintained at >70 mm
Hg, and 44% felt patients with severe head trauma should be treated at Level I trauma
There are three most commonly used treatment approaches. One is the traditional
approach, which has been to emphasize early surgical treatment of intracranial mass
lesions, meticulous critical care treatment to avoid secondary injury to the brain, and to
minimize intracranial hypertension (including the use of hyperventilation) (Robertson,
The second protocol is the CPP management, based on the vasodilatory cascade.
According to this hypothesis, a reduction in CPP- either a decrease in arterial blood
pressure, an increase in ICP or both- stimulates the cerebral vessels to dilate in an attempt
to maintain CBF. This is the normal pressure autoregulatory response to a decrease in
CPP. Because the increase in CBV that accompanies the vasodilation further reduces
CPP by increasing ICP, this sets up the cycle that leads to ever reducing CPP. An
increase in arterial blood pressure has been observed to break this cycle and reduce ICP.
Thus management is focused on maintaining CPP (Robertson, 2001).
The third approach is the Lund therapy, which emphasizes reduction in
microvascular pressure to minimize edema formation in the brain. The goals of this
approach are to preserve normal colloid osmotic pressure, to reduce capillary hydrostatic
pressures by reducing systemic blood pressures, and to reduce CBV by vasoconstricting
precapillary resistance vessels (Grande, Asgeirsson, & Norstrom, 2002; Robertson,
This study was conducted in a retrospective fashion. Medical records of
children diagnosed with traumatic brain injury that met study criteria were reviewed and
data were collected. Data were collected on a Data Collection Form (see appendix A).
This was a single center study at a large medical center in Northeast Florida. This
facility is an urban tertiary teaching hospital with a Level 1 adult and pediatric trauma
center with immediately available neurosurgical services and is affiliated with a State
A statistically determined sample size of 43 subjects provided the desired
sensitivity for a level of significance of p < 0.05 and a power of 0.95. Medical records of
subjects who met study criteria were identified until the sample size was reached.
The inclusion criteria were as follows:
1. Children less than 15 years of age at the time of injury
2. Diagnosis of non-penetrating traumatic brain injury
3. Traumatic brain injury as a result of non-penetrating blunt head trauma.
4. Glasgow Coma Scale score equal to or less than 8 upon arrival in the Pediatric
Intensive Care Unit.
5. Children with intracranial pressure monitor placement within 24 hours of traumatic
6. Children of both sexes
7. All social, economic, ethnic groupings
The exclusion criteria were as follows:
1. Infants and toddlers with unfused fontanels, i.e. those less than 18 months of age.
2. Children whose cerebral compromise was not the result of traumatic brain injury
3. Children whose cerebral compromise was the result of penetrating traumatic brain
4. Children with secondary medical diagnosis of encephalitis, Reyes syndrome, near
drowning, diabetes, cerebral palsy, or other diagnosis that affect cerebral
5. Children who did not undergo intracranial pressure monitor placement within 24
hours of the traumatic brain injury or if the invasive neurological procedure was
performed at a referring hospital.
6. Other subjects in the opinion of the investigator who were not appropriate for the
Vital signs (BP, P, T), ICP, and CPP measurements were recorded at least hourly
on the intensive care unit flow sheet. These were abstracted for the first 7 days after
admission or until the patient died.
Pediatric Blood Pressure
The major determinant of a normal blood pressure in children was maturation, not
chronological age. Increasing body size as a factor apart from age was needed to judge
the relationship of a child's blood pressure from that of the normal population. It was
therefore important that baseline blood pressure for each child be individually
determined. Blood pressure monitoring occurred by noninvasive cuff pressure and by
invasive arterial cannulation. Noninvasive and invasive blood pressure were measured
using Component Monitoring System (CMS Ml 176A, Phillips Medical Systems,
Healthcare Solutions Group, Andover, Massachusetts, U.S.A.).
Mean Arterial Pressure (MAP)
Mean arterial blood pressure was determined by the cardiac output (CO), systemic
vascular resistance (SVR), and central venous pressure (CVP) which was based upon the
relationship between flow, pressure and resistance: MAP (CO SVR) + CVP. Because
CVP is usually at or near 0 mm Hg, this relationship is often simplified to: MAP = CO -
SVR. In practice, however, MAP is not determined by knowing the CO and the SVR, but
rather by direct or indirect measurements of arterial pressure. At normal resting heart
rates, MAP can be approximated by the following equation: MAP = Pdiastolic + 1/3 (Psystolic
- Pdiastolic) where P is equal to pressure. For example, if systolic pressure is 120 mmHg
and diastolic pressure is 80 mmHg, then the mean arterial pressure will be approximately
93 mmHg. At high heart rates, however, MAP is more closely approximated by the
arithmetic average of systolic and diastolic pressure because of the change in shape of the
arterial pulse pressure (it becomes narrower) (Klabunde, 2001). Mean arterial pressure
was determined by continuous monitoring of blood pressure using Component
Monitoring System (CMS M1176A, Phillips Medical Systems, Healthcare Solutions
Group, Andover, Massachusetts, U.S.A.).
Intracranial pressure (ICP) is a function of the relative space occupied by the brain,
the cerebrospinal fluid, and the cerebral blood volume. An increase in the volume of one
must be accompanied by a reduction in one or more of the other volumes or there will be
an increase in the intracranial pressure (Shackford, 1997). Intracranial pressure
monitoring was obtained using the Camino fiber optic catheter-tip transducer (Camino
Laboratories, San Diego, California, U.S.A.). Placement of the Camino monitoring
device was accomplished in the operating room or at the bedside under sterile conditions
by a qualified neurosurgeon. The monitoring device was placed in the subdural space
with the transducer mounted on the side of the head at the level of the foramen of Monro.
Cerebral Perfusion Pressure
An intimate interdependence exists between the intracranial pressure and systemic
arterial blood pressure (SABP). The link between ICP and SABP is cerebral perfusion
pressure. Cerebral perfusion pressure is calculated as the arithmetic difference of the
mean arterial pressure and the mean intracranial pressure with both referenced to the
level of the external auditory meatus and with the patient in the flat position (Ghajar &
Hariri, 1992). Cerebral perfusion pressure was calculated manually from MAP obtained
through Component Monitoring Systems and ICP obtained through the use of the Camino
intracranial transducer (MAP ICP = CCP).
Heart Rate (Pulse)
Heart rate (pulse) was documented from readings obtained from Component
Monitoring System (CMS M1176A, Phillips Medical Systems, Healthcare Group,
Andover, Massachusetts, U.S.A.).
Body temperature was documented from readings obtained from IVAC
thermometers (CNA Medical, Rockwall, Texas, U.S.A.).
Data Collection Procedure
The Medical Center for data collection is a Level 1 pediatric trauma center. As
such the Division of Pediatric Surgery and Trauma participates in the National Pediatric
Trauma Registry (NPTR). The NPTR accepts data on injured children. On March 1,
2002, because of a lack of funding, the NPTR stopped accepting data. However, the
pediatric trauma program continued to use the NPTR data collection form to track
patients and the data were entered into a Division of Pediatric Surgery and Trauma
computerized data base. It was from this data base that subjects meeting research criteria
were identified. The beginning date was September 30, 2002 working backward in time.
Data collection were transcribed from the patient intensive care flow sheet located
in closed medical record. All data collected were part of the normal and routine care of
children with traumatic brain injury. Data were collected by the principal investigator,
the co-investigators, or by surgical research assistants. Data collected included post
injury day, time clinical values obtained, heart rate (pulse), temperature, cuff blood
pressure, arterial line blood pressure, mean arterial pressure, intracranial pressure, and
cerebral perfusion pressure. Demographic information, which included sex, age, and
ethnicity, GCS and DELTA scores were included in data collection.
All subjects had cerebral pressure devices placed with measurements taken.
Subjects had either an intracranial pressure monitor or an external ventricular drain
(EVD). Several subjects had both. For those subjects with both ICP and EVD
monitoring, the ICP value was used for data evaluation. Additionally several subjects
had blood pressure recorded by cuff, arterial-line or both. For subjects with both cuff and
arterial-line blood pressures, the arterial line blood pressure was used for data evaluation.
Data collected for each subject, systolic blood pressure, pulse, temperature, mean
arterial pressure, intracranial pressure, and cerebral perfusion pressure, were averaged to
a single value for each 24 day. Thus BP1 represents the average systolic blood pressure
for a single subject for a 24 hour period. Data were entered for each subject for each day
the intracranial monitor remained in place. Thus subject data ranged from two days up to
15 days. For purposes of data analysis, a maximum seven days data were evaluated.
This decision was based two factors, one that maximum cerebral edema peaked at 2-4
days and less than half of the subjects had intracranial pressure monitors in place for
longer than 7 days. It was felt this decrease in subject data would present biased results.
Procedure for the Protection of Human Subjects
All patients who met criteria were included for data gathering. As no clinical
interventions were proposed and data collection was by medical record review, a request
was made and accepted by the Medical Center Institutional Review Board (IRB) to wave
the informed consent requirement. Patient confidentially was maintained at all times.
Subjects were identified numerically within the context of discussions within the
Division of Pediatric Surgery and Trauma, Department of Surgery, University of Florida
and the context of a doctoral dissertation.
Descriptive analysis was used to clarify the structure of the demographic data,
obtain a simple descriptive summary, and possibly lead to a more sophisticated analysis.
Analysis of variance (ANOVA) was used to determine the relationship between cerebral
perfusion pressure and outcome in children diagnosed with traumatic brain injury.
Regression analysis was used to determine inverse linear relationship between cerebral
perfusion pressure and intracranial pressure.
ANALYSIS AND RESULTS
The purpose of this retrospective study was to evaluate CPP on outcome in children
experiencing severe blunt head trauma. Subjects were identified through the pediatric
trauma registry. Measurements included systolic blood pressure, heart rate, temperature,
mean arterial pressure, intracranial pressure, and cerebral perfusion pressure. Outcome
was measured using The DELTA disability and injury score.
Fifty-nine pediatric trauma patients were identified via the pediatric trauma registry
who met study inclusion criteria. The timeframe for the injuries spanned November 1995
through September 2002. Of the 59 subjects, one was omitted because the medical
record could not be produced by the Medical Center Medical Records Department. Three
additional subjects were eliminated after a review of their medical record indicated they
died within twenty-four hours of injury and hospitalization. A cerebral pressure catheter
was not placed prior to the child's death or the monitor was in place for less than six
hours. It was apparent from the record review these children suffered injuries so severe
that medical intervention was not a factor in their outcome.
Subject Age, Sex and Ethnicity
The mean age of the subjects was 8.4 years (SD = 4.2), with a range from 1.5 years
to 15.0 years (see figure 4.1). Four subjects were Hispanic (7.3%), thirteen were black
(23.7%) and thirty-eight were white (69%). Thirty-five of the subjects were male
(63.6%) and twenty female (36.4%). The mean DELTA score was -4.7 (SD = 3.1) with
a minimum score of -9 for those children who survived (see figure 4.2). The mean GCS
score was 5.5 (SD = 2.4) with a minimum value of 3 and a maximum value of 13 (see
Figure 4-1. Age Distribution
Count of DELTA
-9 -8 -7 -6 -5 -4 -3 -2 -1 0 (blank)
Figure 4-2. Distribution of DELTA severity and outcome score
3 4 5 6 7 8 12 13 (blank)
Figure 4-3. Distribution of Admission GCS Score
Mechanism of Injury
The majority of children (76.3%,42) suffered injuries from a motor vehicle (see
figure 4.4). Twenty children (36.6%) were passengers in an automobile involved in a
collision, either with another car or an object such as a tree or post. Twenty-two (40%)
children were hit by motor vehicles, either as pedestrians or while riding bicycles,
skateboards, all terrain vehicles (ATV), and dirt bikes. Three (5.3%) children were hit by
objects, such as a baseball or brick, two (3.6%) children were victims of child abuse, and
six (10.9%) children suffered falls. Two children were classified in the other category
(3.6%). One was injured by falling off a horse and one a near hanging. It is apparent
from this small sample that motor vehicles continue to be the greatest cause of severe
head trauma in children.
Car v. Ped v. Hit wl Child Fall Other
Car Car Obj Abu
Figure 4-4. Mechanism of Injury
Data were analyzed using SAS (SAS Institute Inc., Cary, North Carolina).
Descriptive statistics were used to obtain the summary measures for the data. The t-test
procedure, analysis of frequency, multiple regression, logistic regression, and Pearson's
correlation analysis were used to address the research questions.
Question 1. Children with non-penetrating traumatic brain injury with a cerebral
perfusion pressure averaging 50 mm Hg or higher will demonstrate better outcome, as
measured by the DELTA disability and injury score (DS), than subjects with a cerebral
perfusion pressure averaging less than 50 mm Hg.
Data were analyzed by separating subjects into two groups, those with CPP at or
above 50 mm Hg and those with CPP below 50 mm Hg. The purpose was to evaluate
outcome, as measured by DELTA scores, of the two groups. The mean DELTA score of
the group with a CPP at or below 50 mm Hg was -4.1 (SD = 3.84). The mean DELTA
score of the group with a CPP above 50 mm Hg was -4.8 (SD = 3.12). The mean age for
the group with a CPP at or below 50 mm Hg was 8.0 years (SD = 5.37). The mean age
for the group with a CPP above 70 mm Hg was 8.5 (SD = 4.15). Using stepwise
regression the data were evaluated for 7 days. On day of injury, blood pressure (p =
0.0163), temperature (p = 0.00011), MAP (p = 0.0338), ICP (p = 0.0256), and CPP (p =
0.0147) were significant in predicting CPP. On post injury day (PID) 1, MAP (p =
0.0238), ICP (p = 0.0470), and CPP (p = 0.0487) were significant predictors of outcome.
On PID 2, blood pressure (p = 0.0009) and ICP were significant (p = 0.0029). Post injury
day 3 (p = 0.0149) and PID 4 (p = 0.0003) showed only temperature as significant for
outcome. For PID 5, temperature was no longer significant, but MAP (p = 0.0334), ICP (
p = 0.0154), and CPP (p = 0.0183) again became significant. By PID 6, only blood
pressure was significant (p = 0.0028).
Question 2. Children with non-penetrating traumatic brain injury will tolerate a
lower average cerebral perfusion pressure than adults with non-penetrating traumatic
brain injury as measured by DS.
The subjects were again separated into two groups, those with CPP at or above 70
mm Hg and those with CPP less than 70 mm Hg. The mean DELTA for the group with a
CPP at or below 70 mm Hg was -4.1 (SD = 3.34). The mean DELTA for the group with
a CPP above 70 mm Hg was -5.52 (SD = 2.90). The mean age for children at or below
70 mm Hg was 7.7 years (SD = 4.44). The mean age for children above 70 mm Hg was
9.3 years (SD = 3.97). Using t-test to measure outcome, BP (p = 0.0001), pulse (p =
0.0292), MAP (p = 0.0001), ICP (p = 0.0064), and CPP (p = 0.0001) were significant on
the day of injury. For PID 1, BP (p = 0.0001), pulse (p = 0.0083), MAP (p = 0.0001) and
CPP (p = 0.0001), were significant. For PID 2, BP (p = 0.0001), MAP (p = 0.0001), CPP
(p = 0.0001) and temperature (p = 0.0003) were significant. On PID 3, BP (p = 0.0088),
MAP (p = 0.0001), ICP (p = 0.0230), CPP (p = 0.001,) and temperature (p = 0.0342)
continued to be significant. These data indicated that those children with a CPP above 70
mm Hg had a greater window of opportunity for therapeutic interventions which may
have positively impacted outcome. While these data were significant, one cannot make
the conclusion that children tolerate lower CPP pressure than adults. To answer this
question a similar retrospective data review of adults with nonpenetrating traumatic brain
injury needs to be conducted.
Question 3. There is an inverse relationship between intracranial pressure and
cerebral perfusion pressure in children with traumatic brain injury.
Using Pearson's correlation analysis it was determined that there was an inverse
relationship between ICP and CPP (R = 0.45, p = 0.0007).
A second statistical model was utilized with the following parameters. Subjects
were separated into two groups, children who died (DELTA = 0) and children who
survived (DELTA = -1 to -12). All previous variables, with the addition of GCS score,
were compared for these two groups for 4 days. These same variables were then
compared to only those children who survived. The goal was to determine which
variables impact survival and outcome.
Linear logistic regression was used to compare children who died with the children
who survived. The day of injury shows temperature (p = 0.0152) as the only significant
variable. The odds ratio of temperature for children who survived was 11.666 times
higher than children who died. For PID 1, temperature was again significant (p = 0.0167)
with an odds ratio of 6.133. On PID 2, temperature (p = 0.0398) and CPP (p = 0.0356)
were significant variables in determining survival with the odds ratio of 6.470 and 1.125
respectively. For PID 3, only temperature was significant (p = 0.0220) with an odds ratio
of 6.696. The mean temperature of children who died was 37.1 degrees Celsius with a
minimum temperature of 34.6 degrees Celsius and a maximum temperature of 38.9
degrees Celsius. The mean temperature of the children who survived was 38.1 degrees
Celsius with a minimum temperature of 37.0 degrees Celsius and a maximum
temperature of 39.7 degrees Celsius.
Children who died were then excluded from the model. The survivors were
separated into two groups, those with long term sequelae or disability (DELTA <-5) and
those with injuries that will resolve (DELTA >- 4). On the day of injury age (p =
0.0287), CPP (p = 0.0596), and GCS (p = 0.0246) were significant in determining
outcome with the odds ratio of 1.295, 1.060, and 0.619 respectively. On PID 1, ICP was
significant (p = 0.0124) with the odds ratio of ICP for those with improved outcome
1.411 higher than those with poor outcome. Post injury day 2 showed age (p = 0.0400)
and ICP (p = 0.0370) significant with an odds ratio of 2.088 and 1.891. By PID 3, no
variables were significant with relationship to outcome.
CONCLUSIONS AND RECOMMENDATIONS
Most serious childhood injuries are the result of motor vehicle accidents. With the
increased use of skateboards, in-line skates, all terrain vehicles, dirt bikes, etc. children
are at increasing risk for accidents resulting in serious injuries and permanent disability.
While parent teaching and diligent supervision continue to be the best way to decrease
childhood accidents, they will continue to occur. It thus becomes imperative for health
care providers, specializing in the care of injured children, to seek optimum clinical
The purpose of this retrospective design study was to evaluate the variables of
systolic blood pressure, pulse, temperature, mean arterial pressure, intracranial pressure,
cerebral perfusion pressure along with age, Glasgow Coma Scale score, ethnicity, and sex
on outcome, as measured by DELTA disability and severity score, on children with
nonpenetrating traumatic brain injury. Fifty-five children who met criteria were
evaluated by medical record chart review.
Discussion of Findings
Cerebral Perfusion Pressure
Subjects were separated into two groups for data evaluation. One group consisted
of subjects with a CPP at or below 50 mm Hg and a second group with CPP above 50
mm Hg. The mean DELTA score for the group with CPP at or below 50 mm Hg was
-4.1 with the mean DELTA score for the group above 50 mm Hg -4.8. This indicated
that children with a CPP at or below 50 mm Hg had a slightly better outcome than
children with CPP above 50 mm Hg. This value difference in the two mean DELTA
outcome scores is very small and functionally insignificant. This data analysis indicates
the outcome value for both groups is essentially equal indicating a CPP of 50 mm Hg
may not be the benchmark value for predicting improved outcome as measured by
When subjects were divided by a CPP at or below 70 mm Hg and a group above 70
mm Hg the results were similar. The mean DELTA score for the group at or below a
CPP of 70 mm Hg was -4.1. The mean DELTA score for the group above a CPP of 70
mm Hg was -5.5. This result indicates children at a higher CPP exhibited worse outcome
as measured by DELTA score. While the difference in the two DELTA outcome scores
was minimal it is still significant that the children with a higher CPP value showed a
Evaluation of these results indicate concern regarding the value of the DELTA
score with both the 50 mm Hg and the 70 mm Hg models. Children who did not survive
their injuries were assigned a DELTA score of zero. It is probable with this statistical
model that children who died contaminated the DELTA score. Zero is a higher value
than any of the negative DELTA score values assigned to the children who survived.
Based on the statistical model used, children who died were included in the summary
data. These children were likely assigned to the good outcome range.
Both models showed variables affecting cerebral health, BP, MAP, ICP, and CPP
significant towards outcome on PID 0 and PID 1. The values ofBP, pulse, and MAP
were significant for the first three days. In addition, intracranial pressure was significant
the first day. These data results continue to show the importance of early intervention
and appropriate resuscitation of the pediatric trauma patient immediately post injury. The
window of opportunity continues to remain within the first 48 hours post injury.
Temperature was significant on PID 0. Temperature was also found to be
significant by PID 3 and PID 4. This may represent physiological response to open
wounds. Since the majority of subjects were involved in motor vehicle accidents it is
speculated they suffered multiple injuries. The presence of injury to the abdomen or
extremities and the impact of surgical intervention is the most likely explanation for the
significance of temperature. By PID 5 temperature dropped off as a significant variable
which coincides with effective antibiotic therapy.
Survival versus Nonsurvival
Subjects were separated into two groups, those subjects who survived their injuries
and those subjects who died as the result of their injuries. Evaluation of these data
indicated age and GCS score to be significant. The mean age for children who died was
7.1 years and for children who survived was 8.3 years. Review of the literature regarding
childhood accidents indicates younger children at higher risk, especially regarding
pedestrian versus motor vehicle and as passengers in motor vehicles. As this cohort of
subjects suffered the majority of their traumatic brain injury from motor vehicles the data
are consistent with existing literature.
These same groups showed a mean GCS score of 6.1 for children who died and a
mean GCS score of 5.3 for children who survived. The result was unexpected since the
lower the GCS score at admission suggested a more seriously injured patient. The data
indicate that children with lower GCS scores have a higher probability of survival than
those children with higher GCS scores. The GCS score recorded in this retrospective
chart review was the admission GCS score. Brain injured patients often present to
Trauma Centers with compromised neurological values but show significant
improvement within hours of admission. This may have been a factor in these data
results. For future research, it might be advantageous to record GCS score values at eight
hours and 24 hours post injury. Glasgow Coma Scale score values recorded at these
times might represent more accurately the extent of the brain injury and neurological
The data analysis results for this model indicated the single most important factor
for survival is temperature. The mean temperature of children who died was 37.1 degrees
Celsius. The mean temperature of children who survived was 38.1 degrees Celsius. This
result was unexpected. There is much in the traumatic brain injury literature, some of it
covered in the literature review for this study, promoting the benefits of lower core body
temperature in patients suffering from TBI (Tokutomi et al., 2003). Conversely, a study
by Andrew et al. (2002) evaluating multiple variables did not show temperature as
significant in determining outcome, either for mortality or morbidity.
The importance of temperature in determining outcome is a significant finding
from this study. The data indicate improved survival for subjects with a higher body
temperature. This is consistent with the normal and usual care of hospitalized children.
The exception to this demand for warmth has been pediatric traumatic brain injured
patients. They have often been artificially maintained at lower than normal body
temperatures. The rationale for lowering core body temperature was to decrease the
demand for oxygen by the injured brain. The data obtained in this study bring this
therapy into question and provides an avenue for further research.
Good Outcome versus Poor Outcome
The subjects from the previously identified group of children who survived their
TBI were separated into two groups. One group was classified as the good outcome
group. This group was defined by a DELTA score of-1 to -4. The second group was
classified as the poor outcome group. This group was defined by a DELTA score of -5 to
-12. Evaluation of data for these two groups indicates age, admission GCS score, and
CPP on the day of injury as significant prognosticators of outcome.
The mean age of the poor outcome group of children was 9.6 years. The mean
admission GCS score was 4.9. Children with good outcome had a mean age of 6.2 years
and a mean admission GCS score of 6.6. It is not surprising that age and GCS score
have an inverse relationship as older children engage in more risk taking behavior and
suffer more severe injuries. The mean CPP on PID 0 of the good outcome group was 72
mm Hg while the mean CPP on PID 0 for the poor outcome group was 62.93 mm Hg.
This 10 mm Hg variance in mean values of PID 0 CPP is significant. The mean value of
72 mm HG for the good outcome subjects is consistent with the research findings
conducted primarily on adults. The findings of this study indicate that children and
adults have the same threshold of CPP on the day of injury relative to improved long term
outcome. It is thereby concluded from the data obtained in this study that children do
not tolerate a lower CPP value than adults with comparable injuries.
The results of this study differ with the results obtained in a similar retrospective
study conducted by Clifton et al. (2002). The Clifton et al. study concluded that GCS at
admission, age, MAP, fluid balance, and ICP were the most powerful variables in
determining outcome. While both studies concluded admission GCS score and age to be
significant indicators of outcome, the Clifton et al. study did not list CPP as a significant
outcome and this current study only lists CPP as a physiological variable with
significance towards outcome. What is particularly interesting regarding the Clifton et al.
study is omission of CPP form the list of significant physiological variables. Both MAP
and ICP are listed as physiological variables of significance. Since CPP is a function of
both MAP and ICP, it is surprising that CPP did not present as significant. Along this
same thought is the absence of MAP and ICP from the current study. With CPP a
significant physiological variable, one might expect MAP and ICP to also be significant.
These results involving MAP, ICP, and CPP might indicate there is less interdependence
among the three physiological variables than previously thought.
The majority of subjects were injured secondary to motor vehicle accidents. There
was initial concern during evaluation of data that the variables and outcome were
impacted by other injuries to the subjects. The concern was the possibility of
hemodynamic compromise which may influence data. However a careful review of the
raw data show that all subjects remained hemodynamically stable. Fluid resuscitation in
the field, the trauma center, and pediatric intensive care unit was sufficient to maintain
vital signs, with the exception of temperature, within normal physiological range specific
to the age of each subject.
Clear implications for clinical practice cannot be drawn from this study.
Continuing research to clarify data obtained from this study is necessary. Data obtained
from this research study show that temperature plays a greater role in the outcome of
children with traumatic brain injury then previously thought. While the need to keep
patients warm is widely known and universally practiced by health care providers, the
dramatic impact of temperature seen in this study group of pediatric head trauma patients
is unexpected. This is in contrast to some published protocols which seek to maintain
patients suffering from TBI at a lower than normal core body temperature to decrease
oxygen demands. These conflicting therapies need further evaluation.
The finding that the first 48 hours post TBI is critical to outcome is consistent with
current practice. The framework of trauma care is based on the value of immediate
interventions to stabilize patients. Continued diligence is critical to maintain
physiological variables within acceptable range. What continues to remain unclear is the
definition of acceptable range.
Limitations of the Study
Study Design Limitations
While accepted clinical treatment for traumatic brain injury was used for all
subjects there was no standardized treatment protocol used in this retrospective study.
Specific management of children with intracranial injuries varied slightly according to
the preference of the treating physician. Numerous modalities, including sedation,
administration of mannitol, hyperventilation, vasopressors, ventriculosotomy, and
decompressive craniectomy were used with varied frequencies.
A second concern was the time frame of the subject injuries. The fifty-five
children were injured in the time period from November 1995 to September 2002.
During this eight year study period it is possible and likely probable that treatment
protocols changed. It is difficult to assess whether outcome was the result of the
variables studied or subtle changes in treatment.
Retrospective studies always suffer limitations with biases such as the following:
varying degrees of doctor and nurse engagement, and varying quality of general intensive
care at different times. Further, the presence of pre-existing medical conditions or co-
morbidities for each subject was unknown. These factors may have played a significant
role in the clinical outcome of the subjects.
Statistical Analysis Limitations
A major conceptual limitation of regression techniques is they can only ascertain
relationships. Regression analysis can never address underlying causal mechanisms.
Multiple regression is most effective as a statistical model with 10 to 20 times as many
observations (subjects) as one has variables. Using this formula, the number of subjects
should be 60-120. The number of actual subjects was 55. Therefore, the small sample
size is a limitation of the study.
Strengths of the Study
A strength of the study includes location, a Level I pediatric trauma center in a
large metropolitan area. A further strength of the study was the reliability associated with
the assignment of the DELTA outcome score. The pediatric trauma program is
coordinated by a masters prepared clinical nurse specialist who is concurrently an
associate clinical professor in the College of Nursing at the associated University. All
DELTA scores were assigned by this single individual, guaranteeing consistency of
subject outcome evaluation.
Recommendations for Further Research
There has been a lack of research on a standardized treatment protocol for
children suffering traumatic head injury. The majority of research is conducted
on adult patients with the results of these studies extrapolated for care of children.
There is need for a prospective study which clearly evaluates the optimal range of
blood pressure, temperature, intracranial pressure, and cerebral perfusion pressure that
will provide for optimal outcome by decreasing the occurrence of secondary cerebral
injury. The involvement of nursing staff in the development and implementation of
treatment protocols is critical. As observed from this retrospective study, the window of
opportunity for interventions which may affect outcome is small and nursing
participation is essential.
The relationship of specific therapeutic interventions that link physiological
variables such as BP, temperature, MAP, ICP, and CPP to survival and functional
outcome is yet to be clarified. Further investigation is warranted and may require a
multiple center study to recruit a large enough patient population for a meaningful
analysis of the problem.
Children, as with adults, are at increased risk for infection from the invasive
devices that are placed to monitor essential physiological parameters. Recently a study
by Schmidt et al. (2003) has proposed mathematical models to estimate noninvasively
ICP and cerebral autoregulation. Once these mathematical models are validated for use
with adults there must be similar studies on children and infants. This highly fragile
group will benefit tremendously from increased ability to closely monitor critical
physiological variables with limited stress to the child.
Implications for Clinical Practice
The major implications for nursing practice is continued appreciation for the value
of diligent clinical assessment and the recording of accurate vital signs, especially within
the first 48 hours after injury. The prevention of secondary ischemic injury to the brain is
one of the key components in determining outcome. As seen with data from this study
the window of opportunity to positively impact outcome in pediatric head trauma patients
is narrow, with each fluctuation in physiological value critical. Aggressive observation
and assessment followed by appropriate intervention to maintain blood pressure, pulse,
temperature, intracranial pressure, and cerebral perfusion pressure within normal range
Of surprise in the data analysis was the importance of temperature in determining
outcome in the subjects evaluated. Keeping patients warm is a function of nursing care.
Nurses can assure that patients are kept warm. Nursing must take responsibility for
maintaining normal body temperatures for the pediatric patients in their care. Further, it
is imperative that nursing impress upon ancillary services that participate in care the
importance of keeping children warm. Nursing is the gatekeeper and must continue to
exercise its authority to promote optimal care for patients.
DATA COLLECTION TOOL
PatNum Date PID Time BP-A- BP- Pulse Temp MAP EVD- EVD- Camino- CaminoCPP
line cuff ICP CCP ICP
Delta Score -
GLASGOW COMA SCALE (GCS)
Best motor response Best verbal response Eye Opening
Obeying 5 Oriented 5 Spontaneous 4
Localizing 4 Confused 4 To speech 3
Flexing 3 Inappropriate 3 To pain 2
Extending 2 Incomprehensible 2 None 1
None 1 None 1
MODIFIED GLASGOW COMA SCORE FOR INFANTS
Best motor response Best verbal response Eye Opening
Spontaneous 6 Coos, babbles 5 Spontaneous 4
Withdraws to 5 Irritable cry 4 To speech 3
Withdraws to 4 Cries to pain 3 To pain 2
Abnormal 3 Moans to pain 2 No Response 1
Abnormal 2 No response 1
No response 1 -
DELTA DISABILITY AND INJURY SCORE
Function, Central Nervous System (CNS)
-1 Attention deficit, nightmares, fixation and distraction despite otherwise normal
-2 Reading, speaking learning deficit
-3 Objective CNS deficit: obtundation, paresis, seizure, etc.
Function, Musculoskeletal (M/S)
-1 Temporary deficit cast, bandage, etc.
-2 Long-term defect loss of muscle group, scarring dysfunction, limb, etc.
-3 Permanent disability limb loss, wheelchair, walker, dependent, etc.
-1 Finite, short-term drug dose antibiotics, etc.
-2 Lifelong as needed (PRN) medication antibiotics for aspienia, seizure,
-3 Lifelong, ongoing medication
-1 Temporary finite help from cast, dressing, etc.
-2 Special education, care
-3 Custodial care
GLASGOW OUTCOME SCALE
Persistent vegetative state 2
Severe disability 3
LIST OF REFERENCES
Aldrich, E., Eisenberg, H., Saydjari, C., Luerssen, T., Foulkes, M., Jane, J., Marshall, L.,
Marmarou, A., & Young, H. (1992). Diffuse brain swelling in severely head-
injured children. Journal ofNeurosurgery, 76, 450-454.
Andrews, P., Sleeman, D., Statham, P., McQuatt, Corruble, V., Jones, P., Howells, T., &
MacMillian, S. (2002). Predicting recovery in patients suffering from traumatic
brain injury by using admission variables and physiological data: A comparison
between decision tree analysis and logictic regression. Journal ofNeurosurgery,
Bauer, R., Walter, B., Torossian, A., Fritz, H., Scholonski, O., Jochum, T., Hoyer, D.,
Reinhart, K., & Zwiener, U. (1999). A piglet model for evaluation of cerebral blood
flow and brain oxidative metabolism during gradual cerebral perfusion pressure
decrease. Journal of Pediatric Neurosurgery, 30, 62-69.
Bishop, R., Bishop, J., & Rosner, M. (1994). Pressure volume index as a function of head
elevation. In H. Nagai, K. Kamiya & S. Ishii (Eds.), Intracranial Pressure IX(pp.
189-191). New York: Springer-Verlag.
Bouma, G., & Muizelaar, P. (1992). Cerebral blood flow, cerebral blood volume, and
cerebrovascular reactivity after severe head injury. Journal ofNeurotrauma, 9
(Supp 1), S333-S348.
Bouma, G., Muizelaar, J., Bandoh, K., & Marmarou, A. (1992). Blood pressure and
intracranial pressure-volume dynamics in severe head injury: relationship with
cerebral blood flow. Journal ofNeurosurgery, 77, 15-19.
Bouma, G., Muizelaar, P., Choi, S., Newlon, P., & Young, H. (1991). Cerebral
circulation and metabolism after severe traumatic brain injury: The elusive role of
ischemia. Journal ofNeurosurgery, 75, 685-693.
Bruce, D., Alavi, A., Bilaniuk, L., Dolinskas, C., Obrist, W., Uzzell, B. (1981). Diffuse
cerebral swelling following head injuries in children: The syndrome of "malignant
brain edema". Journal ofNeurosurgery, 54, 170-178.
Bruce, D., Raphaely, R., Goldberg, A., Zimmerman, R., Bilaniuk, L., Schut, L. & Kuhl,
D. (1979). Pathophysiology, treatment and outcome following severe head injury in
children. Child's Brain, 5 (3), 174-191.
Bruce, D., Schut, L., Bruno, L., Wood, J., & Sutton, L. (1978). Outcome following severe
head injuries in children. Journal ofNeurosurgery, 48, 679-688.
Bulger, E., Nathens, A., Rivara, F., Moore, M., MacKenzie, E., & Jurkovich, G. (2002).
Management of severe head injury: Institutional variations in care and effect on
outcome. Critical Care Medicine, 30 (8), 1870-1876.
Bullock, R., Chestnut, R., Clifton, G., Ghajar, J., Marion, D., Narayan, R., Newell, D.,
Pitts, L., Rosner, M., & Wilberger, J. (1996). Guidelines for the management of
severe head injury. European Journal ofEmergerncy Medicine, 3 (2), 109-127.
Burke, A., Quest, D., Shien, S., & Cerri, C. (1981). The effects of mannitol on blood
viscosity. The Journal ofNeurosurgery, 55, 550-553.
Campbell, B. (1992). Reducing traffic injury: Size of the problem and lack of research
resources. World Journal of Surgery, 16, 384-388.
Chambers, I., Treadwell, L., & Mendelow, D. (2001). Determination of threshold levels
of cerebral perfusion pressure and intracranial pressure in severe head injury by
using receiver-operating characteristic curves: An observational study in 291
patients. Journal ofNeurosurgery, 94, 412-416.
Changaris, D., McGraw, P., Richardson, D., Garretson, H., Arpin, E., & Shields, C.
(1987). Correlation of cerebral perfusion pressure and Glasgow Coma Scale to
outcome. The Journal of Trauma, 27 (9), 1007-1013.
Chestnut, R. (1997a). Avoidance of hypotension: Conditio Sine Qua Non of successful
severe head-injury management. Journal of Trauma: Injury, Infection and Critical
Care, 42 (5), S4 S9.
Chestnut, R. (1997b). Guidelines for the management of severe head injury: What we
know and what we think we know. The Journal of Trauma: Injury, Infection, and
Critical Care, 42 (5), S19-S22.
Chestnut, R. (1998). Hyperventilation versus cerebral perfusion pressure management:
Time to change the question. Critical Care Medicine, 26 (2), 210-212.
Chestnut, R., Marshall, L., Klauber, M., Blunt, B., Baldwin, N., Eisenberg, H., Jane, J.,
Marmarou, A., & Foulkes, M. (1993). The role of secondary brain injury in
determining outcome from severe head injury. The Journal of Trauma, 34 (2), 216-
Christensen, M. (1974). Acid-base changes in cerebrospinal fluid and blood, and blood
volume changes following hyperventilation in man. British Journal of
Aiithe.iih,1gy, 46, 348-857.
Clifton, G., Miller, E., Choi, S., & Levin, H. (2002). Fluid thresholds and outcome from
severe brain injury. Critical Care Medicine, 30 (4), 739-745.
Cruz, J. (1995). An additional therapeutic effect of adequate hyperventilation in severe
acute brain trauma: Normalization of cerebral glucose uptake. The Journal of
Neurosurgery, 82, 379-385.
Cruz, J. (1998). The first decade of continuous monitoring of jugular bulb
oxyhemoglobin saturation: Management strategies and clinical outcome. Critical
Care Medicine, 26, 344-351.
Cruz, J., Jaggi, J., & Hoffstad, 0. (1995). Cerebral blood flow, vascular resistance, and
oxygen metabolism in acute brain trauma: Redefining the role of cerebral perfusion
pressure? Critical Care Medicine, 23 (8), 1412-1417.
Cruz, J., Nakayama, P., Imamura, J., Rosenfeld, K., de Souza, H., & Giorgetti, G. (2002).
Cerebral extraction of oxygen and intracranial hypertension in severe, acute,
pediatric brain trauma: Preliminary novel management strategies. Neurosurgery,
50 (4), 774 779.
DeWitt, D., Jenkins, L., We, E., Lutz, H., Becker, D., & Kontes, H. (1986). Effects of
fluid-percussion brain injury on regional cerebral blood flow and pial arteriolar
diameter. Journal ofNeurosurgery, 64, 787-794.
Dietrich, H. (1954). Prevention of childhood accidents. Journal of the American Medical
Association, 156, 929-931.
DiGuiseppi, C., Rivara, F., & Koepsell, T. (1990). Attitudes toward bicycle helmet
ownership and use by school-age children. American Journal for Disease Control,
Downard, C., Hulka, F., Mullins, R., Piatt, J., Chestnut, R., Quint, P., & Mann, N. (2000).
Relationship of cerebral perfusion pressure and survival in pediatric brain-injured
patients. The Journal of Trauma: Injury, Infection and Critical Care, 49 (4), 654-
Eker, C., Asgeirsson, B., Grande, P., Schalen, W., & Nordstrom, C. (1998). Improved
outcome after severe head injury with a new therapy based on principles for brain
volume regulation and preserved microcirculation. Critical Care Medicine, 26
El-Adawy, Y., & Rosner, M. (1989). Vasodilatory cascade: ICP response to CPP level
and reduction rate. In J. Hoff & A. Betz (Eds.), Intracranial Pressure VII (pp. 842-
844). Heidelberg, Germany: Springer-Verlag.
Feldman, Z., Kanter, M., Robertson, C., Contant, C., Hayes, C., Sheinberg, M., Villareal,
C., Narayan, R., & Grossman, R. (1992). Effect of head elevation on intracranial
pressure, cerebral perfusion pressure, and cerebral blood flow in head injured
patients. Journal ofNeurosurgery, 76, 207-211.